Character Rankings or Character Success in Tournaments

The Version B “Entry” and “Semifinal” Rankings are based on 1072 tournaments, while the “Finals” and “Winners” Rankings are based on 2167 tournaments. The Version C “Entry” and “Semifinal” Rankings are based on 1647 tournaments in Japan, while the “Finals” and “Winners” Rankings are based are 2243 tournaments.

Note: You may have to use the horizontal scroll bar to see both charts



Character Rankings Summary – The colors are arbitrary rankings of a character's success rate in tournaments as either average (yellow), high (dark red), or low (dark blue). The numbers are rankings of a character from a scale of 1 (best) to 17 (worst). Note that the number ranks of 1-17 does not necessarily mean that a character rank 17 is severely disadvantaged compare to a character rank 1.

How do you read the chart? - As an example, I will use Version C Shun: Shun is not a very popular character choice by players entering tournaments, ranking 15th overall. However, the general population of Shun players have above average success in advancing to the semifinals, with a rank score of 2.5. The Shun semifinalists, have a similar rank score in advancing to the finals—3—yet it is only a slightly above average success. Lastly, the population of Shun finalists have average success in winning tournaments with a ranking score of 10.

How were the charts made? - The information above is based on calculations (change in % character distribution and % character advancement) I used on raw tournament results at www.virtuafighter.jp website. The calculations relate to the change in character representation throughout a tournament: 1) at the beginning of tournaments (entry), 2) at the semifinals, 3) at the finals, and 4) in conclusion of tournaments (winners).



The Finals and Winners columns are based on VF5R version A tournaments sponsored by Sega in Japan (2355 total tournaments with 48611 player entries from 7-25-08 to 5-20-09). The Entry and Semifinals columns are based on 1348 total tournaments with 28825 player entries from 7-25-08 to 1-7-09).

nice... very nice.

edit: shun dropped like a rock. heh.

This awesome info akai!! Lots of thanks! Not only is the ranking numbers at each stage and colors helpful, the change in colors across stages also say a lot.

It is interesting to note how lots of Akiras and Laus enter tournaments in Version C but few advance far. Ditto for Pai in Version B.

very, very cool on so many levels. . . . also, honestly didn't realize kage was like that.

Thanks, I plan to update the version C charts on a monthly basis. I think the character rankings in "finals" and "winners" will change quite a bit (except maybe Kage and Eileen), since it is only based on two weeks worth of tournaments.

Great chart Akai, and it's especially easy to read with the color scheme. It goes to show how the revision C changes have a big impact in the beginning as players are learning to adapt to these changes.

i really had no clue there were THAT many kage players in japan.

There are ALOT of kages here. At least from my experience, when people sat down to challenge me and their card was being read, I was sitting on the other side thinking, watch, another kage. 40% of the time I was right. :P

I wonder why Goh is so unpopular in Japan. I mean, there is like a million Goh avatars on Shoryuken.

Where is the love for the bastard child of Akira and Jeffry? (hell, Jeffry is 9th in ver. C?!)

After discussions with Japanese friends, I think Goh is unpopular for two reasons:

1) When he was first introduced in Evo, he was pretty weak and he relied heavily on sabakis and inashis which made him difficult to use.

2) His appearance is too dark and personality-heavy for lots of Japanese VF players. According to my friends, VF players like "thin" or "light" personality characters. Almost all of the VF characters are generic and as far as I know, Japanese VF players like it that way. If they wanted more personality, they would play Tekken.

1) I guess that would explain why Brad is also less popular.

2) Understandable, but from my POV: why play pedo-bait/kawai~i shit when you can go for the Ghold?

I wonder why Vanessa/Shun fall under the same category as well (okay, I can take a guess, but...)

It's interesting isn't it how a character like Lion can be popular in Japan whereas a character like Goh is shunned, and then to see the complete opposite in the West.

But that does show how difficult it is to develop for two different markets with different tastes.

i understand the whole cultural difference thing, but i never really thought about personality when choosing a character for any game. hell, this is coming from a person who used to play tekken competitively and exclusively.

Well, another way you can look at who is the most popular characters are
1) original VF 1 crew
2) difficulty of using the character

Well, I guess they likely mean looks and personality as seen through the characters moveset, win quotes, etc.

I'm guessing you probably don't really consider that because you play competitively. If you play mostly for fun, then you're going to want a character that you like, I would think.

Personally, I just like judo.

With Lion, I honestly also think that one of the main reasons he's unpopular in the west is because of his shitty voice. Seriously, he sounds annoying. In Japan, if they had an annoying sounding Japanese language character, I bet that character would be unpopular too. As it is, I'm guessing that they just don't understand (viscerally that is) how goddamn annoying Lion is.

hey now... take it easy on lion. he's twice the man he used to be.

Below is character rankings based on 2000 tournaments with a total of 66595 player entries during November 2, 2006 to July 23, 2007 (Pre-Version C). Due to less information available per tournament, only finals and winners categories can be calculated.



Comparing this chart versus the chart from the first post, the two charts are relatively similar in ranking characters. Does this mean that one-month-worth of tournament data is an accurate ranking of characters success in tournaments?

Below is a line graph looking at Goh's, Kage's, and Wolf's monthly "Winners" ranks.



For most of the months, the character rankings of Goh, Kage, and Wolf remain relatively at the same position to each other. However, character rankings in January is pretty skewed. Thus, one month worth of data may not be an entirely accurate measurement.

