Probabilistic Prediction of Football Matches - A Discussion

[Rushian]

Re: Probabilistic Prediction of Football Matches - A Discussion lunatism, I'm trying to understand and I may have got this entirely the wrong way, so please excuse me, but a few of the points you are making have me confused:

if somebody provides the percentages for likelihood of h/d/a then i would presume that they aren't doing so with the sole intention of taking longshots. But if that were the case' date= we'd soon notice and suggest they don't take part in this, because we're looking at consistent week by week bla bla bla?! If a system is always suggesting short odds, then it's likely that unless very very good, they wont be successful in the long term....so what? Isn't the point of this to see what's good and what's not? :D

The way I see it, the estimated probabilities have nothing to do with the bets placed unless you know the odds. For example, if according to a model the probability of a home win is 50%, then a bet will be placed only if the odds on offer are higher than 2.00. If the odds available are only 1.91, then a bet would not be placed (since your guess of the probability is 50% and hence in the long run, that bet would return a negative yield). Furthermore, when you say that you wouldn't want someone whose system highlights the lonsghots, I am not sure that I follow: maybe someone's system produces average strike odds of around 3.00 with a relatively low strike rate (say 35%) whereas someone esle has a high strike rate but with much shorter odds. I don't understand why you suggest that the latter is desirbale ("consistent" as you put it) but the former is not.

Comparisons should just be made on what the system users provide based on this season' date= and then we'll get a clearer picture of how they perform as the season progresses. I was always under the impression you wanted every selection from every week - and so number of games shouldn't come into it too much. Obviously if someone only has results from start of jan, then that's their choice to include from then, or we could say a certain date is the cutoff point, but shouldn't be too much hassle. We could even start from scratch and assign a 0% yield for each to start with - so a level playing field (to ensure nobody is cheating...which i'm sure would'nt happen anyway)

When I mentioned the number of bets, I was suggesting that it would not be logical to compare two systems, one which produced a 10% yield based on 10 bets (i.e. the bets that were picked after comparing the estimated probability with the odds on offer, as described previously) and another which resulted in 3% yield based on 200 bets. Maybe both systems estimated probabilities for 700 matches in total, but the odds on offer provided value for 10 and 200 bets respectively (hence the fewer bets).

all that has to be done is that we use the b365 from football-data - that is neutral to this and would immediately provide clarity i think. that is what i have used (since i wouldn't have had the odds on file anyway). And that's what i'd recommend everyone else do for the purpose of this experiment. Choice of bookie doesn't affect strike rate' date= but we do need to look at odds and yield, otherwise how else could you differentiate between short odd favourites and the long shots?!

Again, using odds from one bookmaker may highlight a certain number of bets while using the best available odds accross the board would result in more bets, for the same set of estimated probabilities. This would mean that both strike rate and yield would be dependant on the choice of bookmaker. I agree though that, should people decide that a system's league is needed, the comparison could be based on the odds offered by a commonly agreed bookmaker. Finally, something that was not mentioned in my previous post, the staking plan would also result in different yield, even if the estimated probabilities are the same, i.e. fixed level staking would lead to different results to something like the Kelly criterion etc... Finally, as for my system, I haven't updated for some time now as it is still under development ;). At the last update of data, the overall yield was -3.4%, which was broken down as follows: Premiership -35.9% :wall, Championship -3.1%, League One +4.0% and League Two +14.4%. I don't have figures for the Liverpool - Everton match, but if I find some spare time over the weekend, I'll try to update the data and get some probabilities for the next set of matches. Maybe I got confused, maybe not. Just ignore me if I am completely off the mark :rollin Rushian

[lunatism]

Re: Probabilistic Prediction of Football Matches - A Discussion TOMORROW is Saturday guys! Won't you be submitting any percentages? So far only mines and Grex's are in the spreadsheet, still waiting on Rushian and . HooLooVoo Rushian...I appear to be talkig jibberish, and the reply box on my screen is too small to fit all of your response in at the same time :unsure

[Grex]

Re: Probabilistic Prediction of Football Matches - A Discussion C'mon, Rushian and HooLooVoo...... Loon has put a lot of hard work into this, so the least we can do is to avoid delays to getting the ball rolling.

