MMAPredictions

Warning - do not read this post unless you want to better understand how predictable mixed martial arts events are. If you are curious about MMA and how easy (or not) it is to correctly pick fight outcomes, come along with me intrepid reader.

So, for those who are new to MMAPredictions.com, here’s what you need to know: we allow users to make their selections of fight outcomes on MMA events, allowing them to determine how well they know MMA. What that also does is provide data on whether an individual MMA event was easy to predict or not because we can look at how folks did on average with their picks.

For example, if the average MMAPredictions user correctly predicts the outcomes of 90% of the fights on the upcoming UFC 84, we can say, “Mister MMA Promoter, you’re not doing a good job of putting together events because they’re way too predictable”.

What if the opposite happens? What if the average person correctly predicts 10% of the fights on a card correctly? If it turns out that MMA fighters deliver upsets that often, what the heck is it that we are watching anyway?

Okay, to continue this mental exercise, where do we draw the line? At what point do we say, “that fight card was too predictable”, or “that MMA event was too full of surprises and made no sense”?

I think we would all agree that if everyone gets 90% of a card correct, that’s too predictable. At that point it’s almost scripted. But is an average prediction accuracy of 80% for, say Tuesday’s DREAM 2 acceptable? That’s pretty boring too. But you get the idea.

Honestly, I don’t know where to draw the line. But I thought that we could take a look backwards at the events we’ve held to date on MMAPredictions.com and see if there’s anything to learn there.

This blue chart depicts where all 66 of the fight cards we’ve had to date fall in terms of predictability. Thankfully, we have had no events yet where the average MMAPredictions.com user got more than 90% of the card correct. We have also had no events where the average person got less than 19% of the card right.

Here are a few things I come away with when I look at the graph to the left:

• Caleb needs a real graphing software because doing that in Photoshop took forever and still doesn’t look very good.
• Thirty-one events were unpredictable (that is, the average prediction accuracy is below 50%) and thirty-three were predictable (that is, the average prediction accuracy is above 50%). The fact that it’s so close is comforting, and jives with the information on our all time site stats, which shows that historically across all fight cards, an average user will likely get just a little over 50% of the fight card correct.
• If an MMA fight card is going to be unpredictable, it’s probably going to be just a tad unpredictable, falling into the 40-49% range. It’s pretty rare for a fight card to be less predictable than that, as only 8 of 66 events have been less predictable.
• If an MMA fight card is going to be predictable, it’s pretty safe to say that it will have an average prediction accuracy that falls between 51% and 69%, a wider distribution.

I think the cool thing about the graph above is that we have a better idea of what to expect moving forward. While it doesn’t tell us any absolute truths (does such a thing even exist?), we can look at a particular MMA fight card and say “Hey whoa that’s weird!” with a little more confidence.

My question for you is, do you care if an event is unpredictable? I think we agree that too predictable is bad, but is there such a thing as too unpredictable? Is it okay, or even a good thing if a card is so full of upsets that the average prediction accuracy is somewhere between zero and 20%? Please let me know your thoughts by posting in this thread in the forum!

- Caleb

Want text messages of MMA fight results sent to your cell phone for free? Register now!