DATA – Bayonetta – A detailed statistical breakdown of Smash 4’s most controversial character.

INTRODUCTION

I am Barnard’s Loop. I have collected tournament data at the regional and national level across most prominent countries for two years, and have used it to engineer a scoreboard for showing what characters do the best, alongside their meta saturation.

I have done a number of stats & data based projects for Smashboards & Reddit, as evidenced by the content on this blog. Some were with help, others without. I feel like I can call myself very qualified to discuss the issue at hand from a data perspective and have plenty of evidence to support it if asked.

I began to write preliminary versions of this article during May-June, possibly sooner, but priorities shifted and it became delayed. By late December, I decided to begin work on it again, and I have accelerated that work after GENESIS 5 and Frostbite 2018.

The reason, of course, is because the community has become pretty tumultuous as of late. Twitter, Smashboards, and Reddit are rife with discussion about this character with many top players or people who manage the scene chiming in.

I figured after doing experimental projects with Freezie like OrionRank, after analyzing character diversity across multiple titles, and after doing short-form analysis of character data for Kirby & Marcina, I decided it would be a good idea to extensively research the most discussed character in Smash 4.

This article will largely be based in factual data and reasonable interpretations based on that data. I will include sources and methodologies at the bottom.

TABLE OF CONTENTS

Section 1 – Tournament Data

  • 1.0: Tournament Results Breakdown
  • 1.1: Methodology of Line Chart Data
  • 1.2: The Top Ten
  • 1.3: National Data Progression – Phase 1-6
  • 1.30: Bayonetta National Progression 2016-2017
  • 1.4: Regional Data Progression –  Phases 1-6
  • 1.40: Bayonetta Regional Progression 2016-2017
  • 1.5: Total Data Progression – Phases 1-6
  • 1.50: Bayonetta Total Progression 2016-2017
  • 1.6: Mitigating Factors in 2016
  • 1.60: Adjusting for Mitigating Factors
  • 1.61: Adjusted Progression – Regional
  • 1.62: Adjusted Progression – National
  • 1.63: Adjusted Progression – Total
  • 1.7: Tournament Win Rate 
  • 1.8: Conclusion of Section 1
  • 1.80: Interpretation of Section 1 Data

Section 2 – Player Data

  • 2.0: The Players
  • 2.1: Player Histories 
  • 2.2: Player Success Breakdown
  • 2.3: Success Prior to Bayonetta
  • 2.30: Understanding the Midwest Pre-2016
  • 2.31: No Prior Records
  • 2.32: Salem’s Varied Record
  • 2.33: Louisiana’s Rise
  • 2.34: Ontario’s Anomaly
  • 2.4: “Carried” – An Analysis of a Term
  • 2.5: National Success Rate of Bayonetta Mains
  • 2.50: Adjusted Chart – Accounting for Falloff/”Fluff” Placements
  • 2.51: The Great Filter
  • 2.52: Conclusion on Mistake
  • 2.6: Power Ranking Data
  • 2.7: Conclusion of Section 2
  • 2.70: Quantum Foam Hypotheses
  • 2.71: Only One Option

Section 3 – Match-Up Data

  • 3.0: Top 10 vs. Bayonetta
  • 3.1: Character Breakdown
  • 3.2: Player Breakdown
  • 3.3: Learning from the Match-Up Chart
  • 3.4: Conclusion of Section 3

Section 4 – Meta Knight Comparisons

  • 4.0: National Ranking Comparison
  • 4.1: Power Ranking Comparison
  • 4.10: Prominence of Meta Knight
  • 4.11: Dominance of Meta Knight
  • 4.3: Reaching Critical Stages
  • 4.4: Conclusion of Section 4 

Section 5 – Pro & Anti Ban

  • 5.0: The Case Against a Ban
  • 5.1: Contextualizing Frostbite & GENESIS
  • 5.10: Frostbite 2017 vs. Frostbite 2018
  • 5.11: GENESIS 4 vs. GENESIS 5
  • 5.12: What could be the cause of scene decline?
  • 5.13: Conclusions based on Frostbite/GENESIS analysis
  • 5.2: The Data Argument
  • 5.3: The Case in Favor of a Ban 
  • 5.30: Speculative arguments
  • 5.31: Concerning data points
  • 5.32: Subjective arguments
  • 5.4: Conclusion of Section 5

Section 6 – Addressing the Community & Solutions

  • 6.0: Calming Down
  • 6.1: Monthly Polling
  • 6.2: Playing the Waiting Game
  • 6.3: The EVO Prophecy
  • 6.4: Improvements from the Playerbase
  • 6.5: Continued Observation 
  • 6.6: Conclusion 

Section 7 – Sources & Methodologies

  • 7.0: Section 1
  • 7.1: Section 2
  • 7.2: Section 3
  • 7.3: Section 4
  • 7.4: Section 5

 

SECTION 1 – TOURNAMENT DATA

1.0: Tournament Results Breakdown

Tournament results are one of the better ways to extrapolate a character’s viability. It isn’t a surefire outcome since results do not carry the many nuances that exist in movesets and player or character interactions, but there is a strong correlation between viability and strong results.

Sometimes perceived character viability and results can be skewed by public perception that is often determined by one player. Villager, Mega Man, Pikachu, Shulk, Duck Hunt, and other characters are typically mained by 1-2 strong, top level mains, and may live and die on player attendance.

If Ranai begins to slip, you can expect that Villager’s tier placement may decline with him, as his upper/high tier status was largely built on Ranai’s 2016 performances. Results numbers follow this in a more absolute sense, as no results = no numbers.

Based on tournaments gathered over the course of nearly two years, I have split several segments of time into “Phases”.

  • Phase 1: March 15th-May 15th, 2016
  • Phase 2: May 15th-August 31st, 2016
  • Phase 3: September 1st-December 31st, 2016
  • Phase 4: January 1st-April 30th, 2017
  • Phase 5: May 1st-August 31st, 2017
  • Phase 6: September 1st-December 31st, 2017

Phase 7 began on January 1st, and will end at the end of April. I will not be using it in my post here because the January data and the February data thus far paint extremely contrasting pictures, likely due to the fact that first month data on any given phase is filled with bizarre outliers since it has the least amount of data to work with.

The lead-up here is to ask a simple question: How prominent is Bayonetta across these six phases? Well, we’ll be doing a couple of things:

  • Look at results phase by phase.
  • Break down phases to regional and national levels.

This will allow see where Bayonetta is most prominent by separating regional data from national data, but also gives us a more complete picture by combining the two. By separating phases, this will allow us to see her progression over time.

  1. National Data is considered Categories 4-6. (Think A Tier > Upwards)
  2. Regional Data is considered Categories 1-3. (Think B Tier < Lower)
  3. For pre-2017 methodology, Category 1 is considered regional while Categories 2-4 are considered national.

1.1: Methodology of Line Chart Data

For the purposes of ease of collection & interpretation, we will be comparing Bayonetta to her most consistent peers – Cloud, Diddy Kong, and Sheik, with “Other” representing the cumulative points of the rest of the cast.

The basic idea: If Bayonetta was dominant, you would expect to see her comfortably edge out her peers by large margins on a consistent basis.

This national data was taken from:

  • Category 2-4 events during 2016.
  • Category 4-5 events during 2017.

This represents Majors, essentially, with some variance due to collection methods in 2016-2017. Several Category 3 tournaments classified as large regions were not used for the 2017 data, but I do not believe they would significantly affect the outcome.

We will examine:

  • National Progression (Only major data)
  • Regional Progression (Only regional data)
  • Total Progression (All data)

1.2: The Top Ten

As 2016 and 2017 went on, it became increasingly clear as I scored characters that ten consistently came up as the best.

  • Bayonetta
  • Diddy Kong
  • Cloud
  • Sheik
  • Sonic
  • Fox
  • Mario
  • Zero Suit Samus
  • Rosalina & Luma
  • Mewtwo

The composition of these characters per data sometimes differ, but there tends to be a gap between the 4th spot and the 5th spot. Sheik has declined over time to become a strong but distant 4th to her better peers while Mewtwo has sporadically lost his spot to Marth, Meta Knight, Ryu, and others over certain periods of time.

When I refer to the “top 10”, I’m referring to the ten listed above. For line charts, I will use 4 characters + other. For pie charts, I will use 10 characters + other.

1.3: National Data Progression – Phases 1-6

complete.png

You can see that these four characters have some volatility. The total collapse of “Other” is due to the smaller number of major events recorded for this methodology during the last four months of 2017. Due to how my methodology for collecting works, the 2GG Championship itself was not used since it is not a traditional tournament bracket.

Breaking it down, let’s remove Other and look at the 4 characters themselves:

bayonettacomparison.png

Here, we see that Bayonetta per national level events was roughly on the same level as her peers during phase 1 before dropping during phase 2 due to her nerfs affecting her perception. This gradually recovers in phase 3, and stays stagnant in phase 4 despite a strong performance for the character at GENESIS 4.

Starting phase 5, during the Summer of Smash, Bayonetta hooks upwards due to the large number of high-point value events that took place. In phases 5 and 6, she marches in near lockstep with Cloud, with Diddy in third and Sheik in fourth.

This has been known to be the case for a long time. Sheik has stagnated and declined since her nerfs, Diddy has fallen behind, but Bayonetta and Cloud appear to have a similar share of representation.

Had I included Category 3 data in 2017, I doubt this would have changed, as all Sumabato/Umebura events are considered Category 3, as well as a number of B-Tier equivalent events where Cloud performed about as strongly at a casual glance.

bayocloud.png

The effective lockstep between the two characters during 2017 is pretty remarkable. However, I can’t call it unexpected, as there have consistently been three Cloud players in the top 20.

1.30: Bayonetta Progression Between 2016 and 2017 – National

phase1to3national

phase4to6national.png

Phase 1-3 – National

  • Diddy Kong: 8.5%
  • Sheik: 8.3%
  • Cloud: 7.9%
  • Mario: 5.8%
  • Sonic: 5.6%
  • Fox: 5.4%
  • Bayonetta: 5.2%
  • Zero Suit Samus: 4.9%
  • Mewtwo: 4.1%
  • Rosalina: 3.9%
  • Other: 40.3%

Phase 4-6 – National 

  • Bayonetta: 9.9%
  • Cloud: 9.9%
  • Diddy Kong: 8.7%
  • Sheik: 6.5%
  • Fox: 5.6%
  • Sonic: 5.5%
  • Zero Suit Samus: 5.2%
  • Mario: 4.6%
  • Rosalina: 4.4%
  • Mewtwo: 3.3%
  • Other: 36.3%

Changed Values between 2016 and 2017 

  • Bayonetta +4.7%
  • Cloud +2%
  • Mario +1.2%
  • Rosalina +0.5%
  • Zero Suit Samus +0.3%
  • Diddy Kong +0.2%
  • Fox +0.2%
  • Sonic -0.1%
  • Mewtwo -0.8%
  • Sheik -1.8%
  • Other: -4%

We can see the effect sudden prominence has on Bayonetta’s influence over the meta. While I will be discussing mitigating factors (as the 4.7% figure really isn’t accurate for certain reasons) it’s clear that she experiences a big jump between the end of 2016 and the end of 2017.

Now, the jump is easier to visualize in part because Bayonetta’s national results in 2016 were quite weak. Indeed, community discussion regarding Bayonetta as a character really ceased after May 15th and stayed silent for a period of time.

Some concerns were brought up when Salem began to perform well regionally, but Salem was not nationally successful until mid-2017 on any consistent basis. I would argue that GENESIS 4 marked the first major, unquestionable success for the character.