I am unsure how the "Tougeki Tier List" was derived, but for fun, comparing it to my character tournament "winners" ranking data, the Tougeki Tier List overrated Lei and underestimated Aoi.

As much as I try to make a fair, unbiased character rankings chart, there are always variables that can skew the data such as 1) differences between each tournament, 2) individual player skill level, 3) same character matches, and 4) possibly biased assumptions and errors on my handling of the data. So take the information with some reservation.

As for the "Above Average" and "Below Average" ranks, they were determined for the most part by the standard deviation between all the characters. For example, the above and below average rankings listed in this post is approximately plus or minus 5-10% from the average (50% winning percentage)

Next week or two, I should have the first full month of Version C Rankings up.

Great job with these stats Akai. Really interesting data. I always wonder if AM2 does something similar to this.

damn akai...

Thanks Akai, this is some really interesting stuff... I like the hard numbers approach vs. the intuitive approach most have used before.

Seems to me like a lot of the Pai players were hacks... at least the Kage players can win when it counts.

Looking forward to further updates.

So the moral of the story is "Kage owns to high heaven."

at least 10 feet high...

Looks like little has changed since VF4FT, game balance wise that is.

No dude. The balance in FT is way better imo.

I'd rather against play FT Lei-Fei as well.

lol

Awesome, interesting info as always akai. A few questions I always wanted to ask:

- What do the numbers listed in the table actually mean? How were they normalized -- can you also publish the raw data?

- By standard deviation to determine above/below average rankings -- does this mean that if a character's win rate is say, 70%, he or she is above average, and if 30% it is below average assuming 50% is the average? Or something more complicated?

- Is the data singles tournaments only? Or is team tournament data also included?

I do however think your first set of data is more informative, in that it shows a progression of how a character performs. For example, in the complete chart you can see that Eileen's usage is actually only average but that she performs extremely well in the finals/winners stage. This suggests to me that she is probably an above average character, as opposed to say Jacky who a lot of people use but not as many win with.

Ice-9 - This is going to be a long post. Hope it answers your questions!

Edit: The more I look at this post, it could be pretty confusing...


Single tournaments only.



The numbers in the ranking charts are based on two calculations: % character distribution and % character advancement. For each calculation, whoever have the best score is given the rank of 1 and the character with the worst score is given the rank of 17. The ranking charts numbers averaged the ranks from the two calculations. For example, Kage rank 1st in % character distribution score and rank 3rd in % character advancement for an average ranking score of 2nd.

The Calculations
The two type of calculations differs on what is being compared. The purpose of using two types of calculations to rank characters is that each calculation may have certain “biases� to characters that are either highly represented or lowly represented in tournaments. Thus, combining two different types of calculations can alleviate some of that bias.

% Character Distribution – If all characters are balanced having no advantage or disadvantage against other characters, you would expect that at every portion of the tournament the % character distribution will remain the same. Hypothetically, if character A is the choice of 5.9% of the total tournament entries, then it should make up 5.9% of the total semifinalists, total finalists, and total tournament winners. However, if certain characters have more or less success in advancing to semifinals, finals, and winning the tournament there will be a change in % character distribution during the later portions of the tournaments.

% Character Advancement – If all characters are balanced having no advantage or disadvantage against other characters, you would expect that each character will have the same chance as other characters in advancing to the next portion of the tournament. Hypothetically, a character has ~12.5% chance of advancing to the semifinals, then 50% chance to advance to the finals, and then 50% chance to win the tournament. However, if certain characters have more or less success in advancing to semifinals, finals, and winning the tournament compare to to other characters, their % chance to advance in the tournament may increase or decrease.

Below is both the raw data I gathered from http://www.virtuafighter.jp/eve_all.html and the calculations I used to make the Version B Ranking Chart in the original post (June 11th, 2007 to July 23rd, 2007). Note that it is not very easy on the eye.







For the most part, yup. If you look at the calculated data chart above, the last row is the standard deviation (SD) value for each column.

Approximately:
High/Low rankings is values > 2 SD
Above Average/Below Average rankings is values > 1 SD
Slightly Above Average/Slightly Below Average rankings is values > 0.5 SD
Average rankings is values < 0.5 SD



I made calculations at various parts of a tournament assuming that at the beginning of a tournament, the players’ skill can range from novice to advanced. The later portions of the tournament should have less novice players (eliminated from the tournament) and more advanced players; thus, the calculated data of the later portions of the tournament may reveal a character’s ability more than data from earlier portions of the tournament.

In hindsight, due to tournament format, rankings for later portions of tournaments are not as statistically sound compare to calculations for earlier portions of tournaments.

Only way to make it more statistically sound, is to increase the data set which is the reason I did the second chart for version B (November 2006 to July 2007). Unfortunately, the raw data is most complete for the last two months of tournaments. The website removes the tournament brackets for the older tournaments and list only the top 4 finishers in each tournament. Thus, I am unable to calculate "usage" and "semifinals" ranks. However, the "finals" and "winners" rankings are as informative as the first version B chart.

Edit: Just to clarify on what each column in the ranking charts represent.

Usage - Ranks which characters were chosen the most or used the most in tournaments.

Semifinals - Of all the characters chosen, what characters used by players have more success in advancing to the semifinals.

Finals - Out of all the semifinalists, what characters used by players have more success in advancing to the finals. Basically this column does not take into consideration the "Entry (Usage)" data and only considers the "Semifinals" data to determine ranks.

