In the past two years (and one week), we’ve seen 8 UFC shows take place in Brazil, with this amount steadily increasing — 1 in 2011, 3 in 2012, 7 in 2013 (3 still to come).
Given that a large majority of UFC shows and their fighter population reside outside of Brazil, how “fair” is it to these fighters that so many of their fights take place in Brazil against Brazilian opponents? When we say “fair”, we refer to two things — the “hometown advantage” and the matchmaking.
The “hometown advantage” is something we looked at very briefly a couple of years ago, when some bookmakers like Bwin started to claimed that in sports such as Poker, it can give you what you need to win. We found just under a 2% advantage for the home fighter (a hometown fighter with say… 98 rating points would have a 50/50 chance versus an outsider with 100 rating points). However, this included a slew of contests that involved neighboring entities (example — US vs. Canada) and I think we can all agree a United States vs. Brazil or even more extreme, Japan vs. Brazil at Brazil is a far cry from the aforementioned border wars. Plus, the database is much, much more mature now. A detailed re-visit of this analysis of this is better saved for another day.
Using the fighter “nationality” setting you see on our ranking pages, I isolated 60 fights from the 8 previously mentioned UFC shows that involved a Brazil vs. non-Brazil fighter in Brazil, that resulted in non-draw, non-no contest outcomes. Brazilian fighters won 47 of these fights, a win rate of 78.33%. If we expand this window back to 1/1/2010 and include all organizations where we know the nationalities of both fighters (399 fights), this win rate drops significantly to 62.15% – though is still sizable.
|So, how much of this is the mysterious “hometown advantage” and how much of this is matchmaking?|
To figure this out, I took the 399 fights and made a subset of fights where each fighter had at least 5 prior fights (excluding “no contests”) and both were at least 3 times above the “starter rating” just prior to the bout rating calculation. This criteria typically creates the most suitable environment for accuracy without sacrificing too many samples. Unfortunately, this annihilated the sample size in this case, reducing it all the way down to 95 fights — but we’ve got rid of all the noise and still have something to work with. Though, the set needs to be a bit smaller — we have to pull 4 outliers, fights where the difference between the two ratings exceed the mean difference standard deviation times 3.
After doing that, we’re left with 91 fights where the Brazilian fighters won 70.32% of the fights and held a 24.98% pre-fight rating advantage.
In the same date range where fighters had the same nationality and the rating advantage was between 20% and 30% for one fighter (all the other aforementioned fight qualifiers apply as well), the higher rated fighter only won 61.4% of the time, much less than the 70.32%.
Although our sample sets are small, we’re not aiming to be published in a scientific journal with this analysis – but the trend is clear. The success of the Brazilians is partly a result of matchmaking and partly a result of the hometown advantage. It’s tough to express how “strong” the hometown advantage is, but keep in mind that a 70% win probability is not observed until you get into the area of almost 100% rating advantage — far more than the 25% we were looking at here.
|Experienced Brazilian fighters hold a 55% hometown rating advantage* over experienced outsiders since 2010.|
Notice the asterisk though… because I often wonder, is there a third factor that we can’t quite quantify? I think so. Brazil has long been the “wild west” of MMA as far as record-keeping goes. It is very likely that the popular fight databases are missing a massive amount of fight data, but it is uncertain how this affects this study. Furthermore, the dataset is not as large as I would like, so I have to think this is on the high side and will come down a little bit. I may come back and expand the window back further in time.
In case you were wondering, I ran the query for the larger dataset (the 399 fight one for Brazil — which yielded a 62.15% win rate) for some other countries — with their rank in most shows since 1/1/2010 in parentheses (Brazil is #2).
- Poland (#7) – 70.85%
- Russia (#5) – 68.90%
- Japan (#4) – 55.84%
- Canada (#6) – 55.24%
- Germany (#8) – 52.34%
- UK (#3) – 52.30%
- Australia (#9) – 50.85%
- Mexico (#10) – 37.50%
The USA is one big melting pot at this point and I omitted them for just that reason. Given the above, things don’t seem quite right in Poland or Russia… and the scene in Mexico seems incredibly weak.
In totality, I think the moral of this story is — think twice before you take that fight offer in Poland, Brazil, or Russia. Try Mexico instead.
Posted on September 2, 2013 by jcs