Salem winning Collision XIV was a big deal at the time, but his inability to replicate this at multi-day major events was a major roadblock in the character’s progression. This is similar the case for Pink Fresh’s win at KTAR Saga.

1.4: Regional Data Progression – Phases 1-6

completeregion

Regional data was easily the biggest headscratcher for me to collect. Since this removes all major-level data, the number of oddities that exist at lower level play really come to the forefront. R.O.B., for example, was (and still is, to an extent) an extremely prominent top 10-15 placing character due to a lot of good non-major performances.

That aside, this is the most volatile section of the data and requires a better breakdown.

bayoregion.png

This is the first time we can see Bayonetta having something of an edge when compared to her peers in the line chart, with a slow but steady progression starting from phase 3 onward. She far exceeds her initial #1 status in phase 1, but meta shifts and differing collection methodologies are the best explanation for weird discrepancies with 2016 data.

Diddy notably dominates the regional data during phase 5. Considering this includes a number of Category 3 events, ZeRo is present and partially responsible for spike in Diddy Kong’s favor. I can’t explain the Cloud drop at this point in time, and Sheik (as always) lags behind to be just ahead of other top tiers but definitively lower than the so-called “Top 3.”

While Bayonetta has a spike at the end of the year, it’s hard to determine if this will continue. As you can plainly see from the charts presented thus far, the data over the course of the last 2 years is very chaotic and sometimes inconsistent.

Nowhere is this more prevalent than regional data, where standards for collection have shifted the most. The reason for a region-exclusive section was to figure out if the character dominated the regional level, and this piece of evidence suggests she doesn’t.

1.40: Bayonetta Progression Between 2016 and 2017 – Regional

phase1to3regional.png

phase4to6region.png

Phase 1-3 – Regional

  • Diddy Kong: 6.9%
  • Bayonetta: 6.5%
  • Cloud: 5.8%
  • Sheik: 5.4%
  • Fox: 5.4%
  • Sonic: 5.4%
  • Mario: 4%
  • Rosalina: 3.6%
  • Zero Suit Samus: 3.1%
  • ROB: 3.0%
  • Other: 51.1%

Phase 4-6 – Regional

  • Bayonetta: 9.1%
  • Diddy Kong: 8.7%
  • Cloud: 7.9%
  • Sheik: 6.6%
  • Sonic: 5.1%
  • Fox: 5.1%
  • Mario: 5%
  • Zero Suit Samus: 3.2%
  • Rosalina: 3%
  • Mewtwo: 3%
  • Other: 43.3%

Changed values between 2016 and 2017

  • Bayonetta +2.6%
  • Cloud +2.1%
  • Diddy Kong +1.8%
  • Sheik +1.2%
  • Mario +1%
  • Zero Suit Samus +0.1%
  • Fox -0.3%
  • Sonic -0.3%
  • Rosalina -0.6%
  • Other -7.8%

For a time save, I didn’t bother looking at R.O.B.’s decline (R.O.B. will be the subject of a later article…) or Mewtwo’s rise between the years because I forgot and really didn’t care considering the subject matter.

So, while Bayonetta’s lead over her peers is more definitive in the regional category than it is in the national category, her rate of increase is smaller. It’s easy to see why – she did not lag behind regionally in 2016 the same way she did nationally.

You might attribute this to Salem, but only Category 1 events under the 2016 methodology were considered regional for this project. This means that events at the near bottom saw Bayonetta doing quite well that year.

I at least partially attribute this to Europe, which dealt with prominent Bayonetta players. Additionally, the full usage of a top 16 at a category 1 event (a practice I would later drop because it’s dumb and stupid) also meant that a lot of “straggler” data happened, so Bayonettas scoring 9th or 13th got the character a bonus 1-2 points on occasion.

This should be uniform among most of the top tiers, so it still means Bayonetta was especially prevalent regionally.

The increase itself is… more mundane, honestly. We’ll cover this in greater detail during mitigation & adjustments, but she only has a 0.5% edge on Cloud and an 0.8% edge on Diddy when it comes to increases. The common factor here is that the top 4 are regionally eating up more and more spots at the expense of meta diversity as evidenced by the sharp decline of “Other.”

Consider this: “Other”, that is, characters outside of the Top 10, had a much less sharp decline on a national level (higher skill level) than they did on a regional level (lower skill level.)

1.5: Total Data Progression, Phases 1-6

bayograph1

There’s not much to distinguish this from the above two charts except small spikes in phase 5 and Bayonetta hitting the bottom during phase 2. As stated earlier, “Other” represents all other character scores, meaning it shifts with the amount of points accumulated from phase to phase.

Removing Other, we can see the top 4’s progression:

bayprpogress2.png

Both Diddy and Bayonetta experience a spike during phase 5, which took place over the summer of 2017, followed by a steep drop for Diddy and a less than steep drop for Bayonetta.

However, she does not extend past her peers in any significant fashion. She experienced a brief but significant decline during phase 2 after her nerfs but swiftly recovered. Sheik has notably stagnated below her peers since phase 3, but Sheik will be the subject of a future article.

Bayonetta’s overall lack of huge movements can be seen on the Bayo-Other chart:

bayoother.png

Her progression is noteworthy, but she does not race past the top tiers even at her best on data recorded so far. It’s plausible that a number of factors damaged her phase 6 numbers and that phase 5 is more representative of her capabilities on the national level, but we’ll need phase 7 data to determine that.

1.50: Bayonetta Progression Between 2016 and 2017 – Total

phase1to3total

phase4to6total

Phase 1-3 Total

  • Diddy Kong: 7.6%
  • Sheik: 6.8%
  • Cloud: 6.8%
  • Bayonetta: 5.9%
  • Fox: 5.4%
  • Sonic: 5.3%
  • Mario: 4.7%
  • Zero Suit Samus: 3.8%
  • Rosalina & Luma: 3.6%
  • Mewtwo: 3%
  • Other: 47%

Phase 4-6 Total

  • Bayonetta: 9.3%
  • Diddy Kong: 8.7%
  • Cloud: 8.6%
  • Sheik: 6.5%
  • Fox: 5.5%
  • Sonic: 5.2%
  • Mario: 4.8%
  • Zero Suit Samus: 3.9%
  • Rosalina & Luma: 3.5%
  • Mewtwo: 3%
  • Other: 40.9%

Change values between 2016 and 2017

  • Bayonetta +3.4%
  • Cloud +1.8%
  • Diddy Kong +1.1%
  • Mario +0.1%
  • Zero Suit Samus +0.1%
  • Fox: +0.1%
  • Mewtwo = %
  • Rosalina -0.1%
  • Sonic -0.1%
  • Sheik -0.3%
  • Other -6.1%

This exists as the logical in-between from significant growth (national) and small growth (regional.) Not much else to talk about, it speaks for itself and the detailed breakdowns were above.

1.6: Mitigating factors of 2016 data

While we’ve established a good amount of data to show Bayonetta has risen, my concern is that direct comparisons between 2016 and 2017 progression are flawed.

  • 2 months and 15 days are missing from the 2016 data. One of those months include a period in which the character did not exist. This is compared to a full 12 months in 2017.
  • Bayonetta’s meta was effectively halted for a short period of time after May 15th. While this has obviously reversed at least partially by KTAR Saga, phase 2 scores do not accurately represent the character’s meta presence in all other phases.

1.60: Adjusting for mitigating factors

In order to balance out the score, I decided to adjust her data in phase 2 by making the % the mean average of her representation in phase 1 and 3. I did this for Regional, National, and Total, and then re-compared the 2016 data to the 2017 data.

This isn’t a perfect way of managing the situation and still does not account for phase 3 in and of itself being slightly lower in terms of representation. However, it’s as far as I’m willing to go and gets the idea across and at least will demonstrate that the rise isn’t as significant as initially indicated.

1.61: Adjusted Progression – National

adjustednational

Compared to the phase 4-6 charts, we get the following:

  • Bayonetta +2.9%
  • Cloud +2.1%
  • Rosalina +0.6%
  • Zero Suit Samus +0.4%
  • Diddy Kong +0.4%
  • Fox +0.3%
  • Sonic = 0%
  • Mewtwo -0.7%
  • Mario -1.1%
  • Sheik -1.7%
  • Other: -3.3%

Change in Bayonetta Progression after adjustment – Minus 1.8% 

By modifying phase 2 data, this strengthens her 2016 position to better mesh her with phases 1 and 3. In this case, it makes her data stronger for the year, softening how strong her improvements look from 2016 to 2017.

While still holding an edge over her peers, her rise isn’t much higher than Cloud’s. This is still the most concerning figure we’ve bumped into so far and will come back up later as one of the several “small detail” instances that indicate Bayonetta could be a issue.

While the exact reason for making these adjustments will be clear by Section 4, the point right now is to demonstrate that you can’t reasonably expect the same progression rate year to year. Meaning, you can’t expect a 4.7% progression by the end of 2018 just because we saw one in 2017.

1.62: Adjusted Progression – Regional

adjustedregional

Compared to the Phase 4-6 charts, we get the following:

  • Diddy Kong +1.9%
  • Cloud +1.9%
  • Bayonetta +1.3%
  • Sheik +1.3%
  • Mario +1.0%
  • Zero Suit Samus +0.2%
  • Sonic = 0%
  • Fox -0.2%
  • Rosalina -0.5%
  • Other -7.0%

Change in Bayonetta Progression after adjustment – Minus 1.3%

This is probably her worst data point in the entire article. There’s not much to discuss here because no character can even manage a 2% progression rate, and Bayonetta in particular is totally unremarkable if we’re only looking at the regional data. This is considering that her rate of increase regionally wasn’t that threatening to begin with.

1.63: Adjusted Progression Rate – Total

adjustedtotal

Compared to the Phase 4-6 charts, we get the following:

  • Bayonetta +1.9%
  • Cloud +1.9%
  • Diddy Kong +1.2%
  • Fox +0.2%
  • Zero Suit Samus +0.2%
  • Mario +0.1%
  • Mewtwo +0.1%
  • Sonic = %
  • Rosalina -0.1%
  • Sheik -0.2%
  • Other: -5.4%

Change in Bayonetta Progression after adjustment – Minus 1.5%

The conclusion that can be made at this point is that only the national data indicates a distinct rise. In both Regional & Total, Bayonetta is either tied or explicitly behind her peers.

1.7: Tournament Win Rate

Across 860 tournaments documented from March 15th 2016 to December 31st 2017, we have Bayonetta’s specific win rates:

     National, 2016 (Among 60) 

  • Win: 5 (8.3%)
  • Did not win: 55 (91.7%)

     Regional, 2016  (Among 160) 

  • Win: 18 (11.2%)
  • Did not win: 142 (88.8%)

    National, 2017 (Among 24) 

  • Win: 3 (12.5%)
  • Did not win: 21 (87.5%)

    Regional 2017 (Among 606)

  • Win: 58 (9.6%)
  • Did not win: 548 (90.4%)

The difficulty in comparing 2016 and 2017 methodologies comes into play here, since 2016 has more tournaments considered “national” so I could have even data to work with to begin with.

However, as the number of tournaments collected greatly expanded in 2017, her win rate went down regionally while it increased nationally. If she was dominating lower level of play, you would expect her prevalence to expand to some extent as I began to include monthly events, small regionals, or sampled obscure regions where entrant counts are very low.

As it stands, her win rate is pretty low across all segments.

