Character Tournament Rankings in Japan

Final update to Character Ranking Charts

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....

 

Re: Final update to Character Ranking Charts

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 /forums/images/%%GRAEMLIN_URL%%/smile.gif Any questions, comments, or criticism about the data is welcome.

 

Re: Final update to Character Ranking Charts

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.