If this is your first visit, be sure to
check out the FAQ by clicking the
link above. You may have to register
before you can post: click the register link above to proceed. To start viewing messages,
select the forum that you want to visit from the selection below.
2011 Off topic thread(basketball,movies,etc whatever)
I don't really understand what a "model" is in betting terms.
How does it work?
A model is a method for taking the available data and turning it into an estimate of win probability (or point spread, etc).
The challenge is to figure out which stats (and in what combinations) are the most predictive. Let's take football, for example. An extremely crude model to estimate point spread might be to take the average number of points each team scored in prior games this season, and look at the difference in the two.
Of course, we would find that this is a horrible model. So you might make it more complicated in stages: start by also taking into account how many points each team's defense allowed. Then you'll want to factor in "strength-of-schedule", so you look at the points scored and points allowed for all their opponents.
It turns out that people have figured out that yards are more predictive of future points than points are, and "efficiency stats" that look at yards per play instead of yards per game are better still. So a lot of people come up with formulas that are based on yards per play for running and passing.
A model doesn't even have to just be a single formula. For example, some people try to model baseball outcomes by basically putting in all the stats for each team's players and having a computer program play out 100,000 games or so to estimate the outcome.
I get the impression this isn't really being done for MMA. . .
Yeah, we've discussed this before a bit. I don't think it's really possible to make a comprehensive model for MMA. As far as I know, no one's done it with boxing and betting on that has been around far longer.
I've been working on a model to estimate probability of an MMA fight going the distance (given the initial odds and the current time of the fight).
What I am saying is that no one has been able to do it with boxing and MMA is even harder, given the addition of the clinch game and the ground game. Also, most fighters fight even less.
You could do it in mma but you'd have to make up stats for punching power,TD ,striking, BJJ,wrestling etc.....all the stats are already done in basketball ,football and baseball so all you have to do is plug them into your formula and see the results
I don't see a huge problem in estimating the stats. I mean, put Carwin at a 9 on the punching power scale and someone like Struve at a 5, etc. It wouldn't be perfect, but I'd be interested to see how it would work out.
I don't know how people think they can beat the NFL/MLB/NBA.
Baseball and basketball are hard but football is much easier to bet than MMA. In MMA if Guillard decides to do something stupid you lose on a big fav, in football if one guy isnt ready to play you have 21 other guys to depend on to try to win....plus in all other sports stat after stat is available at your finger tips.
And SPX is right my sig does not reflect all my bets in football,I've only posted plays 2 units and under on here.....anything bigger I have not.
I don't see a huge problem in estimating the stats. I mean, put Carwin at a 9 on the punching power scale and someone like Struve at a 5, etc. It wouldn't be perfect, but I'd be interested to see how it would work out.
Then get to work, but after you make a rating system you'll have to make a formula to figure out which bets to make or too pass on.
Unless I'm the one with the super computer program I don't like the idea anyway. It's nice that there is a sport that's about good ol' fashioned analysis and intuition.
It's nice that there is a sport that's about good ol' fashioned analysis and intuition.
Yep. If MMA could be easily modeled, most of the value would probably get sucked out of the market quickly and it would be as hard to bet as those other sports.
Comment