1.8: Conclusion of Section 1

1.80: Interpretations of Section 1 Data

  • Bayonetta is more relevant nationally than she is regionally.
  • Bayonetta is not significantly more prevalent in the metagame than Diddy Kong or Cloud are.
  • Her rate of increase either ties, is behind, or only marginally exceeds that of Cloud after adjustments.
  • Her rate of increase may be overestimated even after adjustments due to differing methodologies in tournament scoring & collection between 2016 & 2017, as well as mentality differences within the metagame.
  • She does not win the vast majority of events and never has at any documented level of play, though this does not preclude her from being a problem.

Bayonetta is the clear #1 character based on available tournament data, but this is not but a significant margin in terms of results. Her national progression is the most concerning statistic uncovered by this extensive breakdown of her tournament results.

 

SECTION 2 – PLAYER DATA

 

2.0: The Players

In Smash 4, there are currently eight prominent Bayonetta mains on the Panda Global Rankings. This is as of v4, and in the interest of using as much data as possible, I should call attention to the rest of 2017 and point out 9B. As such, I will also be using OrionRank data to better flesh out the entire year.

The players we will be looking at individually:

  • MVG Liquid | Salem (#2, PGRv4)
  • P1 | Tweek (#4, PGRv4)
  • EMG | Mistake (#13, PGRv4)
  • Abadango (#15, PGRv4)
  • P1 | Captain Zack (#18, PGRv4)
  • ERG | Lima (#24, PGRv4)
  • 9B (#27, OrionRank 2017)
  • JK (#37, PGRv4)
  • tyroy (#40, PGRv4)

There are effectively 9 top 50 players who use the character. However, 9B is currently inactive, and JK is retired. As a result, there are 7 active primary Bayonetta players who are ranked Top 50 who attend tournaments.

For the first portion of Section 2, we will be examining these 9 players’ histories in bracket.

Preemptive notes:

  1. There is a chance the players I’m mentioning are reading this, or at least people who otherwise have good knowledge of their records. If I have made a mistake, please correct it. Pre-March of 2016 data used is liable to be inaccurate as I only started regularly collecting data in mid-late March of 2016.
  2. “National” is defined per Section 1 as an A Tier or greater per PGR methodology or Category 4 or greater per my methodology for 2017 onwards.
  3. Dabuz has a secondary/pocket Bayonetta. He may well earn a spot on this list in the future, but as with Locus, it hasn’t really been prominently featured enough in tournament to make a proper judgement call on yet. There’s not enough data to work with.

2.1: Player Histories

   #1: Salem

  • PGRv4: 2nd
  • OrionRank 2017: 3rd
  • Major Event Wins: 3 (EVO 2017, DreamHack Atlanta, 2GGC: Fire Emblem Saga)
  • Original Main: Various
  • Co-Main?: No
  • Active Before February of 2016?: Yes (Sporadic)
  • Peak Placement pre-Bayonetta pickup: 7th, SKTAR 4
  • Brawl Veteran?: Yes (18th, SSBBRank 2014)
  • First Successful Outing post-Bayonetta: Collision XIV, 1st (7 Months post-Bayonetta pickup)
  • First National Top 8 post-Bayonetta: Frostbite 2017, 5th (12 Months post-Bayonetta pickup)
  • First National Top 4 post-Bayonetta: 2GGC: Nairo Saga, 2nd (16 Months post-Bayonetta pickup)

Note: KTAR XIX, where Salem placed 2nd, is varied by major definition. Per PGR, it is not considered a major. Per less stringent definitions of my database prior to 2017, it constitutes as a Category 3 (2016 definition.) However, I would consider the skill pool of this event to be extremely top-heavy, a trait resembling the bulk of Category 3 & B-Tier events of 2017.

   #2: Tweek

  • PGRv4: 5th
  • OrionRank 2017: 6th
  • Major Event Wins: 1 (2GGC: MKLeo Saga)
  • Original Main: Bowser Jr., Wario
  • Co-Main?: Yes, Cloud
  • Active Before February of 2016?: Yes
  • Peak Placement pre-Bayonetta pickup: Various (2nd, Shine 2017. 4th, EVO 2017. Etc.)
  • Brawl Veteran?: No
  • First Successful Outing post-Bayonetta main: Frostbite 2018, 2nd (4 Months post-Bayonetta pickup)
  • First National Top 8 post-Bayonetta: Frostbite 2018, 2nd (4 Months post-Bayonetta pickup)
  • First National Top 4 post-Bayonetta: Frostbite 2018, 2nd (4 Months post-Bayonetta pickup)

This is considering the sole event thus far at the national level where Tweek has primarily used Bayonetta. He has used the character multiple times in bracket prior to this particular event, but never as a primary on the national level. Frostbite marked an occasion where the majority of his Top 32 sets saw him using Bayonetta.

This is also considering “pick-up” to be when the character first saw use in national events. However, I believe Tweek has commented before that he’s been training his Bayonetta for well over a year, and without VOD access to many pools sets I many be off in my 4 month estimation.

   #3: Mistake

  • PGRv4: 13th
  • OrionRank 2017: 23rd
  • Major Event Wins: 0
  • Original Main: Zero Suit Samus
  • Co-Main?: No
  • Active Before February of 2016?: No
  • Peak Placement pre-Bayonetta pickup: 33rd, Frostbite 2017
  • Brawl Veteran?: No
  • First Successful Outing post-Bayonetta main: 3rd, Low Tier City 5 (6 months post-Bayonetta pickup)
  • First National Top 8 post-Bayonetta: Super Smash Con 2017, 5th (6 months post-Bayonetta pickup)
  • First National Top 4 post-Bayonetta: 2GGC: SCR Saga, 4th (6 months post-Bayonetta pickup)

   #4: Abadango

  • PGRv4: 15th
  • OrionRank 2017: 13th
  • Major Event Wins: 1 (Pound 2016)
  • Original Main: Various
  • Co-Main?: Yes, Mewtwo
  • Active Before February of 2016?: Yes
  • Peak Placement pre-Bayonetta pickup: 1st, Pound 2016
  • Brawl Veteran?: Yes (Unranked)
  • First Successful Outing post-Bayonetta main: 1st, Umebura 27 (2 months post-Bayonetta pickup)
  • First National Top 8 post-Bayonetta: 2GGC: MKLeo Saga (6 months post-Bayonetta pickup)
  • First National Top 4 post-Bayonetta: EVO Japan 2018 (8 months post-Bayonetta pickup)

   #5: Captain Zack

  • PGRv4: 18th
  • OrionRank 2017: 16th
  • Major Event Wins: 0
  • Original Main: Peach
  • Co-Main?: No
  • Active Before February of 2016?: Yes
  • Peak Placement pre-Bayonetta pickup: 25th, MLG World Finals 2015 OR 33rd, CEO 2015
  • Brawl Veteran?: No
  • First Successful Outing post-Bayonetta main: 5th, Clutch City Clash (6 months post-Bayonetta pickup)
  • First National Top 8 post-Bayonetta: 7th, UGC Smash Open (10 months post-Bayonetta pickup)
  • First National Top 4 post-Bayonetta: 4th, GENESIS 4 (11 months post-Bayonetta pickup)

#6: Lima

  • PGRv4: 24th
  • OrionRank 2017: 33rd
  • Major Event Wins: 0
  • Original Main: N/A
  • Co-Main?: No
  • Active Before February of 2016?: No
  • Peak Placement pre-Bayonetta pickup: N/A
  • Brawl Veteran?: No
  • First Successful Outing post-Bayonetta main: 9th, EVO 2017 (9 months post-Bayonetta pickup)
  • First National Top 8 post-Bayonetta: 7th, MKLeo Saga (12 months post-Bayonetta pickup)
  • First National Top 4 post-Bayonetta: N/A

#7: 9B (INACTIVE)

  • PGRv4: N/A (Inactive)
  • OrionRank 2017: 27th
  • Major Event Wins: 0
  • Original Main: Ryu
  • Co-Main?: No
  • Active before February of 2016?: Yes
  • Peak Placement pre-Bayonetta pickup: 5th, Umebura F.A.T.
  • Brawl Veteran?: Yes (1st, SSBBRank 2014)
  • First Successful Outing post-Bayonetta main: 5th, Shots Fired 2 (1 month post-Bayonetta pickup)
  • First National Top 8 post-Bayonetta: 5th, Shots Fired 2  (1 month post-Bayonetta pickup)
  • First National Top 4 post-Bayonetta: 4th, Umebura Japan Major (14 months post-Bayonetta pickup)

#8: JK (RETIRED)

  • PGRv4: 37th
  • OrionRank 2017: 38th
  • Major Event Wins: 0
  • Original Main: N/A
  • Co-Main?: No
  • Active before February of 2016?: No
  • Peak Placement pre-Bayonetta pickup: N/A
  • Brawl Veteran?: No
  • First Successful Outing post-Bayonetta main: 2nd, 2GG: Pay it Forward (9 months post-Bayonetta pickup)
  • First National Top 8 post-Bayonetta: 7th, 2GGC: GENESIS Saga (11 months post-Bayonetta pickup)
  • First National Top 4 post-Bayonetta: N/A

#9: tyroy

  • PGRv4: 40th
  • OrionRank 2017: 46th
  • Major Event Wins: 0
  • Original Main: Sheik, Meta Knight
  • Co-MaiN?: No
  • Active before February of 2016?: Yes
  • Peak Placement pre-Bayonetta pickup: E2C20 (7th)
  • Brawl Veteran?: No
  • First Successful Outing post-Bayonetta main: 3rd, Midwest Mayhem (1 month post-Bayonetta pickup)
  • First National Top 8 post-Bayonetta: N/A
  • First National Top 4 post-Bayonetta: N/A

2.2: Player Success Breakdown

Players that were significantly successful prior to picking up Bayonetta

  • Tweek
  • Abadango
  • 9B

Players that were moderately/mildly successful prior to picking up Bayonetta

  • Salem

Players that did not play prior to her release

  • Lima
  • JK

Not Enough Data 

  • tyroy

Players that were unsuccessful prior to picking up Bayonetta 

  • Mistake
  • Captain Zack

Average month span before success with Bayonetta (Mean)

  • First Success: 5 Months
  • Top 8: 6.75 Months
  • Top 4: 9.83 Months

Average month span before success with Bayonetta (Median)

  • First Success: 6 Months
  • Top 8: 6.5 Months
  • Top 4: 9.5 Months

2.3: Success Prior to Bayonetta

For the purpose of the some of the points being made here, it would be fair to say that we have extensive documentation to prove that both Tweek and Abadango were very good players prior to adding Bayonetta to their arsenal.

Tweek appeared to require very little adjustment in his transition from using Bayonetta infrequently to full on primary usage during Frostbite 2018’s Top 32.

9B was also successful in Japan with Ryu, in particular at Umebura F.A.T. where he defeated Vinnie and Ally. Data beyond this is limited, but he otherwise performed consistently in Japan. His one red mark is a 49th at GENESIS 3.

2.30: Understanding the Midwest Pre-2016

People who are familiar with the Midwest today understand it as one of the best regions in the world. This is due to a combination of factors, but it’s heavily supported by numerous top players moving to Midwestern states in the last year.

However, prior to the Midwest Mayhem series, which is now a staple of U.S. regional events, the region was extremely underdeveloped. It held few large scale regional events and its players saw minimal exposure across the U.S. during 2015. In fact, it was not considered a prime candidate for GENESIS 3’s Regional Crew Battle.

Per the trailer for the G3 Crew Battle, it was one the last added to the competition. In other words, the Midwest developed as a region and began to acquire national exposure as Bayonetta was released.