Winners - Out of all the finalists, what characters used by players have more success in winning the tournament. This column does not take into consideration the "Entry (Usage)" data and "Semifinals" data. This column only looks at the "Finals" data to determine ranks.

I have sooooo got to play that game!! im dyin to play it!!!

Character Rankings Version C – August

For comparison, Version B ranking chart from original post is next to the updated August chart.



The most dramatic changes in a character’s rank is listed below -

Semifinals Rankings -
Wolf (-6 ranks)
Pai (-7.5 ranks)

Final Rankings –
Jeffry (+13 ranks)
Brad (+9.25 ranks)
Jacky (+6.25 ranks)
Akira (-6.5 ranks)
Pai (-7.5 ranks)
Lei (-8 ranks)

Winner Rankings -
Wolf (+10.75 ranks)
Jacky (+6 ranks)
Lei (-5.75 ranks)
Lion (-6.25 ranks)
Goh (-10.25 ranks)
Shun (-12.25. ranks)

Whether the difference in the two charts are due to month-to-month variance or version B-version C is debatable. Next update will be October.

Nice, thanx!

Will it be possible to ask you to steadily release not only ranking charts, but raw data too? I think they are interesting too

Wow, akai... you rock o_O. Thanks for all the info!

But what the HECK happened to Goh in ver.C?? I've read he was barely even touched by change... Maybe just a bad month for Goh players...

Unicorn, sure -



Jigohro - I think its mostly a bad month. Characters that are played less will move up and down the rankings a lot easier than characters that are played more.

akai, thanks for the answers. One more quick clarification question: under the "semis" column, do the numbers stand for the number of players that made it to the semis, or the number of players who made it but then lost at the semis? I.e. if I want to know the total number of Akira players who made it to the semis and beyond, is it 69 or is it 69 + 30 + 30 = 129? I'm guessing the latter based on the way data is presented in virtuafighter.co.jp

I needed a break from work and decided to recut the data. See below for my results, Version B only:

In terms of winning percentage (for 1st place) relative to the number of entries:
Kage 6.5%
Shun 5.9%
Eileen 4.8%
Akira 3.6%
Lau 3.0%
Lei 2.9%
Brad 2.6%
Jacky 2.5%
Pai 2.5%
Vanessa 2.4%
Goh 2.3%
Aoi 2.1%
Blaze 2.1%
Lion 2.0%
Wolf 1.9%
Sarah 1.4%
Jeffry 0.7%

In terms of winning percentage at the championship game (to determine 1st from 2nd place):
Shun 63%
Eileen 62%
Kage 56%
Vanessa 54%
Brad 52%
Blaze 50%
Goh 50%
Akira 49%
Lion 48%
Lei 48%
Jacky 47%
Aoi 46%
Lau 45%
Pai 43%
Wolf 38%
Sarah 32%
Jeffry 29%

Which is more indicative of character strength? Both lists are relevant.

The winning percentage relative to the number of entries is a good stat for forming a tier list for tournament level players. But is tournament level necessarily elite level? I think the two are highly correlated, but there's no question there are lots of players who enter tournaments with no chance of winning.

The winning percentage for the championship game on the other hand is more relevant for elite players, because presumably whoever can get to the championship game of a tournament is an elite level player. I also think the sample set is large enough for the results to be meaningful.

How do the above compare to the Tougeki tier list? Here they are again as a reminder:
1st - Eileen, Kage, Lei, Pai, Shun
2nd - Brad, Goh, Jacky, Lau, Vanessa, Wolf
3rd - Akira, Blaze, Lion, Jeffry
4th - Aoi, Sarah

Tier list as suggested by winning percentage relative to the number of entries:
1st - Kage, Shun, Eileen, Akira
2nd - Lau, Lei, Brad, Jacky, Pai, Vanessa, Goh
3rd - Aoi, Blaze, Lion, Wolf
4th - Sarah, Jeffry

Tier list as suggested by winning percentage for the championship game:
1st - Shun, Eileen
2nd - Kage, Vanessa, Brad
3rd - Blaze, Goh, Akira, Lion, Lei, Jacky, Aoi, Lau, Pai
4th - Wolf, Sarah, Jeffry

What stands out? Sarah and Jeffry are consistently rated low. Aoi and Akira may be better than we thought, while Pai and Wolf may not be as stong as advertised. Shun, Eileen, and Kage are definitely top tier.

Ice-9 - In the raw data charts, the semifinalists are the players who made it to the semifinals but lost at the semifinals.



I agreed with the last sentence in your quote. Beginning of tournaments--the skill level between players will vary considerably. The further into tournament, the skill level between players will be less.

I just made a Ver B (PS3 version) statistic of how succesfull every character wass in reaching top4 players on tournaments in compare to how many characters start at that tournaments. I think top 4 of avarenge 32 players at tourney is good point where luck and so playes not-so-big role

Statistic are: Name / starting / top4 / % succes
Kage / 1516 / 305 / 20,12
Shun / 578 / 89 / 15,40
Akira / 1121 / 158 / 14,09
Jacky / 1043 / 147 / 14,09
Lau / 1055 / 140 / 13,27
Lei / 823 / 104 / 12,64
Eileen / 880 / 105 / 11,93
Pai / 1268 / 147 / 11,59
Wolf / 889 / 102 / 11,47
Goh / 439 / 47 / 10,71
Aoi / 561 / 59 / 10,52
Blaze / 771 / 80 / 10,38
Lion / 657 / 67 / 10,20
Brad / 532 / 52 / 9,77
Sarah / 724 / 68 / 9,39
Vanessa / 587 / 52 / 8,86
Jeffry / 591 / 42 / 7,11

omg Jeffry just sucks

This is based on the Version B finalist (1st and 2nd place) match-ups that I used to make the Tournament Ranking Charts.



