I’m guessing that most readers of this blog will be familiar with a tier listing for a game with matchups and a choice of characters or races with different ability sets. Admittedly this is mainly only fighting games where I’ve seen this kind of thing happen, but it could be doable for other games too, for example, you could theoretically try to ‘tier’ the various different Civs for games of Civilization Revolutions, or tier the different characters in Sega Superstars Tennis or many other sports games.
But in case you aren’t familiar at all, the basic idea is to take the very best player(s) with that given character & game, and have them rate their chances against each other character in the game in a theoretical matchup of ‘players of equal skill’ – in this case, meaning the very best player(s) with that other character also. This rating is generally given as a score out of 10, which is also sometimes seen as a rough guide to the number of wins out of 10 matches that would demonstrate the players are of equal skills were they to win that amount. The “points scores” are then added up, to give the theoretical ‘best characters’ in the game, and the weaker ones, and these are often clumped together into groups of characters that are at a similar points score, and these groups are called names like Top/Average/Low etc; hence the use of the name tiers. Please note there is usually plenty of disagreement about the definitions of how to do these points scores, and naturally, even if that can be agreed, things are argued often and tier listings can change over time as new techniques and information come to light about a game. Where the two top players disagree, eg: the top Ryu thinks Ryu is a 6-4 over Guile, but the top Guile player thinks Guile is a 5-5 even match with Ryu, the scores are usually then averaged, so the score will be shown as a 5.5 advantage to Ryu in the table, or +0.5 for Ryu, -0.5 for Guile.
No-one has really tried to put together a SSF2T HDR tier listing yet, partly due to there being no major tournaments with it yet, and also due to the unfortunate bugs in the game, but the latest one for Super Turbo can be found here:
Now one thing people tend to forget is that even the tier score listing can give a very poor idea of who’s the ‘best character’ to pick, as the final score is unweighted, and thus assumes all characters are played equally, which is of course not the case. In a tournament setting, how the character fairs against the other (at least perceived) top-tier characters is going to matter more, or for XBL or online play in general, how the character deals with the more popular online characters (such as the shotokan scrubs) or the more lag-friendly characters (such as Dic, Claw and Blanka) may also matter more. So you have to take all of that into account depending on the situation as much as, or even more than the raw number. Again, as I am sure most readers are familiar with, this is what’s known as a metagame.
In a relatively unknown metagame, or in a knock-out type of situation, such as some tournament formats, then a characters ‘worst’ matchup can matter as much as, if not more than their overall scoring. As also commented very accurately on the New Arcadia link above by NH2, this is what makes Dictator a much better character than his raw numbers, as his worst matchup is still only a ‘3’, meaning you have at least a reasonable chance against an unknown field in all matchups.
For a personal example, both of these factors contribute to why I am now maining, and doing better with Deejay than Guile these days, as I find he is better ‘vs the field’ in the XBL metagame, and especially, has fewer nightmare matchups for me. It’s really only vs Brouki (& maybe Dhalsim) when I feel I have very little chance to win with Deejay.
Even if one understands a metagame though, people still get confused with tier lists and what they really mean. Often new players confuse it with characters they find are easy to play with and/or learn. I find a good example on HDR is Bwanka, especially online. He’s got a huge throw range, an un-techable throw, some really high priority jumping attacks, a fairly easy crossup, is generally good at aggressive rush-down style of play that’s more effective in lag, and super-fast specials that become almost impossible to react to in heavy lag; basically, overall he is a very easy character to play with to an average level, and generally gets better online, hence why he is known as Bwanka to me 😉 . I’d even like to use my own play as a good example! I really don’t have a clue what I am doing with Bwanka. I’m such a ‘n00b’ I don’t even know how to perform all his moves; the hop-about move is a mystery to me still, as I never really play him seriously ever, yet, using Street Fighter basics, I can get generally decent results with Bwanka despite my huge lack of skill & knowledge with him. I’m equally as weak a player with say, Boxer or Dhalsim, yet I don’t do nearly as well with them as I do with Bwanka. Sometimes this ease of play at a low or intermediate play level, is indicative of the same situations at high level play, but here in this example, it really isn’t true, as Blanka is generally low-tier overall, yet Boxer and Dhalsim are right at the top. They just have a much steeper learning curve, in my opinion. Blanka maybe relatively easy to play at an average level, but conversely he’s actually really difficult to master and win with at the very very top.
But this brings up another interesting point or ‘problem’ with tier listings like this. They only apply at the very top only. For example Guile is listed as a 6-4 advantage vs Blanka. If I remember correctly, Muteki Guile (probably the best Guile player in the world) actually rated it as an 8-2 advantage to Guile himself. Now at this ‘uber-top-pro level‘, having watched some of their games, I can actually easily believe this score is honestly reflective of what he experiences.
The trouble is that tier listing won’t mean much for your regular or even perhaps a very very good player in some cases. For example, for me, an above-average Guile (well, perhaps not, but I was at least good enough for top 300 in the first few weeks of HDR, when I was maining Guile almost the whole time in ranked – before the ratings bug hit), playing against an equally skilled above average Blanka is a nightmare. Even though I don’t think I’m good enough to accurately rate it as a 10 point score, I’d certainly rate it as a ‘large disadvantage’ to Guile, so I’d put it as low as a 2-8 probably. Part of the ‘proof’ is when I switch to another character that I am generally weaker with, know less about, yet find I do better & win more easily against the very same Blanka player.
What you’d really need to express this issue into figures, is multiple tier listings based on the skill level of the players involved. This would enable you to see how a character’s tier changes as your skill level progresses, effectively demonstrating a characters learning curve, rather than only showing how good that character is at the very peak of that learning curve, which very few people will attain.
To do this you’d need an accurate measure of the players skill – the closest thing we have to this would be a players ranked skill score with a particular character. A game would then need to collect all the stats of who won against who, along with that players ‘skill rating’, and then only using results where players were of comparable skill to compare to create a matchup chart and thus the tiers. Put that data onto graphs and charts; and you’d be able to see actual learning curves, matchups vs skill rating, and ‘top end’ character ability displayed graphically. Perhaps it’s only me, but personally I’d find that kind of thing absolutely fascinating!
About the current state of play (and data) with tiers in general though, I’ll leave you with this final thought which summarises parts of this article into a great post from Chaghatai, from Capcom Unity, (with corrections and emphasis added):
The difference between the people is likely* to be greater than the advantage or disadvantage imparted in a matchup. And in any case those listings assume both players know all the counters and counter counters inside out for that matchup.
Amateurs as a group and usually as individuals are more inconsistent than pros. So what for a pro makes a real difference, to an amature, they may not notice the difference at all.
It’s like expensive golf clubs – indespensible for a pro, but most amateurs wouldn’t notice the effect on their game.
People, don’t worry about your matchups too much, just have fun.
“But Otochun says this match is a disadvantage….”
Are you Otochun? Is your opponent Daigo?
Then don’t worry about it.
Tiers are real in a sense and can be interesting tools to analyze a tournament or whatever, but on a day to day basis, they’re nothing to get overly hung up on.
* I’d personally note this depends a lot on the game itself! I find HDR to be pretty good, and certainly much better in this regard than Super Turbo, due to the rebalancing attempted.
Of course, if you believe M’s infostation, you can pretty much disregard this article. An M-tier listing is very simple, since every character will either win or lose to each other character; it’d obviously be a 50/50 chance.