This is a long-winded way of explaining why people had probably never heard of tyroy. People are likely extrapolate that he only rose due to picking up Bayonetta, but what little data exists indicates he did fine at the same Chicago weekly events players like JJROCKETS and Ned played at.

Tyroy pre-Bayo Tournament Record 

Poplar Creek Bowl Series:

  • 7: 3rd
  • 8: 2nd
  • 9: 9th
  • 10: 4th
  • 11: 4th
  • 28: 2nd
  • 29: 2nd
  • 31: 2nd

Other:

  • E2C20: 7th
  • Ignition 14: 1st

He essentially falls into a similar category as JK and Lima, except the small data suggests that he built himself to become one of Chicago’s best players prior to Bayonetta’s release. With Sheik being considered #1 at the time and Meta Knight considered a plausible top 5 or even 3 character pre-nerf, we can tell that tyroy prefers to run the highest tiered characters.

2.31: No Prior Records

It is impossible to determine how well Lima or JK would’ve done before they picked up Bayonetta. As far as I can find, JK first appeared at a Vegas weekly he won in late February of 2016, and Lima’s attendance data seems to have started during DFW’s Shockwave series in November of 2016.

As either player could be reading this, feel free to contact me if I’m wrong and I will add an addendum. However, I don’t think it will significantly change the conclusions made by the data currently available.

2.32: Salem’s Varied Record

Salem is a pretty weird case overall. He’s the only player to win majors as a Bayonetta main and his kickstart was a serious of extremely consistent regional performances during Fall of 2016. However, these did not translate into top 8s at majors until Frostbite 2017 and it took until Nairo Saga for him to become the significant meta threat we know today.

He had a long incubation period and continues to be the most consistent player of the character. His record prior, circa 2015? Kind of a mess: He used a various assortment of characters as far as I can find and had results all over the map, and this messier results table extended into the first few months of him maining Bayonetta.

2.33: Louisiana’s Rise

Zack and Mistake typically raise eyebrows the most, as both became prominent players after picking up the character. I described in a brief history in a reddit post on the different between Zack and Mistake, but the short of it:

  • Mistake’s successes were back-to-back in the same month.
  • Zack’s rise was much slower, going on a more natural scale of progression from good placements > upsets > good tournament performances.

Zack’s “profile” above really doesn’t show anything out of the ordinary, and I believe he only catches attention for it due to his character choice. His results prior weren’t even terrible for what it’s worth, but look at his rise paired by Samsora over 2016:

  • CEO 2016: Samsora 49th, Captain Zack 97th (Unknown if Zack used Bayo or Peach)
  • Clutch City Clash: Samsora 2nd, Captain Zack 5th
  • 2GGT: Abadango Saga: Samsora 65th, Captain Zack 9th
  • Aftershock 2016: Samsora 4th, Captain Zack 2nd
  • The Big House 6: Samsora 25th, Captain Zack 17th
  • KTAR XIX: Samsora 5th, Captain Zack 13th
  • UGC Smash Open: Samsora 17th, Captain Zack 7th
  • 2GGT: ZeRo Saga: Samsora 17th, Captain Zack 25th
  • GENESIS 4: Samsora 49th, Captain Zack 4th

Both are players hailing from the Louisiana scene who sort of just came into existence as national threats at Clutch City Clash despite both suffering from poor performances at CEO. The placement disparity until GENESIS 4 set a standard isn’t very significant sans an outlier in Abadango Saga.

The criticism toward’s Zack’s skill as a player seems fairly unfounded as he developed into a top player over the course of several months after picking up Bayonetta and seemed to develop as a player closely alongside his Louisiana peer Samsora, who uses a Mid/Upper Tier character.

2.34: Ontario’s Anomaly

Among the nine players used who are ranked and demonstrate a capacity for success, Mistake has the least number of factors going for him to indicate likelihood of success.

  • He was not a Brawl veteran.
  • He did not have a significant track record prior to 2017.
  • He used a high/top tier to middling or no true success.
  • He did not come to the forefront slowly.

Some have gone on to describe Mistake as a player who only succeeds because he mains Bayonetta. In order to deconstruct this idea, we need to break this into two sections:

  1. Breaking down the term “Carried”.
  2. Breaking down hard Power Ranking data and the national success rate of Bayonetta mains in general to see if players like Mistake commonly occur in the metagame.

Let’s start with #1.

2.4: “Carried” – An Analysis of a Term

What does the term “Carried” mean?

Carried is a term generally directed at players who win with a character who is extremely specialized or obscure, or directed at players who attain significant success with a very powerful character (often after not having as much success prior to using them.)

It can be subjective from player to player and is probably used in some form or another across multiple E-Sports. However, I have mostly seen it in one specific context. This is when it is directed at Bayonetta players.

What is the term supposed to mean when applied to Bayonetta players?

It’s usually intended to demean or denigrate that player’s accomplishments by boiling down their success exclusively to their character. This is meant to imply that, if that player did not use Bayonetta, they would not be successful.

bayoimage
Reddit example #1.
redditmoron3216
Reddit example #2.

This argument is primarily used by people who have a particular distaste for the character. The idea, for those touting the concept, is that these players (of whom many of the poster do not like) only are relevant due to a character that particular poster dislikes.

I think most people following along at this point might see the issue. This is not a judgement call based on analysis, it is based entirely in emotion. In particular, it is resentment that these players are successful at all, as these posters would greatly prefer if they weren’t, and want a single target reason they can pinpoint.

This means:

  1. They can imply these successes don’t count and denigrate the skill of the players using them.
  2. They can advocate a pro-ban position for the character, believing it will fix their problem with the metagame.

The latter is supported by more aggressive and direct comments specifically directed at the most controversial Bayonetta players. If you’ve been on Reddit or Twitter, especially in the last month, you’ve seen these, so I won’t exhaustively reference every example.

Of course, not all pro-ban people apply this argument, but it’s not uncommon.

The most immediate problem is that by stripping any player of their main, their results will likely decline. This may not necessarily be the case if someone is exceptionally prominent with multiple characters, such as MKLeo, where one stripped part could still feasibly see him taking a major.

However, the majority of players in the Top 100 are primary solo mains.

meta-chart
Source: OrionRank 2017

Most players are character specialists. They might have secondaries, but for good reason do not primarily use them in bracket outside of specific match-ups, either character or player based.

If the point being made by the non-analytical statement is that a tournament post-ban would see a drop in results for Bayonetta mains, you’re probably right. But they want to use that to prove something when it proves nothing, as it would likely apply to ESAM if you banned Pikachu, apply to Ranai if you banned Villager, etc.

You’d see those players gradually improve over time as they adjusted to new mains, but it’s possible that:

  • The character they were robbed of is the one that clicks with them the most, meaning they may never achieve their prior peaks.
  • They may simply lose interest in the game if they had a particular love or dedication to the character they can no longer use.

As Bayonetta mains are players, the same circumstances uniformly apply to them. This makes the implication that their results would decline with Bayonetta an inane statement, as it isn’t evidence of them lacking in skill as players. However, this is the intended statement being made by people who use the “Carried” argument.

It is only evidence that people dedicate themselves to one character, generally speaking, and losing a character you have invested hundreds or thousands of hours into over the course of 1-3+ years is going to damage your ability to see your peak results. Especially immediately after losing that character.

Hence, as stated before, saying somebody is “Carried” is inane. It would only be applicable if Smash 4 was determined by who could play an Iron Man the best.

However, what if they had the data to back it? What if they could find an example of somebody who saw success after switching from one high tier to Bayonetta? Does this prove or provide evidence that the character will “Carry” you to the top?

We can use data to examine that idea.

2.5: National Success Rate of Bayonetta

There are 162 on-record Bayonetta mains who are power ranked. I took this chart and boiled it down to who was and wasn’t “successful”. Now, success is a relative term. If I were to go at event and go 2-2, it would be a success for me.

But this is not the case for the level we are discussing. We are comparing absolute success to players who consistently place top 8. This is an arbitrary metric that sort of just came to be, but it’s effective.

We have a few categories for our first chart:

  • Top 32 – Placed 25th or 17th
  • Top 16 – Placed 13th or 9th
  • Top 8 (Once) – Placed Top 8 once.
  • Top 8 (Consistent) – Placed Top 8 at least twice.

national1.png

  • Unsuccessful: 88.3%
  • Top 32: 4.9%
  • Top 16: 1.2%
  • Top 8 (Once): 2.5%
  • Top 8 (Consistent): 3.1%

Success Rate at all levels: 11.7% 

Partial Conclusion: Successful Bayonetta players are rare compared to her install base.

Depending on your exact definition, they may be extremely rare. Among players who have made Top 8 on multiple occasions, the percentage gap of success becomes 3.1% to 96.9% if you have a particularly stringent definition of the term. and only apply it to consistent top 8 placements.

This is in and of itself being very generous to those pessimistic about the character. This considers Abadango and Zack “Consistent” Top 8 presences which is not really true, and this is not contextualizing the status of Bayonetta mains who managed to make Top 32.

Let’s talk about those Top 32 appearances.

  • Chag (17th, GENESIS 4, GENESIS 5)
  • Sells (17th, DreamHack Atlanta)
  • Mystearica (17th, 2GGC: Fire Emblem Saga)
  • JeBB (17th, GENESIS 5)
  • ikep (17th, Umebura S.A.T.)
  • Calculus (25th, GameTyrant Expo)
  • Lagnel (25th, The Big House 7)
  • Child (17th, GENESIS 4)

Now, one of the defining factors of certain Bayonetta mains is their ability to make a significant upset on their way top that Top 32 position. If we were to boil this list down to only the players who made significant upsets to make their placements, we are left with;

  • Chag (Larry Lurr/GENESIS 4, Raito/GENESIS 5)
  • Mystearica (VoiD/Fire Emblem Saga)

Every other listed Bayonetta either made minor regional-level upsets or were heavily benefited by chaotic brackets.

In Lagnel’s case, there’s no assurance he even used Bayonetta significantly at The Big House 7 as he is also a prominent Zero Suit Samus main. Even if he did, his most significant set win was vs. Umeki.

In JeBB’s case, he made regional-level upsets of Rayquaza, Sinji, and Seth. All notable, but none really compare to more bombshell upsets made by Mystearica & Chag in order for them to place Top 32.

Similarly this is applicable to Child, who defeated Pugwest and Brood at GENESIS 4. This was a benefactor from ZeRo losing to Brood and Mew2King DQing. Not a bad performance, but not on the level of the big 2 listed.

Calculus could be considered arguable, as he defeated Captain L at GameTyrant Expo 3-0 after defeating Day. Captain L is not ranked Top 100 via OrionRank, however, he could very well be that good. Nonetheless, scraping the 70-100 position in wins still does not compare to the above 2 cases seen with Mystearica and Chag.

Sells made a number of very regional upsets at DreamHack Atlanta over Sonido, RoguePenguin, and Eldin. Lost to Salem & Ally. Not a poor performance, but no significant upsets.

Ikep has a few accomplishments under his belt, but none of them are at majors, and we are around 17-18 months separated from a 17th at S.A.T. that included no upsets. Again, no bad losses –  Mr. R and Ranai – but nothing significant beyond the raw placement.

The defining factor here is the fact that both Chag and Mystearica upset top 10 players to make top 32.

Additional adjustments:

  • Pink Fresh and Saj are placed in a Falloff category, as both have suffered a significant decline in results since 2016 and saw no extended success in 2017.
  • There is now a placement for Top 4 and Top 8 rather than two Top 8 sections to better represent peaks.