How to read the charts:
Looking at the Character's "Rows," you are looking at his/her wins or win ratio, respectively. Looking at a Character's "Column", you are looking at his/her losses or loss ratio, respectively.

For example, let's say we are interested on how Wolf players fair against Kage players for 1st place. Going down the rows in the first chart, I find Wolf's Row. Next, I looked to the right and find Kage's Column. One can see from the Wolf's Row / Kage's Column entry that Wolf have won 6 times against Kage. Looking at the second chart, those 6 victories give Wolf a measly 0.16 win ratio (16%) against Kage. If you want to see how much losses that actually entails, you can go to the first chart and look for Wolf's Column / Kage's Row entry (Wolf loses to Kage 29 times).

Note: The color "rankings" is somewhat deceiving due to low amount of certain character specific matches. So for the most part, I would just ignore the colors. Enjoy!

The most dramatic changes in a character's rank (September from August) is listed below -

Semifinals Rankings -
Blaze (+6 ranks)
Eileen (-5 ranks)
Wolf (+8 ranks)

Final Rankings –
Aoi (-7.5 ranks)
Blaze (+7 ranks)
Goh (-9 ranks)
Jacky (+5.25 ranks)
Sarah (+7 ranks)
Vanessa (-7.25 ranks)

Winner Rankings -
Aoi (+5 ranks)
Brad (-10.75 ranks)
Lau (+6.75 ranks)
Lion (+5 ranks)
Sarah (+8.75 ranks)
Shun (+5.25 ranks)
Vanessa (-5.25 ranks)
Wolf (-6 ranks)

WOOT Ver C Sarah!!!
=)

I wish these jap eileen players would post more youtube vids so I can learn from them !!! I'll just have to go to more tournies

You're going to the Univ of Maryland on 11/17, right?

I'm trying to go to the Philly one on nec 8th. Seeing if I can convince my sparring buddy to come. I'm kinda broke so I can't afford to fly to maryland *tear* I really need a job. Do a lot of good eileen's go to the univ. of maryland one?

no idea how many Eileens will be at Univ of Maryland, but it's shaping up to be the biggest east coast VF5 tourney so far...

When exactly is nec 8th anyway? I don't see it on my calender

supposed to be 12/1-12/2 here's the thread. I've been trying to get out to some tourny to see how good I really am ya know but money really gets in the way of you know traveling lol.
http://virtuafighter.com/forums/ubbthreads.php?ubb=showflat&Number=171171#Post171171

Magnificent data! thank you very much!

Most dramatic changes in character rankings (comparing October to September)...

Semifinals:
Aoi (+7.5 ranks)
Blaze (-9.5 ranks)

Finals:
Akira (+5.5 ranks)
Blaze (-6 ranks)
Goh (+6 ranks)
Jacky (-9.5 ranks)
Vanessa (+6.25 ranks)

Winners:
Akira (-9.75 ranks)
Aoi (-11.75 ranks)
Brad (+10.75 ranks)
Eileen (-8 ranks)
Goh (+10.25 ranks)
Jacky (-5.75 ranks)
Jeffry (+9.75 ranks)
Lei (+10.5 ranks)
Pai (+7.5 ranks)

sorry for my noobness but what does that mean for kage!?

The charts rank the usage and success of the character in singles tournaments in Japan. The raw data is from virtuafighter.jp website.

Basically, the chart suggests that -
1)Kage is a very popular character used in tournaments.
2)General population, semifinalist population, and finalist population of Kage players have high or above average success in tournament advancement.

The original post (and subsequent posts) explains how the charts should be read.

~ 4 months of tournaments, the rankings in "Total" Chart is becoming more stable. I do not think their will be any major fluctuations.

General trends that I noticed -

1) In the next "Total" Ranking Chart, the Goh's (an increase)and Eileen's (a decrease) ranking may change.

2) Comparing preliminary version B and version C data, it appears that version C does not necessarily make VF5 "more balanced" than version B. I will post up more when I have more solid data on this observation.

It's interesting to note that based on your first set of data, Wolf went up and Eileen went down just as predicted by Arcadia (forgot which issue).

Surprised to see how high lau is in the list. Since ver. C he's been really pulling his weight

1,323 tournaments in 4 months time, Japan is like, the promised land. Thanks again for posting these up man.

What's crazy is, the popularity of the characters from the last Version B polls and the latest C polls don't vary much. I think 3 is the highest change I counter, being with Akira going from 5 to 2. The rest were only 1 or 2 spots if any change at all. To me that's amazing because it really shows tiers and tourney results arn't swaying player choices much. That's pretty rare in fighting games.

Also, Shun's difference between Finals and Winners is pretty crazy. Wonder why he's having such a hard time in the finals versus others. Maybe Kage and Lau are bad matchups for him lol.

It's weird how one of the most annoying characters on live, Pai is so low in the finals ranking. Maybe it's just me?