[ozangumus]

Re: Probabilistic Prediction of Football Matches - A Discussion Everything depends on probabilities in soccer-betting. Any person betting on any team with any odd should think that this bet is a valuebet. Maybe he does not know but he is making an estimation in his mind and compares his calculation result with odd, and he decides betting to that situation or not. Every outcome have a probability in soccer matches even Rapid Wien can win Salzburg with 0-7 at away:) So important thing is if odd * probability passes 100 or not. If you only bet valuebets you will probably win in long term there is not another way for winning in betting, you have to make an estimation by using inputs (past stats, special cases, injuries ext) and compare your result with odds announced by bookies. Please examine my page to make easier your estimation job, probs there are estimated by only analsysing results of past matches, you can add effects of other special situations to make probabilities more optimal/reliable.

[Rushian]

Re: Probabilistic Prediction of Football Matches - A Discussion Just a quick note: I'll try to find some time to update my system data, hopefully tonight or early tomorrow at the latest and send my estimated probabilities. (Unfortunately my work is such that most days, I can only post at night...) :sad

[Grex]

Re: Probabilistic Prediction of Football Matches - A Discussion

Please examine my page to make easier your estimation job' date= probs there are estimated by only analsysing results of past matches, you can add effects of other special situations to make probabilities more optimal/reliable.

Thanks, Ozangumus, for your input..... looks like we´re all wasting our time, lads ;)

[Rushian]

[Hooloovoo]

Re: Probabilistic Prediction of Football Matches - A Discussion Hi guys, sorry for the delay. Have been ill for a few days :puke:cry Will post the percentages before KO tomorrow :ok

[lunatism]

Re: Probabilistic Prediction of Football Matches - A Discussion

Hi guys, sorry for the delay. Have been ill for a few days :puke:cry Will post the percentages before KO tomorrow :ok

sorry to hear that mate - will we get the percentages in time so that I can post the weighted averages?

[Hooloovoo]

Re: Probabilistic Prediction of Football Matches - A Discussion Will try to post them tonight :unsure

[Hooloovoo]

Re: Probabilistic Prediction of Football Matches - A Discussion Ok no fancy table but nevertheless, here are the probs: Match, H%, D%, A% Birmingham - Man City, 37, 30, 33 Bolton - Arsenal, 19, 20, 61 Derby - Fulham, 34, 32, 34 Portsmouth - Wigan, 60, 24, 16 Reading - Blackburn, 32, 28, 40 Sunderland - West ham, 34, 28, 38 Man Utd - Aston Villa, 64, 22, 14 Chelsea - Middlesbrough, 72, 20, 8 Tottenham - Newcastle, 61, 24, 15 Liverpool - Everton, 52, 26, 22 :ok

[Grex]

[lunatism]

[Rushian]

Re: Probabilistic Prediction of Football Matches - A Discussion Thanks for the update lunatism. Good work :ok Regarding the use of bet365 odds and the fink tank data, both good ideas (and I have to correct you: the fink tank data wasn't my idea :unsure) Interesting figures: I am wondering though, in the same sense that we are willing to back a team which represents value according to our probabilities, are we willing to use these figures to offer odds on betting exchanges? For example, I would be very reluctant to offer odds of 50.0 on a Derby-Fulham draw, as lunatism figures would suggest (although who am I to judge, following lunatism's succesful systems :clap) Overall, this may be of interest to follow :hope

[Grex]

[Grex]

Re: Probabilistic Prediction of Football Matches - A Discussion By the way, there´s one thing I can´t get my head round.... for some of the above matches, the Draw probability is less than both the Home and Away probabilities. Take the Birmingham-Man City predictions for example..... Man City are (marginally) the favourites 37%. Surely the next most probable result should be a draw, and yet a Birmingham win (34%) is predicted as more probable than the Draw (29%). I appreciate that it depends on the specifics of the system, but "logic" suggests it should be otherwise.... perhaps I´m missing something (or, at least, my brain is) :wall

[muppet77]

Re: Probabilistic Prediction of Football Matches - A Discussion this is a great thread people - keep it up. what about measuring the opinion of the masses? i.e. betfair prices (which for the prem on a sat will be very close to a 100% book)

[Hooloovoo]