2.50: Adjusted Chart – Accounting for Falloff/”Fluff” Placements

national2
National defined as Category 4-6 events. See Section 1 for breakdown.
  • Falloff: 1.2%
  • Unsuccessful: 92%
  • Top 32: 1.2%
  • Top 16: 0.6%
  • Top 8: 2.5%
  • Top 4: 2.5%

Success rate at all levels: 8% (-3.7% decline after adjustments) 

We could adjust this further by pointing out the fact that the only Top 32 placements left in all three instances went 0-6 in games after reaching winners side Top 32. However, the difference between this chart and the prior one even after adjustments isn’t very significant.

There are some random things I never really thought about mapping out or didn’t have the time to, but these are some positive examples for the character of players going well beyond their level to make upsets:

  • Black Yoshi and AeroLink technically meet the Top 32 qualification and upset criteria by defeating Abadango and ESAM, respectively. However, neither are power ranked anymore and both peaked at 25th in their tournament performances.
  • Dom defeated Kameme at EVO, meeting the upset criteria, and he IS Power Ranked in Upstate New York. However, he placed 65th at the event.
  • Some players used in the 162 number have not attended national events, but hail from historically weak regions and are assumed to almost certainly be in the “No success” side as all of their non-Bayonetta peers PR’d higher perform very poorly, ranging from 65th to 257th in major placements. There is a potential margin of error in this area, albeit it is very unlikely.

2.51: The Great Filter

As this section draws to a close, I thought of something: How often is it that an unranked player breaches the Top 16 to begin with? What if the prominence of highly ranked players acts as a barrier that filters out?

This veers more into a hypothetical idea because I did not have the time to actually put this into raw stats, but what we’ve seen in the player data thus far is a correlation between the ability to make Top 8 if you’ve made Top 16.

Why is this?

Well, I think the reasonable explanation for a number of Bayonetta mains capping at 25th or 17th boils down to the skill pool progressively increasing as tournament brackets go on.

bracket

Let’s use GENESIS 4’s bracket as an example. The picture here is a bit small, but let’s follow the projected seeding of the event alongside their rankings at the time of the event.

GENESIS 4 Top 32 Seeding (With PGRv2 Rankings)

  • VoiD (#6) vs. Kameme (#11)
  • Ranai (#14) vs. Ally (#2)
  • ZeRo (#1) vs. Tweek (#17)
  • ANTi (#9) vs. Abadango (#7)
  • Larry Lurr (#5) vs. Captain Zack (#20)
  • Mr. R (#10) vs. Nairo (#3)
  • Salem (#13) vs. MKLeo (#7)
  • Dabuz (#4) vs. komorikiri (#12)

By the start of Top 32, your general major event is going to have the top 20-30 seeded against each other, meaning that even if you can coast by top Top 32 on a chaotic bracket, you have to generally fight a top 10, 20, or 30 player in order to progress to Winner’s Quarters.

Considering the seeding didn’t even remotely match up to what really went down at GENESIS 4, the second barrier that keeps unranked players outside of the Top 16 is the fact that if you’re approaching from either winner’s or loser’s, upsets make it likely for you to run into a top level player.

These factors across events with a large number of top players make it very hard for any prospective Bayonetta main trying to be the next big thing to succeed, and it’s no surprise that a number of 17th or 25th placements are on brackets where no top player upsets were required.

The idea is that if you can cross into Winner’s Quarters or Top 16 through loser’s side, there’s a very high likelihood you can continue to move in bracket because you’re almost required in most cases to have beaten a top player to get there.

If you’re capable of doing that, it’s likelier that you have the skill to progress even further. Yet, even that’s not an assurance, as our two remaining Top 32 examples left both did pretty poorly once they reached Top 32.

Not only is Bayonetta’s national success limited, it’s prohibitively difficult to break into national success. You must be skilled at the game to do so, as it requires at least one, often more, noteworthy upsets on top players. Fluking your way is impossible.

2.52: Conclusion on Mistake

We’ve sort of built to this point, but the conclusion to be made after dissecting the flaws in the term “carried” and by looking at national success rates:

  • Picking up Bayonetta will not ensure you success, let alone the level of success seen in Mistake.
  • Being nationally successful in general is very difficult.
  • The terminology used is already a fallacy by its very nature.

Ergo, no, Mistake is not carried. Neither are any of the other 8 players listed. Mistake is what’s referred to as an outlier.

While he did not have factors indicating he would be successful, he was anyway, but the 1% outlier doesn’t give credit to the concept that anybody can be successful with the character without adequate fundamentals. Bracket theory and her success & progression rates back that.

2.6: Power Ranking Data

One way to examine the character is to look at her presence in power rankings. This can tell us if she’s dominating the regional aspects of the game since she’d pop up an inordinate amount and frequently top power rankings.

So, let’s start by looking at who is and isn’t a Bayonetta main.

2.60: Prominence of Bayonetta

bayonettachart4
5.7% of the ranked player base mains Bayonetta.

Out of 2803 players, 163 were Bayonetta mains. Secondaries are difficult to confirm in this case, and as this is a case study on player mains, I would often search twitter bios or VODs to clarify in cases where it was unknown if Bayonetta was prominently used by a player per smash.gg, the Power Ranking Database on Smashboards, etc.

This number is actually lower than that of her results saturation in tournaments as documented in section 1, and I struggle to find this a concerning number even when accounting for a cast of 58.

Bayonettachart1
53.8% of Power Rankings have at least one Bayonetta.

This is another one of those candidates for “worst statistic” for the character. I suppose that’s subjective in a sense as long as I don’t have comparative data for other characters due to time constraints, but what little digging I did indicates characters like Cloud & Mario probably aren’t far behind in frequency.

If we’re thinking about the character as a dominant force, the most we can manage here boils down to her being absent on 111 separate power rankings.

Beyond that, you might wonder how she does on the Power Rankings where she actually exists. To do this, I split it into Top 3 and Top 1.

2.61: Dominance of Bayonetta

Bayonettachart2
19.2% of Power Rankings have at least one Bayonetta in the Top 3.

In contrast to the charts above, I’d say this is somewhat worrying, and definitely the best statistic for her in Section 2 as a whole. While a large chunk of power rankings lack Bayonetta, a chunk of the ones that do often have her in the upper echelon of that region.

Now, does this point to dominance? Well, we have one more chart to account for that idea.

Bayonettachart3
6.3% of Power Rankings have Bayonetta at #1.

This is being a bit generous and accounting for Tweek and Salem being the de-facto “0th” of the Ohio and Central Florida rankings, respectively.  That in mind, she’s only #1 on the following PRs, with two duplicates:

  • Salem, 0th (Central Florida)
  • LuxDre, 1st (East Tennessee)
  • Tweek, 0th (Ohio)
  • BlazingSkie, 1st (St. Louis)
  • Captain Zack, 1st (New Orleans)
  • Lima, 1st (Texas)
  • Lima, 1st (Dallas-Fort Worth)
  • SilentRain, 1st (Oregon)
  • Buckstrom, 1st (Alaska)
  • Rush, 1st (Hawaii)
  • Mistake, 1st (Southern Ontario)
  • Mistake, 1st (Durham, Ontario)
  • FRauDF!SH, 1st (Bahamas)
  • Clokke, 1st (Colombia)
  • Laina, 1st (Magallanes Region)

Removing the fluff, you have 13 players that are ranked #1 (or would be ranked #1) in their own region out of over 200 individual regions. I don’t want to be dismissive, but outside of a decent number of them being top 3, this seems fine for a character that’s generally agreed to be #1.

She’s clearly not dominant at a notable level when it comes to the variety of regions that exist.

2.7: Conclusion of Section 2

Data indicates the following:

  • Bayonetta’s success is limited to the top percentile of her mains.
  • 5/9 of her top level mains have factors that would indicate future success, such as Brawl Ranking status or prior success in Smash 4.
  • Bayonetta does not dominate regional Power Rankings.
  • Bayonetta is not present in nearly half of documented Power Rankings.
  • Bayonetta is a common character, but does not dominate any Power Ranking.
  • The concept of her “carrying” players is not supported by any actual data and stems from emotional arguments.

2.70: Quantum Foam Hypotheses

There is an alternative point made that Bayonetta could theoretically be dominating a level lower than that of what is documented across 240 Power Rankings, few of which have overlap.

Her significance would be invisible to feasible collection methods at this level. However, I do not find this argument to be credible.

  1. It exists on a level that cannot be verified or documented in any effective manner. The burden of proof is on the person making the claim, and the claim is often made with bias (e.g. their own local scene versus the potential hundreds or thousands that exist at this “quantum” level.)
  2. It contradicts the fact that the character sees the bulk of her success among a select few players.
  3. It contradicts the fact that even obscure regions with players who go 0-2 at nationals do not seem to be encumbered by her existence in any fashion, with most Bayonetta players ranking outside of the Top 3 and the vast majority not ranking 1st at all.

I am using PRs from extremely obscure regions, too. This is not a cherry-picked list in which I only list the best: Places ranging from the Bahamas, Wyoming, and Iceland are all listed alongside your more noteworthy regions. These are places that you would expect to have a low skill capacity. Despite this, Bayonetta does not dominate those lower skill level PRs.

I would be inclined to argue she does not dominate any PR at all. Excluding exceptional 50/100 man PRs seen in Chile & Japan, most traditional 10-15 man PRs, at the very best, have 3 active Bayonetta players. The main exception is Maine, where there are 4, but none are even top 3 PR in their region one one is listed as an Honorable Mention.

2.71: Only One Option

Bayonetta’s top level players are very clearly good at the game, and the claims stating otherwise are doing the following to come to their conclusion:

  • Ignoring PR data.
  • Ignoring her national success rate.
  • Ignoring the records of the majority of her mains and the context behind them.
  • Cherry-picking individual stocks, sets, or entire players removed from the context of Bayonetta mains as a whole.

This has impact on a pro/anti ban debate in the sense that you can now reasonably conclude that their success is fully earned and that the character isn’t so powerful that anybody can use her to instant success.

This is something most top players were aware of to begin with, of course, but it’s become clear to me over the last month that large portions of spectators have a fundamental misunderstanding of the metagame (or competition in general) or simply choose to be selectively outraged.

Defining how “Good” they are can only really be done in the gameplay itself, which this article isn’t really dedicated to. But with all the data being analyzed, you might consider that this is all reflected pretty strongly in the fact that only a select few do well.

 

SECTION 3 – MATCH-UP DATA

 

3.0: Top 10 vs. Bayonetta

For this segment, I took the 3 best players of each character in the data-defined Top 10 and made a set chart spanning from March of 2016 to Frostbite 2018 and pitted them against the 9 best Bayonetta players per Section 2.

Considering all the stuff used, I can’t/don’t know how to display this on WordPress, so I will provide a link to the Google Sheet containing the information in nifty visual form. This is taken directly from a Pikachu chart where I simply applied the method to the data indicated top ten.

LINK TO MATCH UP SHEET (IMPORTANT) 

https://docs.google.com/spreadsheets/d/17x6pri-VpxOr8Wv86KxLz85YZEKRZzJK0KJvffeEk1k/edit?usp=sharing

This only accounts for sets where the two relevant characters fought. For example, Nairo has won sets vs. Salem with Diddy, but this is not accounted for in the Nairo/Salem set record since I only used ZSS-heavy sets.

3.1: Character Breakdown

     In order of effectiveness

  • Diddy Kong (66.2% of matches won)
  • Cloud (61.2% of matches won)
  • Sheik (61.0% of matches won)
  • Rosalina & Luma (60.9% of matches won)
  • Fox (46.20% of matches won)
  • Mewtwo (45.90% of matches won)
  • Mario (45.70% of matches won)
  • Sonic (45.20% of matches won)
  • Zero Suit Samus (44.40% of matches won)

One immediate thing you might point out is the fact that, since characters are mained by players, player data itself might have a huge swing on things. You’re right, so we’ll take what we have here and extend it into the next portion of this section.