His matches against Kage and Lau were not too bad. His record against Akira and Jacky are worse. Sample of Shun's version C character-specific matches in tournament finals -

versus Lau - 43% (10-13)
versus Kage - 41% (9-13)
versus Akira - 27% (4-11)
versus Jacky - 20% (3-9)

Oh wow, yeah I could see that. Just guessed Lau and Kage due to their frequency in the finals, but those two make more sense to me as far as troubling matchups for him.

Seems like Kage is by far the strongest character... One of the other "S-Tier" characters, Vanessa, seems to be relatively weak. Is that correct?

Kage is a ninja, everyone else is not.

Not exactly. You can indirectly infer that the characters are weaker or stronger, but the most appropriate interpretation of the charts are that Kage players have more success in single tournaments than Vanessa players based on the data that was available.

As for Vanessa being a "S-Tier" character, it depends on which source you are talking about. I think both Arcadia and Tougeki Damashi (I think they are two different magazines?) released a tier list where one listed Vanessa as S-tier (based on Jide's post) while the other listed her as A-tier (Markman's scan), respectively.

I think it is best to just take the Japanese tier lists and these ranking charts with some grain of salt. There are advantages, disadvantages, and biases on how each list are made up.

Most tournaments in a month, but least amount of players per tournament (about 1000 less player entries)

Poor Goh..

Glad to see more Sarah usage though.

Aoi and Pai. Interesting.

Poor Jeff.

My girl had a good month! =)

Thanks for the continued stats Akai. These posts are really interesting.

Note: You may have to use the horizontal scroll bar to see all the images -



The Version B “Entry” and “Semifinal” Rankings are based on 1072 tournaments, while the “Finals” and “Winners” Rankings are based on 2167 tournaments. The Version C “Entry” and “Semifinal” Rankings are based on 1647 tournaments in Japan, while the “Finals” and “Winners” Rankings are based are 2243 tournaments.

Concluding Remarks –
Well, the above charts and analyzing the raw data was a 6 month long side project I started due to (cough) VFDC being down for approximately 2 weeks during later Summer 2007 (cough). The charts were done with the intention to be non-biased and based on the methods describe in the original and subsequent posts in this topic (see the updated original post for explanation on how to read the charts). In this final update post, I will give my biased interpretation of the data (mainly because I been looking at the data for over 6 months -_-;;).

The character ranking charts have some uncontrolled variables, but correlate well with the “buffs” and “nerfs” of characters from version B to C. For example, Pai and Shun version C changes, for the most part, were considered “nerfs.” In the character ranking charts, both Pai and Shun success in tournaments were less in version C compared to version B. Characters such as Jeffry and Sarah were considered to receive mostly “buffs” in version C. Likewise, they had more success in version C tournaments than in version B. Based on this, I assume that the character ranking charts, even with some uncontrolled variables, is a fairly good way of measuring character strengths.

On Tier Lists –
To make it clear, I do believe there are definite character tiers in the game; but I do not think the published tier lists are 100% accurate. They are based on the opinion of high level players and not based entirely on facts (though if you are going to make up a tier list, opinion of respected high level players carry a lot of weight). So one of the original purpose of my project was to see if there is a correlation between the tier lists and the success of players as measured by their character use/success in arcade tournaments in Japan.

The character ranking charts, depending on what part of the data (semifinals, finals, or winners) you look at, correlate well with the tier lists with a few major exceptions—such as Lei and Pai. The main thing I took from comparing the published tier lists and the tournament character ranking charts is that any single “List” alone should be taken with a grain of salt. However, when you have multiple “Lists” based on different criteria producing similar rankings, aka Kage, then it strengthens the listed tiers of characters.

Is Version C “more balanced” than Version B? –
A general assumption is that Version C is a more balanced game than Version B. When I think of a game being more balanced , I assume “tiers” are reduced or less noticeable. However, depending on how you want to look at the data on character success in tournaments, it is actually highly debatable. One can even interpret it as being worse.

Below are line charts plotting out the two calculations (change in character advancement and net change in character distribution) of version B and version C “winners” data. What you should look and compare between graphs is the spread and clustering of the data points.



Well, if you actually read this whole post or been following the thread since it started--thanks for reading and hope there was something of interest to you! Time to start on another side project until VF5:R comes out....

thank s alot for all the work you put into this man really good stuff.

Akai, you are my data god. So unbias, so thurough haha. Thank you for all of this man. It truly is interesting to breakdown some of it and give it good thought.

akai: good work my man! Are these the final stats that you're going to analyze? Would love to get the raw data...

Any major tournament announced? March is usually time for either a Kakutou Shinseiki or SBO...

It's all great stuff, but what happened to Wolf's advancement?

Make that vanessa and wolf.

I assume there was some error selecting the last two rows of the excel chart -- i just can't believe nobody using those two characters qualified for a tournament!

Note: You may need to use the horizontal scroll bar to see all the images.













ice-9 - here is everything...I think. Yeah, I am not taking any more time to collect raw data...maybe when a new revision with game play changes come around I will start doing it again (aka VF5:R). Note that with the raw data, there were two overlapping data sets (italicized versus non-italicized) used to determine the charts. Basically, the "italicized" set was used to determine entry and semifinal rankings while the non-italicized for the finals and winner rankings.

franz - there is a tour going on for the month of March (topic in this arcade forums), but that is the only thing I am aware of.

ash - Wolf got a slightly improved success (from below average to average in "finals" column and below average to slightly below average in "winners."

Thanks again for those who have been browsing this topic Any questions, comments, or criticism about the data is welcome.