Re: Probabilistic Prediction of Football Matches - A Discussion

By the way, there´s one thing I can´t get my head round.... for some of the above matches, the Draw probability is less than both the Home and Away probabilities. Take the Birmingham-Man City predictions for example..... Man City are (marginally) the favourites 37%. Surely the next most probable result should be a draw, and yet a Birmingham win (34%) is predicted as more probable than the Draw (29%). I appreciate that it depends on the specifics of the system, but "logic" suggests it should be otherwise.... perhaps I´m missing something (or, at least, my brain is) :wall

I've noticed the same strange pattern. As you say, if one team is the favourite to win the match, the next most probable result would logically be the draw. This is certainly due to the specifics of the system and this is something worth looking further into when we get some real results coming in. My system threw up a couple of these oddities as well. Statistically, two teams could have a mutual history of a very low draw-percentage, but that would be more due to the small-sample nature of mutual matches, almost never more than 15-20 matches, which is way too small when it comes to any statistical certainty. If two teams met a gazillions times there would be the highest percentage for the best team to win, THEN the draw, THEN the worst team to win. I can say for myself that i have some difficulties with my system as it throws up a rating for each team, and i obviously can't translate these ratings into 1-X-2 precentages correctly. Oh well, time will tell By the way, feeling much better today :)

[lunatism]

Re: Probabilistic Prediction of Football Matches - A Discussion

Thanks for the update lunatism. Good work :ok Regarding the use of bet365 odds and the fink tank data, both good ideas (and I have to correct you: the fink tank data wasn't my idea :unsure) Interesting figures: I am wondering though, in the same sense that we are willing to back a team which represents value according to our probabilities, are we willing to use these figures to offer odds on betting exchanges? For example, I would be very reluctant to offer odds of 50.0 on a Derby-Fulham draw, as lunatism figures would suggest (although who am I to judge, following lunatism's succesful systems :clap) Overall, this may be of interest to follow :hope

don't take the Derby-Fulham odds too seriously - that is a system anomaliy and happens once in a month or so. But since i let my system generate my percentages for everything automatically, it's a matter i don't fancy resolving at this moment in time. And then way they calculate, it would mean changing everything most likely.

[KevinTHFC]

Re: Probabilistic Prediction of Football Matches - A Discussion Nice thread guys.

By the way, there´s one thing I can´t get my head round.... for some of the above matches, the Draw probability is less than both the Home and Away probabilities. Take the Birmingham-Man City predictions for example..... Man City are (marginally) the favourites 37%. Surely the next most probable result should be a draw, and yet a Birmingham win (34%) is predicted as more probable than the Draw (29%). I appreciate that it depends on the specifics of the system, but "logic" suggests it should be otherwise.... perhaps I´m missing something (or, at least, my brain is) :wall

How about this? Think about games where there is a strong favourite. If they don't win than logically the most likely result would be a draw. This can be seen by looking at the odds and predictions for Man U, Arsenal, Chelsea and Spurs. They are all over 50% and are odds on to win. The next favoured result is a draw. In your example of Man City and Birmingham, there is not much to choose between the teams and probably the first to score will go onto win the match and that is why the odds for either team to win is lower than the draw. Might be rubbish but sort of makes sense to me;) Kev

[lunatism]

Re: Probabilistic Prediction of Football Matches - A Discussion

this is a great thread people - keep it up. what about measuring the opinion of the masses? i.e. betfair prices (which for the prem on a sat will be very close to a 100% book)

good point again - maybe next week's "version" will not only have fink tank and bet365 a part of it, but the betfair odds too :tongue2 ...and a point to note - perhaps as people see how this goes, they might want to "join" in. Let's say that the cutoff for "joining" is TUESDAY, people already in it can post percentages whenever they like (as long as not too much later than Friday midnight). This will mean i have time to adjust the spreadsheet. Agreed?