Before we do that, it’s pretty easy to take note of the fact that no character among the top 10 actually score below 44%. We aren’t making direct MU ratios for these, of course, because this doesn’t account for nuances in gameplay that actually determine how good a match-up is (or isn’t.)

3.2: Player Breakdown

     Win rate among players vs. Bayonetta

  • MKLeo (83.3%)
  • Dabuz (81.5%)
  • VoiD (77.3%)
  • ZeRo (73.3%)
  • Zinoto (69.6%)
  • komorikiri (61.5%)
  • Charliedaking (60%)
  • WaDi (60%)
  • KEN (56.3%)
  • Tweek (56%)
  • Mr. R (54.5%)
  • Light (50%)
  • ANTi (50%)
  • Nairo (48.3%)
  • Abadango (46.7%)
  • Ally (45.8%)
  • Falln (44.4%)
  • Choco (41.7%)
  • MVD (41.7%)
  • Marss (38.5%)
  • Zenyou (37.5%)
  • 6WX (36.4%)
  • Manny (25%)
  • Javi (20%)
  • Kirihara (20%)
  • Rich Brown (14.3%)

Now, if you’re on mobile or otherwise lack access to the chart or even simply missed some of the quirks here, there are important details to note that break down some aspects of this list.

3.3: Learning from the Match-Up Chart

1; KEN’s win rate is inflated by a 7-0 record on 9B. He is otherwise fairly weak against Bayonetta, sporting a 1-1 record vs Lima and one win on JK. He has 0-2 records versus Salem, Tweek, and Abadango.

2: VoiD’s record is actually really great all things considered, but it’s also inflated by a 6-0 record on JK. He does go even with Salem and has a dominant 4-0 record on Zack, something I was never aware of.

3: Marss’ record, already mediocre, is actually worse than it seems. Two of his sets on Salem were in early 2016 just after Bayonetta was released.

4: Of the many players listed, only ZeRo (11-7), Dabuz (8-2) and Zinoto (3-2) hold winning records over Salem. This is excluding Marss for the noted reasons above, although in that case he’d be with ANTi and VoiD as the only two players to tie with Salem.

5: MKLeo’s win rate is based solely on 6 sets with Abadango. Don’t put a lot of stock in it since Marth has been his go-to, alongside his seemingly retired Corrin pick mid last year.

6: Even after some slip-ups in early 2018, Dabuz is very clearly the most historically dominant anti-Bayonetta player active. Falln has a mixed record and Kirihara has the work record of any player when you account for the number of sets he’s had comparative to Rich and Javi.

7: Salem’s primary weakness among the top 10 in particular is Diddy Kong, and this doesn’t account for the two occasions Nairo defeated Salem with Diddy and extends into a set win from MVD and 3 set wins from Zinoto.

8: Salem is really good, in case that sheet doesn’t emphasize it enough. A lot of win rates among certain characters would look significantly better if Salem’s record wasn’t there. For example, removing Tweek’s losing record would put Cloud’s win rate at a staggering 80%. Salem has this effect in numerous areas.

9: Going back to KEN, if we remove his record over 9B, Sonic’s record at the top level becomes the only particularly bad one at just under a 30% win rate. Part of this is Salem being the best anti-Sonic player worldwide, but Sonic seems to pretty universally struggle based on the chart.

10: Beyond time constraints restricting my ability to gather further match-up data, it becomes increasingly difficult to get any useful information as a lot of players who use characters below this point are the sole top level user of that character. It would be hard to get a lot of Villager data, for instance, since there’s a massive gap between Ranai and whatever player might be considered the second best active Villager user.

11: Because this lacks a number of players and character users and doesn’t account for counterpicking characters, there are other nuances that have an effect on player win rates. This is why I won’t be comparing who has a solid record vs. who doesn’t, since factors not listed here help determine those things.

3.4: Conclusion of Section 3

  • This is, so far, some of the stronger evidence indicating that while Bayonetta is good, she does not really dominate most of the top cast or top players.
  • While certain players perform very poorly versus her, you have mirror image instances that demonstrate the match-up can be won, even consistently. At that point, it’s indicative of a potential player problem if they regularly lose to Bayonetta.
  • Sheik, Diddy, and Cloud continue to be very relevant characters in the discussion and clear top 4 candidates along side Bayonetta.
  • Considering Sheik’s decline and VoiD’s inflated record, you could also feasibly point to her as a character in decline, leaving Diddy Kong & Cloud to mesh with Bayonetta as the “Top 3.”
  • ZeRo’s departure could have a big effect on the metagame, as he was one of the best three anti-Bayonetta players in the world alongside Dabuz and MKLeo.

Based on all of this, I’d argue there’s a data-based case to suggest what has long been assumed – that Cloud, Sheik, and Diddy all potentially win or go even versus Bayonetta. Only time will tell as meta developments occur, but data thus far supports it.

 

SECTION 4 – META KNIGHT

 

4.0: National Ranking Comparison

Brawl and Smash 4 are very different games in both mechanics and roster balance. However, considering Smash’s unique status, there aren’t many places to go for comparison. It’s hard to find something, say “hey that’s where it becomes broken”, and make a judgement call.

Brawl is really the best we have for the sake of comparison, as it represents an extreme meta state where one character very clearly centralized the game around it. While I won’t make conclusions on what a comparison might mean yet, I think direct comparisons using the data we have should be made.

2014 SSBRANK Meta Knight users:

  • 1: Nairo
  • 3: ZeRo
  • 4: Otori
  • 5: Mew2King
  • 8: Ally
  • 11: ANTi
  • 13: Tyrant
  • 14: RAIN
  • 16: Kakera
  • 25: FOW
  • 40: Bloodcross
  • 41: Dojo
  • 42: Orion
  • 45: Seibrik
  • 46: Koolaid
  • 53: Holy Nightmare
  • 56: Toronto Joe
  • 57: Red Halberd
  • 59: Jbandrew
  • 63: Tearbear
  • 66: Shadow
  • 67: Zex
  • 71: Lee Martin
  • 82: K9sbruce
  • 91: Jtails
  • 98: Chrim Foisch
  • 99: Bjai
  • Hidden Boss: Edge

27% of Brawl’s Top 100 comprised of Meta Knight, with 80% of its Top 5 comprising of the character. I personally place Edge as a hidden boss because he was very clearly a top 100 player based on Japanese performances towards the end of the Brawl scene.

PGRv4 and OrionRank 2017 considerations: (Highest ranking used)

  • 2: Salem (PGRv4, July-December)
  • 5: Tweek (PGRv4, July-December)
  • 13: Mistake (PGRv4, July-December)
  • 13: Abadango (OrionRank 2017, January-December)
  • 16: Captain Zack (OrionRank 2017, January-December)
  • 24: Lima (PGRv4, July-December)
  • 27: 9B (OrionRank 2017, January-December)
  • 37: JK (PGRv4, July-December)
  • 40: tyroy (PGRv4, July-December)
  • 82: ikep (OrionRank 2017, January-December)
  • Hidden Boss: Hiro (OrionRank 2017, January-December)

Here, we see the nine players dissected in earlier sections, with limited noteworthiness beyond that. Ikep is 82nd but declined from his 2016 rank of 60th, while Hiro is a probable top 100 player that simply doesn’t travel enough.

Hiro, for those unaware, is a Kyushu Bayonetta that has a winning set record on Shuton. He attended NicoNico Tokaigi 2017 after winning a qualifier, but ran into Abadango whom he lost to 1-2 before getting thrown in the Dabuz meatgrinder. It’s hard to gauge his potential based on sets versus a moderately proficient anti-Bayonetta player and the best anti-bayonetta player in the world.

Anyway, you have a comparison of 27% +1 Hidden Boss and 10% +1 hidden boss with half the players prominent in both the top 5 and top 20. Due to the skill barrier that exists in the top 10, it could be even harder for a Bayonetta player to crack into the top 5 and stay there. Tweek, for example, was already in the top 5 upon using Bayonetta as a co-main.

4.1: Power Ranking Comparisons

With limited time, I chose to put my PR research efforts into Meta Knight instead of Cloud & Diddy in Smash 4. This gives us a decent point of comparison to a character people considered banworthy while solid Cloud, Sheik, and Diddy Kong data has already been pretty extensively displayed in Sections 1 and 3.

I think I made the correct decision. Brawl had a significantly smaller number of PRs, many of which are state combinations representing scene that felt more like large scale regions, but Meta Knight was prominent everywhere.

4.10: Prominence of Meta Knight

metaknightplayers
16.7% of the ranked player base mained Meta Knight.
prswithmetaknight
76.1% of Power Rankings had at least one Meta Knight present.

It’s pretty easy to see just how prominent Meta Knight was based off of this, being a persistent fixture in Brawl’s metagame and appearing on the bulk of PRs. However, this isn’t even accounting for the staggering number of secondary examples where Meta Knight was used, which would bring both numbers up even more.

4.11: Dominance of Meta Knight

metaknighttop3
42.2% of Power Rankings had at least one Meta Knight in the top 3.
metaknightatone
26.7% of Power Rankings has a Meta Knight ranked 1st.

This probably isn’t a shock to anybody, but Meta Knight was indeed an extremely prominent character in Brawl’s meta game to the point of centralizing it. Not shown are other factors that I don’t think warrant charts, but are worth discussing some.

1: Tristate in particular had 4/5 Meta Knights in its Top 5. Comparatively, a Kanto ranking –  likely Smash 4 ‘s best region in terms of depth – would have 1 at best in 9B. The Midwest could potentially have 2 between Tweek and tyroy, but Tyroy as a Top 5 Midwest player is somewhat debatable.

2: None of this account for Japan, as Japan had no Power Ranking as far as I’m aware. If they do, as I’m writing this, it’s too late to really change any of the above, just keep in mind that the above charts in fact underestimate his prominence as players like Kakera, Otori, Edge, RAIN, Aki, & more aren’t considered. This marks another extremely strong region with a large number of Meta Knights in its upper echelon, a factor not currently seen in the Japan of Smash 4.

3: I did not do data for the Ice Climbers, who appear to also be very prevalent on statewide Power Rankings towards the end of the Brawl era with consideration that 9B was tied for #1 as an Ice Climber main.

4.3: Reaching Critical Stages

You may be wondering at this point if we can determine how quickly Bayonetta could reach a level of dominance seen in Brawl. Well, we can give a rough estimate, using Brawl and Smash 4 character data compiled.

Brawl National Results 2017
Meta Knight sits at 29.5%.

We’re going to somewhat simplify this. With a roster size roughly 50% larger than Brawl’s and with Meta Knight’s peak representation at the National level reaching roughly 30%, you would need Bayonetta to reach 15% to have a rough equivalent when accounting for differing roster sizes.

  • National Progression Rate, Adjusted: 2.9% vs. 9.9% Base (Equivalent achieved by Fall of 2019 assuming no changes)
  • Regional Progression Rate, Adjusted: 1.3% vs 9.1% Base (Equivalent achieved by Late 2022/Early 2023 assuming no changes)
  • Total Progression Rate, Adjusted: 1.9% vs. 9.3% Base (Equivalent achieved by 2021 assuming no changes)

The national progression rate is by far the fastest even after using the adjusted method. However, this assumes no mitigating factors occurred to dilute Bayonetta’s results during 2016. While I adjusted my best for the sake of accuracy, the general malaise surrounding her meta state was prominent until GENESIS 4.