OMG Kage owns
Wow.
There should be no doubt, ever, that Kage is the son of Dural.

I like how version C Jeffry moved up though
Nice job, Akai, thanks!

from looking at the raw data it's clear why kage wins so much. It's just because there is an overwhelmingly large amount of them compared to everyone else. Kage is def. a outlier who messes up the data a lil bit just because there are SOOO many of them comparatively.

Damn. That's a lot of Kages. Also, thanks akai for the Wolf confirmation.

man aoi does not have alot wins does she

Poor Goh. Below 2 standard deviations for semifinal winners and final winners... that's bottom 2.5%, pretty much nobody wins with Goh

Makes me sad.

So much for balance, stupid god damn ninja...

what the h, man?!? What's up with the Goh haters?

If you look at entry level raw data, both Pai and Kage players start off at a similar high number. However, while one has great success, the other has horrible success. The ranking charts took the raw data and calculated "success rate" rather than "who is played the most" or "who wins the most."

If you subtract Kage from the calculations for the ranking charts (the way I collected the data, only Winners Rank can be figured), the placement of the remaining characters remained nearly the same - Lau, Eileen, and Sarah players at the top; Lion, Goh, and Pai players at the bottom. The "colors" are different since standard deviation is altered due to subtracting Kage players from the calculations. Also for the sake of comparison, the highest rank minus Kage, starts at 2.

ah okey, there goes my hole. Glad you explained it out like that with pai since she isn't doing quite as well. I was hoping pure randomness of percentages was the explanation. glad/sad to hear i'm wrong

Could someone help explain to me how to read the matchup charts? I think I know, but I'm kind of uncertain. Really, really interested in those the most I think.

I knew Goh vs Eileen was a shitty match-up!!!

Correct me if I'm wrong but even on tournament level it looks like Goh only beats Eileen 17% of the time. That means that Eileen beats Goh 83% of the time. So much for game balance...

The balance in VF5 isn't really from character to character at all. There are actually VERY unbalanced match-ups in the game. However, I think the balance comes in with the fact that you can punish someone so hard with good yomi.

Eileen has a HUGE advantage over Goh in general but this is lessened by the fact that if you know what the other player is going to do you will win. It's not that the characters are really that balanced compared to each other, it's that good yomi can used to punish so hard with any character that it has a balancing effect on the game...

How to read the "Winners" Character Specific Matchup Charts:
Looking at the Character's "Rows," you are looking at his/her wins or win ratio, respectively. Looking at a Character's "Column", you are looking at his/her losses or loss ratio, respectively.

For example, let's say we are interested on how Goh players fair against Eileen players for 1st place in Version C Tournaments. Going down the rows in the first chart, I find Goh's Row. Next, I looked to the right and find Eileen's Column. One can see from Goh's Row / Eileen's Column entry that Goh won only 1 time against Eileen. Looking at the second chart, that 1 victory against Eileen is a measly 0.17 win ratio (17%). If you want to see how many losses that actually entails, you can go to the first chart and look for Goh's Column / Eileen's Row entry (Goh loses to Eileen 5 times). So in the 2243 Version C Tournaments, Goh and Eileen fought against each for the championship, 6 times.

So Gernburgs, you read it correctly; however, since that 17% is based only on 6 fights, even if it is true Eileen has an advantage over Goh, that data is not "strong" enough to support it, IMHO.

Great, thank you man.

Thanks for all the time spent collecting this data, really interesting. It's kindah odd there's so few dedicated Kage players in the west compared to Japan where he's by far the most used and most efficient character in tourney.

That is not saying much. =P

Since a new version of VF is out, I am collecting tournament data again. Most likely update this once every 4 weeks.

The original and subsequent posts explains how the charts should be read.

Good stuff as always akai!

I never would have thought Jean would be so popular.

I'm very happy to see akira, eileen, and vane getting the amount of love that they are in R

Damn, what did they do to Lau? lol

Was looking like they couldn't tame/nerf Kage or Lau's results up until now.

Didn't think Wolf would be sitting where he is with the new mc throw damage adjustment.

Taka and Jeffry are skrong. Damn skraight.
The Evo boys ain't getting no love, then again neither is Taka but I bet he will be a favorite soon.

Yeah, Goh and Brad were always at the bottom in character usage.

I am more surprised to see Kage and Pai being used considerably less in the recent tournaments. Every time I collected the data in the past--Kage, Pai, Akira, Lau, and Jacky were always the top 5 characters used in tournaments. Not sure if the Kage/Pai players are switching over to Jean/Taka in VF5R or were there gradually less Kage/Pai players at the end of the version C days (when I stopped collecting data)...

Nice to see Jeff get a boost up... Lei and Kage kinda dropped didn't they? And I agree with Tony, the EVO boys need some beefing.

Thanks for this akai!!

