TUF 18 Finale

[Svino]

There's an easy way to tell if "value is overrated" among combat sports bettors. If it is, it will be consistently profitable to bet all the favorites -- this doesn't seem to be the case.

As for variance, I don't think it's that easy to say which kind of betting has higher variance. It depends on what variables you are assuming "stay fixed". For a fixed-swing bet, the maximum variance is for even odds. For a single-unit bet, the max variance is for the longest odds. For a bet to win one unit, max variance is for the most favorable odds. What you'd really want is the expected-variance to expected-value ratio, but that's kinda hard to figure out.

[SPX]

Some people take the approach of betting for value but i think there is too much variance in it.

For me it just depends on the fight. I don't bet every line that I think holds SOME value, but if I think the line is way off I'll bet it.

Gustaffson clearly had value, even in a loss.

[MMA_scientist]

I threw out the pi "help" symbol last week to you svino. But then Syriam terrorists deleted it.

In any event, I figured it out.

[SPX]

There's an easy way to tell if "value is overrated" among combat sports bettors. If it is, it will be consistently profitable to bet all the favorites -- this doesn't seem to be the case.

That makes sense. It actually brings up another question: Would it ultimately be profitable then to always bet the dogs?

As for variance, I don't think it's that easy to say which kind of betting has higher variance. It depends on what variables you are assuming "stay fixed". For a fixed-swing bet, the maximum variance is for even odds. For a single-unit bet, the max variance is for the longest odds. For a bet to win one unit, max variance is for the most favorable odds. What you'd really want is the expected-variance to expected-value ratio, but that's kinda hard to figure out.

No clue what the fuck you just said here.

[MMA_scientist]

That makes sense. It actually brings up another question: Would it ultimately be profitable then to always bet the dogs?

No. I think someone (probably svino) did a test on it and both strategies lost money last year (or maybe the year before). Even if it won in an given year, it would probably correct the next year, so you would never know when it would work.

[trotterz]

No. I think someone (probably svino) did a test on it and both strategies lost money last year (or maybe the year before). Even if it won in an given year, it would probably correct the next year, so you would never know when it would work.

I would like to see such an analysis for post ufc fighters in smaller orgs. I think they often loose and they are usually hight favorites

[Svino]

No. I think someone (probably svino) did a test on it and both strategies lost money last year (or maybe the year before). Even if it won in an given year, it would probably correct the next year, so you would never know when it would work.

Right. On the whole, bettors seem to have the concept of value (as opposed to picking the favorite) down across the board, there might be a couple areas that stand out:

1) Even though you can't just bet blindly here, bets in the +105 to +140 range tend to be better than most others. (And conversely, bets in the -120 to -160 range have done the worst.)

2) It's hard to tell because of the small sample size, but extreme underdogs (i.e. guys over +1000) have done very well in recent history. Think about Griggs over Lashley, Newton over Mo, Beltran over Gracie, "Whoever that was" over Bubba Jenkins, maybe Edgar over Penn. Those "one in a million" underdogs seem to win in more like one time in ten.


So that raises a question: "If underdogs become a good value at mere 10 to 1 odds, why do I see so many fights where it is clear that one guy wouldn't even win one time in forty?"

I think the answer is that it is just not easy as we think to tell which fights those will be in advance. I thought Jones / Gustaffson might look like one of those, for example. If you think of three different sets of betting lines:

A) The lines that you would set given perfect information about the fighters
B) The lines that you would set given all reasonably-accessible public information about the fighters
C) The actual sportsbook lines.

it is almost certainly true that the lines in "A" would swing much more extreme than "C" (there would probably be plenty of 40-1, 50-1 lines), but this fact is actually not relevant to bettors, because we do not have access to "A". All we can do is try to make money on the differences between "B" and "C", and there is really quite a bit of uncertainty about fighters in general.

[MMA_scientist]

2) It's hard to tell because of the small sample size, but extreme underdogs (i.e. guys over +1000) have done very well in recent history. Think about Griggs over Lashley, Newton over Mo, Beltran over Gracie, "Whoever that was" over Bubba Jenkins, maybe Edgar over Penn. Those "one in a million" underdogs seem to win in more like one time in ten.

It seems like MMA lines are getting steeper on obvious favorites. I was really surprised to see Jones at -1000 over Gustaffson, not like he was fighting Bob Sapp. I can't imagine that he would have been a 10:1 favorite 5 years ago. Just looking at boxing or football lines, this seems to be the trend. If a team is clearly favored to win, the line is absurdly steep compared to MMA lines in the past.

This also confirms my original hypothesis about MMA, that it is super dangerous thinking that anyone can ever be more than 80% likely to win, given a competitive matchup (throwing about Bob Sapp, and other squash matches).

We talked about Taleb a while back and betting what "can't happen" will happen, and these extreme dogs seem to give some merit to that approach if you have the patience, which I don't.

[SPX]

It seems like MMA lines are getting steeper on obvious favorites. I was really surprised to see Jones at -1000 over Gustaffson, not like he was fighting Bob Sapp. I can't imagine that he would have been a 10:1 favorite 5 years ago. Just looking at boxing or football lines, this seems to be the trend. If a team is clearly favored to win, the line is absurdly steep compared to MMA lines in the past.

