| |
|
 Subscribe in a reader
Hughes-St. Pierre III By the Numbers
Check out our article at MMA Weekly examining the match-up between Matt Hughes and Georges St. Pierre. The article analyzes three levels of material:
Primary Research: How have these fighters done against each other?
Secondary Research: How have these fighters done against common opponents?
Tertiary Research: How have these fighters done against different opponents.
Programming Announcement
Due to the holidays, there will not be a new report up tomorrow, but stay tuned during the coming week for coverage both pre- and post-UFC 79.
How Does One Score an Atomic Butt Drop?
The fight between Wanderlei Silva and Mark Hunt, the report for which is released today, presented two new challenges for FightMetric. The first is expressed by the title of this post: How does one score an Atomic Butt Drop? It doesn't fall into any ready strike category that we're familiar with. This is a case where FightMetric principle #3 really comes in handy. It doesn't matter what you strike with, be it fist, knee, or derriere, only where you strike to.The second new challenge was the question of what to do with a yellow card during a Pride fight. FightMetric has a mathematical way to deal with a point deduction in fights contested under the Unified Rules, because that penalty must be factored into any discussion of fight judging. But a yellow card does not directly affect the judging of the fight. The penalty for receiving a yellow card is purely financial; the penalized fighter loses 10% of his pay. While, the action leading up to the yellow card - be it passivity or a foul - must be factored into a judge's decision, it is not the yellow card in itself that is important. For that reason, we have chosen not to change any scores based on yellow cards during Pride fights. Given the inconsistent and capricious fashion in which the cards are given, it's probably for the best.Agree or disagree? Let us know at info@fightmetric.com.
FightMetric Gives You a Major Upgrade
Today marks a huge advance for FightMetric as we release our new interactive fight report format. All new reports will be done in this format and our existing reports will be converted into this dynamic format over the course of time. The first one to get the upgrade is for the fight between Stephan Bonnar and Keith Jardine.The goal of this new format is to put a massive amount of information into your hands, but in an easy-to-consume package. Make no mistake, this report contains a huge amount of information: 842 unique data points, to be precise. The navigation is simple and intuitive and should be familiar to anyone who has surfed the web. The PDF format was a good start, but it has its limitations. As a static medium, you can only see exactly what we put on the page. In addition, there's a limit to how much information one can cram onto so many pages. A Flash format removes those limitations, allowing unfettered access to the entirety of FightMetric's stats for a particular fight.FightMetric's founding principles assert that context is key. A strike is not just a strike, a takedown is not just a takedown. Everything that happens in a fight has several identifying qualities. What was the strike's target? When did it happen? From what position? With what force? FightMetric's new report will allow you complete access to all of that information. Using the selectors and charts, you can instantly find out how many times a fighter threw a power strike to the head on the ground in the second round. Or how many successful slam takedowns from the clinch in the third round. More importantly, by viewing stats separated by these distinct qualities, we can come to a better understanding of a fighter's performance. We hope you enjoy the new look and functionality these reports provide. As always, we love to hear your comments and questions, which you can send to us at info@fightmetric.com.
FightMetric Breakdown of Huerta-Guida
Check out our article at MMA Madness to see FightMetric's breakdown of the incredible fight between Roger Huerta and Clay Guida.
Survey Says (cont.)
On Tuesday, we revealed the responses to the first two questions in our survey. Today, we get to the third and most difficult question.
3. Not counting opponent quality, what are the factors that most contribute to the quality of a victory?
(Rate the following on a scale from 1-5, with 5 being the highest)
Time
Effective Offensive Volume
Accuracy
Effective Defense
Method
You'll have to forgive us a little when it comes to the data for this question. Our intention was to force people to rank these five factors in order from 1-5. Instead, the survey asked them to rate them. The resulting data is not as precise as we'd like (some people just don't like giving 1's or 5's), but the conclusions should be valid just the same. 
It's no surprise that the highest scoring factor is effective offensive volume. MMA matches are all about successfully attacking your opponent. It's true, some fighters (specifically BJJ fighters) will let their opponents take them down or punch them because it serves their ultimate purpose of setting up an attack of their own. In general, however, you can usually tell who is doing a better job by tracking who lands more HiPer techniques. Nearly 90% of respondents rated offensive volume as a 4 or a 5.
