Between those two events, which represent more than five and half years of monthly competition, there were 58 main events. Perhaps a better comparison is this: Tito Ortiz was champion at the beginning of my data set. A main event is classified as a bout that is specifically called a main event or co-main event. The only way a non-labeled main event can be counted in the data is if the fight is a title fight and the fighters are featured in equal proportion on the fight poster. For instance, UFC 81 showcased Frank Mir and Brock Lesnar, which was not a title fight, but was called a co-main event. Meanwhile, the UFC 79 fight between Georges St. Pierre and Matt Hughes counts because it is both a title fight and both fighters were featured in equal proportion to Chuck Liddell and Wanderlei Silva. On a different note, a fight like Mauricio Rua and Forrest Griffin (UFC 76) would not count even though the fighters were given equal space on the event poster. The fight was not labeled a main event nor was it for a title, so it will not be counted in these listings. Using this definition, UFC 81, 79, 73, 65, 64, 61, 59, 52, 50 and 46 featured more than one main event.

Of these 58 main events, the betting odds favorite won 40 of those events, or 68.96% of the time. These numbers seem to dispute the prevailing attitude in MMA that anyone can win any fight at any time, as they are significantly higher than the 60% usually given. It's also easy to forget that many of these "upsets" by underdogs are really not much of a surprise. Take UFC 79, where Chuck Liddell upset Wanderlei Silva. Liddell had been listed as +100, which offered him a 50% chance of winning. Silva was listed as a -125, which predicted that he would win 55% of the time. The overlap accounts for the money collected by the bookies. At any rate, does a 50% chance beating a 55% chance really represent that great of an upset?

My opinion is that it does not, and therefore, I needed to define what constitutes an upset. If a fighter is listed as more than +150, he is being given less than 40% chance of winning the fight, which matches the original odds given by new fans to just about every underdog. When we go back to the data and look at not just the frequency in which underdogs win main events, but also the frequency in which they overcome +150 odds to win, the numbers look much different. Now, instead of 18 of 58 underdogs winning 31.03% of the time, we have 15 of 58 underdogs overcoming large (+150) odds to win main events. The new fraction translates to a paltry 25.86%, which also happens to be the percentage of televised broadcasts during which Mike Goldberg incorrectly refers to Wanderlei Silva as "The Iceman."

If the odds needed to overcome are changed to greater than a +200 or 33% chance to win, then just 7 of 58 or 12.06% overcome great odds to beat a favorite. These are the true upsets in MMA and the one-hit wonders can easily be named off hand – Matt Serra, Gabriel Gonzaga, etc.

Whether you choose to use raw data (32.1%), conservative estimates (25%), or liberal estimates (12.06%), it's obvious that MMA is far more predictable than the less-than-scientific 40% victory rate often attributed to underdogs.

Even more convincing is the simply fact that when fighters are given a 33% chance to win, they only win 12.06% of the time, a spread of more than 20%. By comparison, when fighters are given a 40% or less chance to win a fight, they win just 25.86% of the time, a spread of nearly 15%. It is obvious that the odds are misguided as the conceptions of fighter ability.

With our statistics in place, it must be asked: would the UFC attempt to shape its company by giving champions favorable fights? On one hand, if the UFC books bouts with an outcome in mind, it allows them to choose (except for that 12.06% of the time) who will win, who will be champion, and most important, who will be the figurehead of the company. However, booking such favorable match ups for its champions such that they never lose creates a professional wrestling feeling in the sport, which I believe Dana White would fight to the death to avoid. Additionally, if fans were aware who is going to win roughly 87.94% of the time (using the liberal numbers), why would they tune in? Sure, it's fun to watch Anderson Silva win, but even I can't watch Silva-Leben more than five times in succession. In the end, I believe that the UFC is not booking matches with ends in mind. Sorry, Mr. McMahon.

Another interesting trend is periods of upsets followed by periods of expected outcomes within main events. Between UFC 40 and UFC 49, there are 6 upsets (raw numbers). This is largely representative of Randy Couture and Chuck Liddell gaining more experience, developing their styles, and finally upsetting top-flight competition. From UFC 50 to UFC 59 (the next batch of ten), there are just 2 upsets. This seems to represent the peak (and plateau) of the old crop of UFC stars, including Couture, Liddell, Andrei Arlovski, and Matt Hughes. For comparison, none of the four currently hold belts, one is inactive, and one is injured -- but for those ten events, they were kings. Between UFC 60 and UFC 69, there were again five upsets. This upturn in upsets is similar to what happened in our first sample of ten events – previously middle-tier fighters began to make their name through better training, preparation, and new skill sets. However, unlike Couture and Liddell, who came from within the organization, most of these upsets were at the hands of defecting PRIDE fighters. The additional difficulty of never having seen these fighters added to the troubles of UFC main event favorites. Between UFC 70 and UFC 87, an 80% larger sample size, there were just six upsets. This is due not to champions finding their niche and peaking, but to adjustments made to the PRIDE fighters as well as an increased hybridization of skills across all fight camps.

If the patterns tell us anything, they suggest that we are in for another period of upsets. When the upsets come, new fans will ogle over the unpredictability of the sport. Just remember: they're wrong.

View the spreadsheet here.