Dogs win a LOT, so back on the discussion of value, we can only say then that dogs are holding more value now than ever before.

This also confirms my original hypothesis about MMA, that it is super dangerous thinking that anyone can ever be more than 80% likely to win, given a competitive matchup (throwing about Bob Sapp, and other squash matches).

Agreed. The fact that bettors will even push a line past -500ish, except for total squash matches like you mentioned, is just crazy to me. I mean, WHO is betting these guys at -600, -700, -800 etc?

I've said this before and I'll say it again: It's good that the sport is volatile enough that dogs win as often as they do because they're keeping the lines on faves reasonable. Every time a dog wins its good for sports bettors as a whole.

[MMA_scientist]

Dogs win a LOT, so back on the discussion of value, we can only say then that dogs are holding more value now than ever before.

Agree generally, but unfortunately, the books don't raise the + odds to the same extremes they lower the - odds. I don't think Gus was ever over +600, even though Jones went as low as -1100 at one point. It makes easy to lose a shit load betting steep faves, but a but harder to win a shit load betting the big dogs.

We need peer to peer sports betting.

[SPX]

Looks like +660 was as high as he got. But even at that, you have to say he held a TON of value. I mean, he seriously almost won. In fact, 52% of people on a poll I read the other day thought he should've gotten the decision.

Based on the results of that fight, Gus should've been like +170 or something.

[MMA_scientist]

Looks like +660 was as high as he got. But even at that, you have to say he held a TON of value. I mean, he seriously almost won. In fact, 52% of people on a poll I read the other day thought he should've gotten the decision.

Based on the results of that fight, Gus should've been like +170 or something.

yeah. I don't think it is right to look back with hindsight and say "he had value". Sure, had we known then what we know now, it is obvious Gus was a great bet. But that information was not available, because Gus had never shown TD defense (and offense) like that before.

I have been reading a lot about a pop philosopher/economist/statistician Nicholas Taleb. Even reading the wiki on his theory kind of sheds some light on some of this stuff for me.

Based on the author's criteria:

1. The event is a surprise (to the observer).
2. The event has a major effect.
3. After the first recorded instance of the event, it is rationalized by hindsight, as if it could have been expected; that is, the relevant data were available but unaccounted for in risk mitigation programs. The same is true for the personal perception by individuals.

The theory (the Black Swan theory) is not meant for small relatively common occurrences like an underdog winning in MMA, but I think a lot of it is applicable. It kind of puts me in the "no value unless they won" camp, because of the whole tendency to rationalize these things as if they could have been predicted. The fact is that no one could predicted Gus's wrestling in that fight. So to look at it now and say yeah, he should have been +170 is wrong IMO. Based on what we knew then, we all thought Jones had the wrestling out if the striking wasn't going well. I mean, I thought Jones would throw him down, drop a few elbows, and call it a night. That said, I would love to get +660 on Gus in a rematch, because now we have more info. But if Jones comes out and ragdolls him with his wrestling next time, we will just rationalize that too, "well Jones was surprised last time, didn't train his wrestling much, was injured, etc."

Bottom line: I am more confused than ever.

[SPX]

The fact of the matter is that the line, if true, is supposed to reflect the fighter's actual chances of winning. Regardless of the information available, or unavailable, it's clear that his real-life chances of winning were not congruent with the line.

[Luke]

There's an easy way to tell if "value is overrated" among combat sports bettors. If it is, it will be consistently profitable to bet all the favorites -- this doesn't seem to be the case.

No no no, thats not what I'm saying. Like X said , lines are supposed to reflect a fighters actual chance to win, but when someone says there's "value" in taking Ben Rothwell
at +600 over Cain.....when his actual chance of winning is probably closer to +10000.....there's isnt any value . For there to be value Rothwell would have to win better than 1 in 6 times which it not realistic to think that he could do that. I'm just saying the many many times people say there's value in a line because "if those fighters fought ten times, he'd win 3 of them" when most fights the same guy would probably win 10 out of 10 fights. In the NFL the Pats may beat Buffalo on Sunday, but if they played 10 times the Bills really would win 3 times so there would be value in taking them at +600..........if any of that makes sense

[Mr. IWS]

No no no, thats not what I'm saying. Like X said , lines are supposed to reflect a fighters actual chance to win, but when someone says there's "value" in taking Ben Rothwell
at +600 over Cain.....when his actual chance of winning is probably closer to +10000.....there's isnt any value . For there to be value Rothwell would have to win better than 1 in 6 times which it not realistic to think that he could do that. I'm just saying the many many times people say there's value in a line because "if those fighters fought ten times, he'd win 3 of them" when most fights the same guy would probably win 10 out of 10 fights. In the NFL the Pats may beat Buffalo on Sunday, but if they played 10 times the Bills really would win 3 times so there would be value in taking them at +600..........if any of that makes sense

I agree homie.

I bet on who I think will win, thats it. I dont see how betting "value" in combat sports can be profitable.