[Grex]

Re: Probabilistic Prediction of Football Matches - A Discussion

Statistically, two teams could have a mutual history of a very low draw-percentage, but that would be more due to the small-sample nature of mutual matches, almost never more than 15-20 matches, which is way too small when it comes to any statistical certainty. If two teams met a gazillions times there would be the highest percentage for the best team to win, THEN the draw, THEN the worst team to win. I can say for myself that i have some difficulties with my system as it throws up a rating for each team, and i obviously can't translate these ratings into 1-X-2 precentages correctly. Oh well, time will tell By the way, feeling much better today :)

Totally agree, HLV :ok. That´s one of the reasons why I don´t take into account head-to-head data in my own system. Despite the success some contributors have using this data on other threads, the sample size is usually very small. Anyway, how can you compare a particular team´s capability in 2000 with that in 2008.... some decline (for example Sunderland or Leeds), others advance eg. Reading ;). PS. Glad you´re feeling better, HLV

[Rushian]

Re: Probabilistic Prediction of Football Matches - A Discussion With respect to the daw probability of a given match (say the Birmingham - Man City match), if they percentages are not too far away from each other, then I don't think it's too uncommon. You can also see this in the bet365 odds for the match, which, assuming that the overround is spread evenly, the value the probability of a draw to be lower than any other result (Odds for H/D/A = 2.4/3.3/3.0). However, as KevinTHFC points out, I would expect it not to happen when you have a clear favourite (i.e. you won't see bookies offering odds for H/D/A = 1.3/12.0/5.0).

[lunatism]

Re: Probabilistic Prediction of Football Matches - A Discussion

Nice thread guys. How about this? Think about games where there is a strong favourite. If they don't win than logically the most likely result would be a draw. This can be seen by looking at the odds and predictions for Man U, Arsenal, Chelsea and Spurs. They are all over 50% and are odds on to win. The next favoured result is a draw. In your example of Man City and Birmingham, there is not much to choose between the teams and probably the first to score will go onto win the match and that is why the odds for either team to win is lower than the draw. Might be rubbish but sort of makes sense to me;) Kev

nope - wouldn't call this rubbish, infact - think you've hit the nail on the head with this one. Let's hope that the games kind of follow that pattern ;)

[Grex]

[Grex]

Re: Probabilistic Prediction of Football Matches - A Discussion Not quite sure how to interpret today´s Premiership results.... none of the so-called "value bets" came in :eyes

As usual, the value on the hot favourites (ManU and Chelsea) is cr*p... surprisingly, Arsenal´s value is reasonable. Based on the above figures, if I were a dedicated value bettor, I would go for the following: Man City (A), Bolton (X) [unfortunately I´ve already put money on Arsenal to win :eyes], Derby (H), Portsmouth (X), Blackburn (A), West Ham (A), Liverpool (X), Newcastle (A) :loon

Good job I didn´t put any money on them ;)

[Grex]

[Rushian]

Re: Probabilistic Prediction of Football Matches - A Discussion The thing is, when coupled with the odds on offer, probability systems tend to suggest value in the outsiders (more often than not). This would seem to contradict what is sometimes called as the "longshot-favourite" bias, which, as I understand it, means that the bookies overround isn't spread equally. According to the "longshot-favourite" bias, rather than having the same profit margin accorss the outcomes, bookies (again, if my understanding is right, and do correct me if I'm wrong here) tend to reduce the price of the outsider to cover themselves. If however, others have noticed this too (i.e. that these probabilistic predictions tend to suggest value in the outsiders), how would one go about to improve the system ... unless of course it's profitable with a low strike rate and high average strike odds... Any views?

[Grex]

Re: Probabilistic Prediction of Football Matches - A Discussion I relly don´t understand bookies´ odds and "over-round" strategy.... so, I won´t make an idiot of myself by commenting on that issue. However, when I have back-tested my own "value odds" systems, I find that there is a specific range of % value where returns leap up... this range is normally between 95% and 115%.... anything above 120% and returns decline.

[Rushian]

Re: Probabilistic Prediction of Football Matches - A Discussion

I relly don´t understand bookies´ odds and "over-round" strategy.... so, I won´t make an idiot of myself by commenting on that issue. However, when I have back-tested my own "value odds" systems, I find that there is a specific range of % value where returns leap up... this range is normally between 95% and 115%.... anything above 120% and returns decline.

That's a quite interesting perspective! I haven't tried imposing an upper limit on the perceived "value" of a bet but, having just played with my spreadsheet, it does indeed improve the results! Of course, more testing is needed but it's a start... Thanks Grex :ok