This also assumes that the Brawl data wasn’t in any way damaging to Meta Knight, but this includes his entire lifespan, something that may inherently include factors that understate his representation. If this is the case, the 15% threshold may actually be 16%, 17%, 18%, or higher.

It’s not impossible to know for sure but researching Brawl data in and of itself is quite difficult. This is a rough estimate that, unchanged and assuming no stagnation, Bayonetta will reach Meta knight levels of comparative dominance sometime towards the end of 2019 at the national level and significantly long after that at the regional level.

While this is a nice idea to mull over, I don’t expect progression rates to remain the same as the number of relevant Bayonetta players would need to continue to grow or accumulate increasingly better results.

Now, when we consider that PR data is regional (the slowest rate of progression) we can almost definitively conclude that Bayonetta will never reach Meta Knight levels of dominance at the regional level like he did in Brawl. The simple reason is time. While we don’t know for certain when a new Smash game will be released, your typical region would have 5 more years according to this data. I would be willing to make the assumption a Smash game will be released within 5 years.

4.4: Conclusion of Section 4

WHOBO 1, Brawl’s fifth national event, saw a Top 8 where 7/8 players mained Meta Knight. The sole exception was CO18, a Dedede main that placed 5th.

This event occurred between April 10th and 12th, 2009, a little over a year after Brawl’s release. Now, compare Bayonetta, where it exceeded dozens of national events and around 16 months before a player maining the character managed to even win a major event at all.

Needless to say, the lead up to this was not top 8s filled with Bayonetta either. Frostbite, which we will shortly discuss, may be Smash 4’s analogous event. But it’s worth pointing out that it took two years and dozens of majors for it to even occur.

The time scale is big and exists in a system far more advanced than Brawl’s metagame. Data is freely shared across a plethora of forums, social media websites, and chat websites. Multiple archives exist. Massive amounts of resources exist. The interest in Smash 4 itself far exceeds that of Brawl’s based on entrant counts.

I’d be constructing a strawman if I said people were regularly comparing Bayonetta to Meta Knight, but there aren’t many data points beyond the national progression rate that even remotely hint at the two being similar.

The common counterpoint used is to argue that Brawl shouldn’t be used as an example since its meta was unhealthy, but I think the comparisons made illustrate that the two aren’t even remotely comparable.

 

SECTION 5 – Pro & Anti Ban

 

5.0: The Case Against a Ban

I will mostly be using data gathered by the article, but there are a number of additional points that need to be made.

5.1: Contextualizing GENESIS 5 and Frostbite 2018.

The purpose of this is to discuss and break down exactly how we even got to this point to start with. It’s relevant to the anti-ban position because a lot of recency bias has been used to justify a number of dubious statements made about the scene as a whole.

5.10: Frostbite 2017 vs. Frostbite 2018

Frostbite 2018’s viewership totaled at a paltry 16k compared to Frostbite 2017’s near 36k. This has been a commonly sourced suggestion that the scene has lost interest in the game, with Bayonetta as a direct cause.

However, how were Frostbite 2017’s numbers in comparison to the rest of 2017’s viewership?

  1. EVO 2017 (348,708 total between Twitch/TV)
  2. GENESIS 4 (76,828)
  3. 2GGC: Civil War (54,118)
  4. The Big House 7 (40,000+)
  5. Frostbite 2017 (35,789)
  6. 2GG Championship (35,124)
  7. GameTyrant Expo (34,601)

These were the seven events that crossed 30,000 concurrent viewers or greater in 2017, a marked rise comparative to 2016’s 4 events that crossed the same threshold.

There were quite a number of events in the 15k-30k range, but viewership noticeably stagnates in many cases as the year goes on. While I do not have direct data for this, “Scene Fatigue” is a long-touted cause, as major events were a near constant in 2017. Too much of the same thing will cause stagnation; this is common sense.

While this is only what I can manage to gather, having top 8s comprising the same players may erode interest in the game if done too often. It also may pinpoint areas of interest to “crazier” top 8s like GameTyrant or The Big House where more exceptional events happened compared to something like MKLeo Saga which was run of the mill.

Frostbite, of course, was a very hyped event with a large number of entrants. It was the largest solo Smash 4 event – but I think this alone does not give it the groundwork an event like Civil War (whose entrants it exceeded.)

Let’s examine key factors that give events more “Hype”. People, after all, like to watch the top level of competition fight. Going back to Frostbite, what did the competition look like in comparison?

Top 20 Present for Frostbite 2017, per PGRv2

  1. ZeRo
  2. Ally
  3. Nairo
  4. Dabuz
  5. N/A
  6. VoiD
  7. Abadango
  8. MKLeo
  9. N/A
  10. N/A
  11. Kameme
  12. komorikiri
  13. Salem
  14. Ranai
  15. N/A
  16. Zinoto
  17. Tweek
  18. Mr. E
  19. KEN
  20. Captain Zack

Total: 16/20 (7/10 of top ten)

Top 20 Present for Frostbite 2018, per PGRv4

  1. N/A
  2. Salem
  3. N/A
  4. N/A
  5. Tweek
  6. Dabuz
  7. N/A
  8. N/A
  9. N/A
  10.  Mr. R
  11. N/A
  12. WaDi
  13. Mistake
  14. N/A
  15. N/A
  16. komorikiri
  17. Cosmos
  18. Captain Zack
  19. Marss
  20. N/A

Total: 10/20 (4/10 of the top ten) 

If you want to consider if unfair that I included ZeRo as #1 given his retirement, fair enough – but ESAM, the de-facto 20th in this case, DQ’d from the event due to illness. With Elegant at 10th under this new definition, nothing changes.

Consider other factors;

1: One of the most viewed VODs in Smash history is the Japan/USA Crew Battle at Frostbite 2017. No crew battle took place this year. While this doesn’t exactly link to Top 8 viewership, it is an extremely effective advertisement for the event that may generate interest.

2: Japan, an important fixture in Frostbite’s success, was near nonexistent at the event. Komorikiri is the only PGR or OR ranked Japanese player to attend, while the rest that attended could best be described as the country’s B-Team.

3: Tsu and clips regarding his impact on the scene at this event stand as some of the scene’s most famous moments in 2017. Now, not only was he not present either, but there was no effective underdog story.

4: Without a compelling narrative and over half of the upper echelon missing, the end result, regardless of entrant count, is going to be an event that struggles to garner interest.

Bayonetta is a factor for the viewership decline at the event. The 16k peak occurred long before Grand Finals, and a drop hit once Fatality was eliminated by Salem in Losers Quarters. How much did this inhibit Frostbite’s potential viewership? We don’t really know, but expecting 20+ thousand people to roll into a top 8 is unrealistic unless something really unusual is happening that might generate “hype”.

Frostbite’s numbers in 2017 were exceptional to begin with, and the event was packed to the brim with top level players in a way its successor simply was not. In the same way less people watch more regionally inclined events, you’re not going to see viewership peak at an event where most of the best are missing.

Adding a top 8 many would consider boring does not help and may explain why the event was limited to the lows of some of the weaker performing A-Tier events in 2017 despite the event’s history, but this event was never, ever going to peak like last year’s did unless something as out-of-the-ordinary as Tsu came along.

5.11: GENESIS 5 vs. GENESIS 4

GENESIS 5 hit roughly 40,000 viewers, the biggest since The Big House 7, but a decline from its predecessor GENESIS 4. However, both entrants to GENESIS 5 and total viewership for all primary Smash games saw a noticeable decline.

  • Melee viewership saw a 30-40k decline overall. This is odd, as Plup going on a huge run should be suited to make it an incredibly watched event. While it did run late, a commonly cited source of low viewership, this doesn’t really explain why both prior GENESIS events peaked in the 110k range despite also running fairly late.
  • Entrance across all game saw a decline. I measured this out in a post on reddit, but the decline ranged from 20%+ for all involved games.

The reason I bring Melee up is because if the more popular game declined for that event, you would expect the same to be the case for other games at the event.

There are a couple of possibilities.

  • Interest in the Smash scene as a whole is declining and things have simply peaked. Smash is a niche right now as all relevant titles are on dead consoles, and the game is expensive to get into unless you’re rolling with 64, which is in and of itself a niche.
  • GENESIS does not hold the hype or sway it used to for viewers or attendees, especially as the number of majors has increased in the last 1-2 years meaning you can have that major experience without travelling as far.

Considering the fact that Frostbite itself outpaced GENESIS 4’s Smash 4 attendance, it may simply have to do with the spread variance of major events across the country. People cite declining attendance largely based on GENESIS 4’s numbers, but Frostbite is very hard to ignore here as it indicates an increase at the same time the scene is supposedly declining.

I can’t make a definitive statement as to whether or not the scene is declining, but it appears that the community seems to be basing this off of only very recent data without accounting for potential nuances. I imagine we’ll have a definitive answer by EVO, but it’s a premature statement to make.

5.12: What could be the cause of scene decline?

So, let’s move forward a bit. What of the scene is actually declining? Is it the result of Bayonetta, or is it other factors that aren’t often discussed, or is it a combination of things?

Well, I’d like to consider a few possibilities:

  • Dissatisfaction with current narratives.
  • Wii U being a dead console, limiting any growth.
  • Gamecube controller expenses.
  • Waning interest in the game itself as it approaches a 3.5 year lifespan, which is fairly long in the context of most E-sports, especially ones with a cessation of support by the developer.
  • Solidifying metagame that reduces interest for people seeking variety, one aspect that makes Smash 4 appealing to many people.

If the scene was actually declining, it’d be really hard to know what the direct cause was, if any, unless viewership started to tank while Bayonetta players simultaneously dominate top 8 composition.

I posit the additional fact that Bayonetta’s presence in the Summer of Smash did not appear to damage viewership moving forward into the latter parts of the year where controversy regarding the character grew. While she wasn’t an especially prominent guest at The Big House 7, GameTyrant Expo, or the 2GG Championship, both events occurred long after EVO.

If people are suddenly turning away from the game, I can only guess that multiple issues are behind it.

5.13: Conclusions based on Frostbite/GENESIS analysis

  • There is not direct evidence of Bayonetta causing a significant or consistent decline in viewership.
  • Viewership decline in late-2017 was most likely caused by an over saturation of events, as this decline did not seem to extend to three of the most relevant events taking place in late 2017.
  • Frostbite’s decline in viewership used as an example of scene decline is ignoring a number of factors and assuming the event’s predecessor wasn’t exceptional when it objectively was exceptional going by comparative viewership statistics in 2017.

I can’t make definitive statements as to whether or not the scene is declining as I haven’t extensively reviewed factors that might hurt or help viewership throughout 2017.

However, these concerns were not commonly touted two months ago, and only seem to have spawned in direct response to a single top 8 of an event whose importance is widely overestimated due to people losing all context on its predecessor.

Viewership and factors that might help or hurt it will continue to be watched and examined. While I do not believe Frostbite is cause for significant concern, it shouldn’t be disregarded, as a potential choking of peak viewers after Fatality’s elimination could provide a case for viewership drop offs in the future.

In that case, which may not be as dramatic, you would at least have some idea of how she’s affecting viewership numbers. But, the causes for a scene decline in Smash 4 at this point are subtle enough that they need to be carefully picked apart and contextualized.

Ergo, shouting “NINTENDOOMED” is probably not the best way of going about this. If we collectively ban a single thing and tout it as the sole cause of the scene’s decline, we’re ignoring other issues that need to be addressed.