I took the raw numbers and crunched them in a different way for an added perspective. What I did was the following:

- Used three groups: CHAMP, TOP 2, and TOP 4

- Calculated winning percentages for each group off of Entry numbers

- Calculated the standard deviation of those percentages

- Used those percentages and standard deviations to form a preliminary tier list

Here are the results (character placement within a tier based on descending order):

TOP 4
More than 2 sd above mean: Jean
1-2 sd above mean: Taka
1 sd above/below mean: Lion, Shun, Akira, Lau, Jeffry, Kage, Eileen, Jacky, Sarah, Aoi, Vanessa, Lei, Goh, Blaze
More than 1 sd below mean: Wolf, Pai, Brad

TOP 2
1-2 sd above mean: Taka, Shun, Jeffry, Akira, Jean
1 sd above/below mean: Lion, Kage, Goh, Sarah, Lau, Jacky, Eileen, Vanessa, Aoi, Wolf, Brad
More than 1 sd below mean: Pai, Blaze, Lei

CHAMP
1-2 sd above mean: Akira, Jeffry, Shun, Taka
1 sd above/below mean: Jean, Lion, Kage, Sarah, Eileen, Vanessa
1-2 sd below mean: Jacky, Lau, Aoi, Brad, Lei
More than 2 sd below mean: Goh, Pai, Blaze, Wolf

How to interpret the above?

In terms of the sd (standard deviation) numbers, for a character to be placed 2 sds above or below the mean is statistically significant. This means that it's not likely due to chance -- this character really is overpowered or underpowered.

If within 1 sd however, there is a big probability that differences in results are due to random sampling rather than fundamental advantage/disadvantage. More than 1 sd and less than 2 sd is a grayer area.

In terms of which group to go with, it depends on your belief in what level of play should drive tier lists. Is it at the highest level where elite players play? In that case, you probably want to weight results for the CHAMP group more than the others.

Perhaps you believe that at the CHAMP level it's more about individual players than it is about character (i.e. Itazan would win tournaments whether with Shun or with Pai) in which case you may want to go with the TOP 4 results. Personally, I am more of this camp. Getting to the top 4 in a tournament is already pretty hard and should be a good enough of a proxy for character strength.

Looking across the results, my conclusions are

- Jean and Taka may be overpowered (or players haven't gotten used to defending against them yet)

- Pai and Wolf may be underpowered

- Shun and Jeffry are strong but not as Godly as many feared (thus far)

- Kage, Lau, Jacky and Sarah are firmly mid-tier (so each character's respective players can stop crying now)

Although virtually impossible to quantify, I'd say this is playing a large factor to their success. I think if we were able to perform the same analysis for VF4:Evo shortly after its release, we may see a similar trend with Brad and Goh.

Although I'm speculating, I don't think that the inability to defend against new characters applies as well to new characters introduced at the outset of a new VF (i.e Shun/Lion for 2, Taka/Aoi for 3, and Lei/Vanessa for 4) as players are also learning new game mechanics as well.

Ice-9: It is always good to look at things from a different perspective. Thanks for putting time in your analysis!

As for Jean and Taka being overpowered...

The skill level between players entering the tournament will vary considerably (uncontrolled variable). I assumed since the tournament format eliminates players that should generally be "less skilled," those making it to the finals should be at a more comparable skill level.

Both Jean and Taka players have a high success in advancing to the semifinals. If you look at only the success rate of Jean and Taka finalists in winning the tournament (winner's column, players not making it to the finals are ignored), they are ranked average at best. So...I don't think they are overpowered, maybe overpowered for the less skilled or inexperience players.

One month of data, especially with a recent system change, is not very reliable, imho. Give about two more months of tournament data and the ranking charts would be more consistent. The chart could look quite different from how it is now.

If anyone have suggestions or ideas to better present or analyze the data (that won't take too much extra work) let me know!

ice-9, I think your list is all well and good, but something like a metagame, doesn't show on it. So there will always be that bias in the data.

Take for example, Pai being so low right now, if you see the characters that are used the most, perhaps now Pai doesn't have as strong tools to battle those characters like before, or maybe those characters are better equipped to deal with Pai now.

If there were more Goh's and Aoi's playing this time around, it might offset Sarah's numbers for example.

Basically what I'm saying is that, a characters usage frequency can offset the characters that are unfavorable matchup wise.

One thing that impresses me so much about VF is the fact even the least played characters, are played quite a bit compared to the rest of the cast. You just don't find that in other fighters. Not in tournament play anyway.

Jin that's exactly what I was thinking. That's a concrete testament to the balance of the game.

You could just tossed them charts out of the Window. The results are based on what YOU do. You control your own destiny if you ask me.

A reply to a question about a Version C chart:

That's a good thing to keep in mind -- these numbers are just numbers and may hide certain things that are going on.

Still though, outside of opinions from a few elite players, this is probably the best way of sussing out a tier list. It certainly at least makes for interesting conversation. Once we get a few more months of data, the results should grow more robust.

NP akai, it was an interesting exercise for me. What would be great is if you could keep publishing the raw data along with your charts, so I and whoever else desires to can cut them in a different way.



I thought about the above for a bit...I think it makes sense so recut the data to reflect it. What I did was take the winning percentage of the number of champs from top 4 finishes.

The results...

1-2 sd above mean: Jeffry, Akira, Shun
1 sd above/below mean: Kage, Sarah, Lion, Vanessa, Eileen, Taka, Jacky, Jean, Brad, Aoi, Lau, Lei
1-2 sd below mean: Goh, Pai, Blaze, Wolf

While no character was more than 2 sds above or below the mean (great news as this means the game is quite balanced), it's interesting to see Jeffry, Akira and Shun at the top -- two of three many here pointed out as top tier. Going with the above, it looks like Akira deserves to be mentioned along with Jeffry and Shun.

Thanks for these. Its really nice to see the numbers of participants per character pretty balanced.

Go go Jeffry!

And what have they done to Lei?

ps. Please have a link to the legend since I never remember how am I supposed to read that chart.

