![]() ![]() Nothing particualy special but these highlight why some fighters fail. The lowest 10 ranked fighters on the other hand had no real advantage and served only to pad out the win rate everyone in the middle of the spreadsheet or each other. Fighters like Ken and Cody got so far thanks to aggressive AI and one powerful Special attack, for Ken being Shoryuken spam and Cody having his Zonk Knuckle which confused AI. Just to show off my point here were the 10 best ranked fighters, their Win% and Match Points.įighters with a standard grappler style out performed everyone else, only losing to another grappler or the odd power house with one exception. In that same tread, if a character had a special move they would use it whenever, and the better that move the better they generally did. The damage is obviously insane and the consistency was stupid, a round was literally Zangief doing 5 SPD in a round. Each fighter had 43 fights(not counting the mirror) had a Win% and a Match Point Total.Ī few things I noticed, Grapplers perform incredibly well since the AI will use their grapples as soon as they're in range, opposing AI can't tech it, won't avoid it, and short of an EX move nothing will beat it. For example Guile who landed in the middle with 51% tied with 4 other character but lost a dozen more rounds then anyone else at this percentage and even lost more then characters at a lower percentage. this was done to have a second measure for fighters in the case of a tie and to look at that fighter's actual consistency. I recorded two things, who won a match, and the number of rounds taken in that match by both sides. The strong ones will usually win and the terrible ones will usually lose. In practice variations are stronger for fighters closer to that 50%(which is rare) win rate, given they had the most toss up match ups, and the farther you get from that percentage the less these toss ups exist. Some fighters had clear superiority over others while others clearly terrible and underperform. While a single match for any given match is never gonna be as accurate as it could running the fights it became clear after a few hundred fights that their was a general outline for who would rank where. The Ultra combo Double being a compromise so that fighters would have all options and avoid rounds being won by a single attack, especially since some fighter AI are incapable of properly using particular Ultra's effectively. A single, 4 round match, the AI difficulty set to Hardest, and each character was to use their Ultra Combo Double. I ran these fights the latest patched, Steam version of Ultra Street Fighter 4. The set up for each fight I did was simple. ![]() To put it simply I had every character in the game fight every other character, and recorded the results and found their relative win rates. This spreadsheet started as a fun diversion I started in early October last year and get it out of my system. Don't know why but I do Street Fighter 4 is probably my favorite, to the point I would do AI tournaments. So let me preface this by saying I've always had fun watching games that allow you to watch the computer fight itself. ![]()
0 Comments
Leave a Reply. |
AuthorWrite something about yourself. No need to be fancy, just an overview. ArchivesCategories |