5.2: The Data Argument

Outside of a few points we will discuss in the pro-ban section, the data as far as I interpret it is overwhelmingly suggestive that the character isn’t dominant, broken, or otherwise a danger to the metagame itself.

She’s prominent, but regions are about as likely to have to deal with Sheik, Diddy, or Cloud.

Her match-up spread is good even against high-tiers, but two years worth of match-up data suggests even when accounting for inflation & outliers that there are characters that are clearly effective against her being used by multiple players.

She’s prominent among top level players, but the panicked interpretation that she carries players or can easily slide into top 8s is wholly inaccurate.

Etc, etc. You’ve presumably read the article and understand all the points that have been made up to this point.

This isn’t after minor analysis, either. This is after dozens of hours of dedicated research to find out what the answer would be. There are still various things I could mull over:

  • Hard data supporting The Great Filter Hypotheses.
  • Comparative PR data with Diddy/Cloud/Sheik.
  • Upset rates.

But bracket logic supports #1, Bayonetta’s PR data mostly isn’t concerning to start with for #2, and #3 doesn’t really matter much if Bayonetta mains aren’t breaking the aforementioned filter.

5.3: The Case in Favor of a Ban

The problem we’re at is kind of obvious if you’ve been reading the article up to this point and didn’t just skip to this particular section: There are no hard data points in favor of this position. There are a few bullet points that are worth discussing that make the character threatening, though.

  • There are speculative problems that are realistic and could be at least partially attributed to Bayonetta.
  • There are data points to indicate that she will continue to become more nationally prominent as time goes on.
  • While her PR data is mostly lax, her being in the top 3 position of nearly 20% of PRs is worrying.
  • ZeRo’s departure suggests a decline in a character that historically does strongly versus Bayonetta, which may further allow Bayonetta results to creep higher and higher.

5.30: Speculative arguments

The most prominent speculative case brought up is the prospect of viewership decline. I’ve discussed the inconsistency of those numbers recently, but if it continues to trend downwards while Bayonetta’s presence trends upwards, you will have a good statistical correlation.

I will discuss things like this further in Section 6.

5.31: Concerning data points

There are a few areas that I find concerning.

  • A 15% critical threshold isn’t that far away and could realistically result in her pure numbers resembling Meta Knight by sometime in 2019 or 2020.
  • Her progression rate does show her a step above the rest, albeit we won’t know until April how much this has progressed further past phase 6.
  • 20% of PRs having Bayonetta in the Top 3 is the only data point that suggests Bayonetta is actually prominent at the regional level. While I brush this off since most aren’t #1 (ergo, she doesn’t dominate) she’s definitely popular in some areas.

5.32: Arguments based on subjectivity

People are going to run with a mindset that often roots for those who are a part of their region, people who they are friends with, or people who are using widely loved characters . The latter case, you become a mid or low tier hero, or even just an underdog if you’re using more accepted Upper/High tier characters like Lucario or Luigi.

The dream for these people is the hope that these players actually take the event, though a scene that’s lasted over three years despite being dominated by the same player is obviously going to find solace in the fact that their favorite mid or low tier player probably won’t win much.

With that, exciting upsets or sets in general will often suffice. People like the sense of variety Smash 4 gives.

Bayonetta stands in opposition to that variety by introducing a character whose advantages aren’t as subtle and present a constant sense of tension due to her KO ladders.

I can’t make people who don’t like watching Bayonetta watch it, and it’s clear how she could drive viewership down. If people were concerned with a solidifying meta of 10-15 characters that are truly viable at the top, what happens when you take that away and introduce more of a single character that is frustrating for many to watch, especially when this frustration is echoed by many players in terms of actually fighting her?

This is one of the areas where the door is left open. If the entire spectator community collectively hates this character and drives viewership down, the scene itself is in danger even if it appears to be a highly irrational response.

In that case, a ban would be warranted, but even that has its risks and may not solve the problems permeating the scene.

5.4: Conclusion of Section 5

  • Viewership numbers are inconsistent, but 2017 saw growth in certain areas. Strongest year-wide viewership, in fact, occurred in the second half of 2017.
  • Reasons for the less watched events-  even late into the year – do not appear to be related to Bayonetta herself.
  • Frostbite has been subject to hysterical recency bias that totally disregards any and all context.
  • There are a few select concerning data points that indicate Bayonetta could reach critical levels within 1-2 years, something that has a small amount of evidence to suggest it could poison viewership.
  • Spectators are not necessarily swayed by data points if they still hate watching the character.
  • Bayonetta should be watched and her potential threat to the game shouldn’t be disregarded.

There’s no definitive answer to whether or not the character should be banned, but data strongly suggests she should not be at this point. However, this data is obviously subject to change, and it should continue to be observed.

 

SECTION 6 – Addressing the Community & Solutions

 

6.0: Calming Down

This is probably the biggest one. If I may provide an anecdote, nobody seemed to really care or be concerned about Bayonetta on reddit in January until GENESIS 5 happened. The usual “this character is stupid” posts came up, that’s fine, but people did not seriously consider a ban.

The ups and downs of how people think about this character have always been pretty reactionary, but this entire Frostbite ordeal should really serve as an example to hold your tongue and not jump to extreme conclusions.

I just can’t help but notice this fervor surrounding the character came into being in a way it never really has since her release, even though we saw a wide under performance at PAX South. Metagames don’t typically undergo apocalyptic shifts in two months without any developer input.

6.1: Monthly Polling

One idea is to come up with a good polling formula and distribute across Reddit and Twitter as a way of gauging public opinion on Bayonetta as a character. Google forms seems to be the best way to do this to hopefully avoid tampering.

My suggestion would for it to be on the 1st of every month. Just use it as an index to survey community satisfaction or lack there of, particularly with Bayonetta. Perhaps do polling after a major on a Monday and then repeat a Monday where one didn’t occur to see how much of it is reactionary.

Freeziebeatz actually held a poll on reddit like this, and while I don’t have the data at the moment as I forgot to get it, I think his idea is one that should be made standard as long as the community expresses concern about this character.

6.2: Playing the Waiting Game

As is plainly apparent, there isn’t really a data based case to suggest the character should immediately be banned. Can this change? Yes. If every top 8 becomes Frostbite, we will have seen a spontaneous meta shift that greatly favors accelerated growth for an already divisive character.

The problem is, of course, that the one time we got an absurdly Bayo-heavy top 8 was the time that half the top 10 were missing, affecting things like seeding among a lot of other things (including viewer interest.)

6.3: The EVO Prophecy

Bayonetta will do well at EVO. I want to point this out in advance, because it wasn’t a surprise when 7 were in the Top 32 last year and it won’t be a surprise when it happens again. But, you may ask, why?

It’s pretty simple, of course. This is an extremely volatile game. Bayonetta herself makes for good upset material, but even more ominously, the specter of constant upsets puts very good players into losers who might otherwise beat a Bayonetta. Remember MKLeo getting 65th? That could happen again, and he lost to false and Mute Ace.

Unless the meta has already shifted in a hyper-favorable position for Bayonetta by August, I would appreciate if people didn’t react to Bayonetta doing well at an upset flooded best 2 out of 3 event with shock, dismay, and apocalyptic predictions. It would be more surprising if she performed poorly.

6.4: Improvement from the Playerbase

I absolutely hate to roll down the classic Dark Souls fandom avenue, but I want you to take consideration of the fact that beyond the data I presented, there are a large number of Bayonetta mains (including many who are PR’d) that regularly drown in pools.

If your frustration comes from fighting her, let Section 2’s data be a lesson: At a certain point, it truly is about player improvement. Most Bayonetta mains are not Mistake or Salem, and most Bayonetta mains do not perform well. If it doesn’t take a Top 50 player to consistently keep out a massive number of Bayonetta players, it’s likely the community itself can manage if they seek the proper resources.

This doesn’t solve the problem with spectators being bored, but it’s a general suggestion for players. If she isn’t dominating regions at a varying number of skill levels, the best solution is to improve your fundamentals. Not just SDI – everything. Roll habits, unsafe attacks on shield, doing the same thing too much, etc.

While a ban could be justified if it’s necessary to salvage viewership, a ban is not justified on the basis that a portion of the lower skilled player base is unwilling to improve to the same level as Bayonetta mains that go 2-2 in pools.

6.5: Continued Observation

While I don’t intend on making another article this size anytime soon, I think we should keep tabs on Bayonetta as a character. Through all the data I’ve presented today, it could change at the drop of a hat in unexpected ways depending on new meta shifts.

We should pay attention and I don’t want to see this article linked in 8 months as an anti-ban argument if Bayonetta actually does begin to swamp the metagame. This is a post that covers 2016 and 2017, and I will continue to update my blog with (among other things) progression in Bayo’s data. I will report on new trends.

6.6: Conclusion

Thank you for reading this monstrously large article. I only finished it completely the morning of me publishing it. It was a difficult and time consuming project to manage, and something I’d still like to add on to in the future with some idea and concepts that didn’t quite make it in.

I hope this article will provide some perspective on the community as a whole and help shift the debate in some sort of directing so the community can stop spinning its wheels in twitter drama and emotional arguments on Reddit. I think we can be better than that. If I didn’t, I wouldn’t be trying to change the landscape of how this is all being discussed.

I know some will disregard this article and potentially the data contained within, and I accept that. I accept if people remain pro-ban despite what’s been shared. I fully grasp that people who hate watching the character may be unable to be swayed, but I hope I at least managed to ward of some of the weaker arguments & points surrounding the character.

Coming soon is an analysis of Sheik’s kill potential using data compiled from top Sheiks. When is it coming? No idea! I’m taking a week off to watch the Olympics with my dad. I will answer any questions regarding this article in the meantime.

SECTION 7 – Sources & Methodologies

7.0: Section 1

Tournament Database:

Database Scorekeeping Methodology:

7.1: Section 2

OrionRank quick reference:

7.2: Section 3

Match-Up chart link:

7.3: Section 4

Brawl Power Rankings:

7.4: Section 5

Viewership Statistics:

MORE SOURCES COMING SOON! Compiling what I have into public files will take a bit.

 

 

 

 

 

 

 

 

Advertisements

4 thoughts on “DATA – Bayonetta – A detailed statistical breakdown of Smash 4’s most controversial character.

  1. Ban Bayonetta February 19, 2018 / 7:56 pm

    TL;DR

    Like

  2. Xuramaz February 20, 2018 / 1:40 am

    Thank you for all of your work! This was a very interesting read, and seeing how it has become popular on reddit, I hope that this will calm the masses, and return my twitter feed to something that doesn’t require the muting of 10 words just to avoid this mess.

    Seriously, I will restate that I cannot thank you for your work enough!

    Also, on calling 15% a critical meta centralisation for Bayo, just because that would give her relative dominance to her competitors vs. Brawl Meta Knight, I wouldn’t say that is the case. Although with a much bigger cast Bayo doesn’t need to dominate as much as Meta Knight to have the same amount of centralisation over the game, that is an issue that doesn’t scale linearly for sure. In Brawl, if you couldn’t fight Meta Knight, you had no chance. In Melee, you got a Fox problem? Good luck. However, It can be guaranteed having trouble against 15% of players is better than having trouble against 30% of players.

    Like

  3. Sage February 20, 2018 / 2:33 pm

    Thank you for doing this. It’s a great service. Hoping vocal top players like @Anti who are wholly pro-ban despite claiming they’re neutral read this.

    Like

  4. Pablo February 20, 2018 / 10:19 pm

    Geez, what an incredible in depth aritcle. Cheers mate. You will do great at Wall Street !

    Like

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s