The second-highest rating went to accuracy, which was a bit of a surprise. Nearly half of respondents gave this a 4, with an almost equal number rating it either a 5 or a 3. The implication is that accuracy is a very important thing, but not the most important thing. We found this unusual because accuracy is a factor you almost never hear about when people talk about fights. This is likely a result of there being no means to comprehensively track accuracy until recently. The implication (which is good news for a service like FightMetric) is that fans aren't talking about this but they really wish they could.
Third on the list was method of victory. While an average of 3.72 looks high and is pretty close to the results for accuracy, that number should be taken with a grain of salt. While 62% of respondents said that method of victory rated either a 4 or 5, the results from question #2 indicate that 67% of fans feel that method doesn't matter that much at all. If they were willing to say that a decision can be worth more than a knockout, how important could method really be? This is where the ranking vs. rating problem is most apparent. Our feeling is that the answers to #2 are a more accurate reflection of fan sentiment. Method doesn't rate as highly as the numbers would indicate.
Next on the list comes effective defense. This was the answer that had the largest spread of responses. A nearly equal proportion of respondents rated defense a 4 or 5 as did those who rated it a 2 or 3. People didn't quite know what to do with defense because it's a difficult thing to assess in a sport like MMA. In sports of possession, like football or basketball, after you score you have to give the ball back to the other team. At that point, the goal is to prevent them from scoring, which is an easy thing to measure. But in MMA, the goal is not so much to prevent your opponent from scoring with his techniques; it's to prevent him from ever attempting techniques at all. MMA is a like pick-up basketball where the game is to seven and the rules are make-it-take-it. The best kind of game is the one where your team rattles off seven scores in a row and never plays defense at all. The best defense really is a good offense.
Last on the list is time. This was ranked much lower than the other factors, with 80% of respondents rating it a 3 or lower. It also received the highest number of 1's, indicating no impact at all. Fans seem unconcerned with how long it takes to win as long as the action along the way is of high quality.
Among the 205 responses were 17 write-in answers for "Other." Factors receiving multiple write-ins included damage inflicted (5 votes), controlling the location of the fight (3 votes), aggression, and heart (2 votes each).
Conclusions: The ratings point to a three-tier breakdown. Offensive volume and accuracy are a ways above method and defense, while time lags far behind.
The calculation will be heavily weighted toward rewarding successful offensive techniques. While the FightMetric system awards points for what one could call defensive grappling actions (like escaping from mount), this calculation will ignore those and focus only on techniques that press the action: Strikes, takedowns, guard passes, and submissions.
Accuracy will receive a higher weighting than initially proposed. Mentions of accuracy have started to pop-up. Witness people's admiration - including quotes from several UFC higher-ups - for Anderson Silva's amazing performance against Chris Leben, in which he connected on every strike attempt. As fight tracking becomes more prevalent, this is a category that will only grow in prominence.
Weighting for method gets depressed because of the contradiction between answers to question #2 and question #3. In general, you can trust someone's answer to a simple yes-or-no question better than a judgment call based on rating 1-5.
Defense and time fall to the bottom of the heap. Still important, they will be included in the calculations, but should receive low weights.
Keep watching FightMetric for the release of this still-unnamed fight quality statistic. The calculations are in the final stages of testing and fine-tuning and will be ready in the very near future.
Who Remembers Ricco-Nogueira?
The swarm of protest following the Hamill-Bipsing decision was probably the loudest reaction to a questionable decision in years. The decision that preceded it was for the fight between Ricco Rodriguez and Antonio Rodrigo Nogueira, the FightMetric report for which is released today. Fans watching that fight saw Rodriguez take Nogueira down three times and maintain top position for more than 15 of the fight's 20 minutes. In America, numbers like those guarantee a victory.
Not so in Japan. Fan displeasure with the judgment was so profound that Pride found it necessary to issue an official explanation via Matt Hume, who was a judge for the fight. In short, Nogueira attempted seven submissions and Rodriguez's strikes never put Nogueira in any danger. That "effort to finish the fight" gives Nogueira the victory without considering what else happened in the match.