Most of those numbers aren't what I expected at all. Looks like Shun is back to his Verion A-B days though

The very first post in this topic explains how the chart should be read. If that does not make sense, let me know and I will try to elaborate on it or something.

Even with all the changes. Goh doesn't look like a popular character at all.

Nice to see that Shara and Lion got a boost!

Damn Jeff. Here come the bandwagoners =)
And people thought Shun would be broke.

So in the 'ranking chart', lower number is better?

In the 'total' chart, the numbers represent the amount of players who achieved that final standing in a tournament using the specific chacater? Or does this chart also use some sort of 'ranking'?

wolf got a downgrade

In "ranking" chart, lower number is better.

The "total" chart, is the raw data I collected that is used to make up the "ranking" chart. Some VFDC users like to see the raw data, so I add it along with the colored chart.

For an alternative cut at the data, here's what I have on the basis of the tournament winner as a percentage of the tourney's final four.

S tier (>2sd): Shun
A tier (1sd to 2sd): Jeffry, Akira
B tier (-1sd to 1sd): Lion, Sarah, Kage, Eileen, Jean, Jacky, Vanessa, Taka, Wolf, Aoi, Lau, Pai, Brad, Blaze
C tier (-2sd to -1sd): Goh, Lei

Conclusion? VF5R is a game so far with a big middle class and a few strong/weak characters.

Shun is clearly top tier and Jeffry is close behind (he just missed the cut off to be 2sds above the mean). Akira is again top tier. Goh and Lei are bottom tier but quite decent; Goh just missed the cut off to be in the middle group and Lei's win percentage is about 1.5sd below the mean which isn't bad.

What was it they did to Lau to drop him so much? I figured he'd sit comfortably in his usual spot still haha.

A little OT but KrsJin happy bday!

Thanks a lot man haha

Everybody got a serious boost, Lau only got an average boost it seems.
He still seems to be more fun to play and that's the most important I guess.

And happy Birthday

If we had data for every fighting game, the way standard deviation is calculated, you will always have a "big middle class" for every fighting game.

If you want to use SD as a measurement of "balance," I think a better way to look at the data is the actual value of the standard deviation compared to another game's standard deviation value. You want to see how spread apart are the characters from one another. Thus, I don't think the data by itself can be used to say how balance the game is.



It is still too early to say confidently, but comparing the character specific match-ups of Lau players in version B, C, and R against Jeffry players:

B - 78% (7-2)
C - 63% (17-10)
R - 17% (1-5)

I think one thing to point out is that the standings for 5R are only based on 2 months of data, and seeing as that Jeffry is a relatively easy character to learn (his moveset isn't as massive as, say, Lei or Eileen) it's possible that a lot of Jeffry players have been able to do better with the new changes than other characters. I think as time passes you'll see Jeffry move down slightly... he won't go down to bottom tier but you might see him end up as an upper-mid tier character.

I don't know, Reno. Your synopsis is entirely possible, but Jeff just looks really mean in 'R, and not necessarily because he is easy to play with. I'd go as far as to say that Jacky and Lau are both about as easy to play.

I think the majority of it has to do with the fact that he can be played safely and average about 85+ dmg per correct guess.

very good insight Reno, I didn't even think about movelist complexity. Though Shun is also doing very well and his movelist is friggn' huge. Maybe shun players just have better memory or something since they're used to a huge moveset?

Hmm, I think I see what you're saying, because by definition the normalization process will always create a big middle class. Let me think on this...



WHOA WHOA stop, I never said anything about balance. Let's not go down that road again...

As usual, the Semifinals, Finals, and Winners rankings is based on success rate (two types of calculation was used). Entry rankings is based on usage in tournaments. Lower numbers (1) is a better score than higher numbers (19).

I do not think most of the characters rankings will dramatically changed in a month. The exceptions, I think, may be the winners ranks, especially, Jacky and characters that are less represented (Goh, Lei, Taka, etc)

Judging from these bum ass charts Shun has the most 1st plce Victories with 78 followed by Jacky's 70.

Some body must'vecnducted this experient with these 2 in mind.
Chat don't mean shit. I control my own fate.

wagghhh pais at the bottom, anyone know why shes ranked so low now

Shun finalists record for winning first place 76-49.
Jacky finalists record for winning first place 70-82.
Jeffry finalists record for winning first place 69-56.

How the "experiment was conducted" is based on the same formulas I used over a year ago for VF5 version B and C charts. The formulas does not change, player's ability to win and game system do change.



Chat...do you mean chart? I do hope you control your own fate, as do the players success in the tournaments control their own fate that make up the charts.

If you have forgotten the post and reply I made to you earlier...



Unless I am misunderstanding what you are trying to say, I think you are not liking the charts based very little on the facts and more on ignorance. If I am wrong, please state more clearly the problems of the charts.

I think the roster is biased in favor of characters whose names begin with letters from the first half of the alphabet.

There! I too can make absurd statements that are statistically true. Do i get a cookie?

Here is your cookie. (but you did not made your absurd statement to be statistically true, though :P)

Yeah "S" is definitely towards the end of the alphabet.

Hmm... I wish there were more Taka players. And Jean does well! That's good news

So Lei Fei is the least popular character (barring taka which we don't have) over there? That sounds like the exact opposite of the xbox live situation

I never said i don't like the charts so please don't get it twisted Akai. I just was mentioning that charts is for playing in general whereas anybody can control their own fate at any giving time.