FightMetric doesn't work that way. Though point values are weighted to give credit for techniques that result in fight endings (such as submission attempts), the system tracks everything, no matter how small. So it tracks the 100 head jabs Nogueira landed on the ground or the 57 body jabs landed on the ground by Rodriguez. But after adding up all the numbers, the score still favors Nogueira, mainly on the strength of those seven submission attempts.
Pride rules may be overly simplistic and FightMetric rules may be overly complicated. It will be fascinating to see how often the two systems produce the same conclusions.
Survey Says...
For the past three weeks or so, we've been conducting a survey to ask the MMA community what it values most when evaluating a fighter's performance. The over-arching question is: What makes one win (or loss) better than another?
The goal of the survey was to accumulate data that will help inform the values behind our next great piece of statistical analysis, a statistic that measures fight quality on a scale between 0 and 100. Think of it as acting like the NFL's QB Passer Rating. It is made up of several statistical components, each weighted differently, and produces a single quality score. While the score this kind of calculation produces will be based on hard numbers and will be completely objective, getting the priorities right in the first place involves fans' subjective opinions. One fan might value speed, another could favor accuracy, while another could give weight to damage inflicted.
So rather than rely on a too-small opinion base (namely, our own), we opened these questions to the MMA community. The statistic will only be trustworthy if its values are consistent with what MMA fans at large already believe. To get a cross-section of MMA fans, we created three copies of the same survey. One was posted to the forums at MMAWeekly.com, one was posted at The Underground, and the last was placed on our blog, which got most of its responses from people who heard our interview on The Jordan Breen Show.
A big thank you goes out to all of you who participated. Thanks to you, we were able to reach our target "n" of 200 (we received 205 complete responses, to be exact) and get some meaningful results.
The survey asked three questions, which we will evaluate in order:
1. Is any win ALWAYS better than a loss, or is it possible that an impressive loss can be worth more than an unimpressive win?
More than three-quarters of respondents answered that some losses can be better than some wins. This is a difficult question to consider because we have been trained to think starkly in terms of wins and losses. After all, that's all a fighter has to show for his efforts and is the only consistent evaluation criteria available.
The principled answer is to say yes, a win is a win and a loss is a loss. The problems arise from the kind of win, which gets to heart of our research of quantifying fight quality. What do you do with a win that come via bogus decision? Should that "winner's" performance really be worth more than that of a guy who barely lost a thrilling back-and-forth war? Or removing decisions for a moment, when a fighter gets dominated the whole fight but manages to win via lucky cut stoppage, whose performance was better overall, the cutter or the cuttee?
Conclusion: Our calculation will not place too much weight on who actually got the win in a fight. To be sure, the winning fighter will get points for getting that W, but not too many. It is even conceivable that the losing fighter will have a higher score than the winner based on the cumulative events that transpire in their fight.
2. Is a finished fight ALWAYS better than a decision win, or can an impressive decision be better than an unimpressive TKO/Submission?
Similar to the first question, this question forces you to make a black-or-white choice. It's not easy and it's not fair, but it's necessary to produce objective results.
The results again point toward nuance, though the gap between the two opinions shrunk. Two-thirds of respondents feel some decisions are better than some finished fights. Here, more fans chose the principled answer. The FightMetric system can't help but agree with them. The first principle of the system is that the goal of every fighter is to finish his fight. Not just win, but finish. In the case of a decision, a fighter has failed to achieve his goal.
Still, things are rarely so black and white. Fedor spent 20 minutes dominating CroCop and Nogueira (twice) while it took Gabriel Gonzaga 14 excruciatingly-boring minutes to knockout Kevin Jordan. The principled answer is that Gonzaga's win was better. It's a pretty safe bet that most fans wouldn't see it that way.
Conclusion: Same as with the first question: The calculation will give some weight to finished fights, but not too much. In this way it is not just conceivable, but likely that many decision wins will have a higher quality score than even a knockout.
Stay tuned for the results to final question in the survey, which we will post Thursday.
|
|
|
|
