Wisdom of the crowd AND favourite longshot bias

[Godders]

Hi all, As the title suggests, this is a system that takes advantage of wisdom of the crowd and the favourite longshot bias across a number of sports. There is a website I use that collates individual predictions for various sporting events and converts into a percentage of each outcome happening. This is where the wisdom of crowd comes in. The favourite longshot bias will be satisfied by only choosing selections at odds on prices. Selections will simply be where the 'crowd' determines there is value in the odds. £100 starting bank. Each selection will be 5% of the bank at the start of each day. Odds taken will be best available from oddschecker at the time of posting. I'll also include the value for each selection.

[Godders]

Selections for tomorrow: NatWest T20 Blast - Northamptonshire (vs Derbyshire) @ 1.73. 27% value Super league - Hull (vs Wakefield) @ 1.14. 10% value Super league - Huddersfield (vs Hull KR) @ 1.25. 13% value Super league - Castleford (vs Widnes) @ 1.17. 11% value MLS - Portland Timbers (vs San Jose) @ 1.57. 10% value

[MPLouis]

As the title suggests, this is a system that takes advantage of wisdom of the crowd and the favourite longshot bias across a number of sports. There is a website I use that collates individual predictions for various sporting events and converts into a percentage of each outcome happening. This is where the wisdom of crowd comes in. The favourite longshot bias will be satisfied by only choosing selections at odds on prices. Selections will simply be where the 'crowd' determines there is value in the odds.

Could you explain your systema little more please Godders? What do you mean by 'the wisdom of the crowd'? What is 'Favourite longshot bias'? Do you mean where favourites are better odds than they should be? Have you used this system before or is this an experiment? :)

[Bebe]

Hi all, As the title suggests, this is a system that takes advantage of wisdom of the crowd and the favourite longshot bias across a number of sports. There is a website I use that collates individual predictions for various sporting events and converts into a percentage of each outcome happening. This is where the wisdom of crowd comes in. The favourite longshot bias will be satisfied by only choosing selections at odds on prices. Selections will simply be where the 'crowd' determines there is value in the odds. £100 starting bank. Each selection will be 5% of the bank at the start of each day. Odds taken will be best available from oddschecker at the time of posting. I'll also include the value for each selection.

Do you have a prototype for you site or it's just an inovation of yours? thanks

[Dylan Lynch]

Hi all, As the title suggests, this is a system that takes advantage of wisdom of the crowd and the favourite longshot bias across a number of sports. There is a website I use that collates individual predictions for various sporting events and converts into a percentage of each outcome happening. This is where the wisdom of crowd comes in. The favourite longshot bias will be satisfied by only choosing selections at odds on prices. Selections will simply be where the 'crowd' determines there is value in the odds. £100 starting bank. Each selection will be 5% of the bank at the start of each day. Odds taken will be best available from oddschecker at the time of posting. I'll also include the value for each selection.

"wisdom of crowd" :) Betting is meant to be a business, a mathematical head ache with formulas, or at least it should be organised logically, not based on wisdom...Though it's considered a luck matter, the winners occur to be the smart guys not the wise ones.... but i like your quote:ok

[Paparazzo]

Godders how do you measure the wisdom and the longshot advantage?

[Godders]

I'll attempt to explain the system a little more. Wisdom of the crowd refers to the collective opinion of a group of individuals, rather than the judgement of just one person. The idea is that using a collective opinion reduces the 'noise' in individual judgements. As an example - I may think team A has an 80% chance of winning against team B. Someone else may think they only have a 60% chance. Who's right? Probably neither, but if you take the judgements of many people, the wisdom of the crowd theory says that the average of all the guesses will be very close to the actual chance of team A beating team B (in a particular wisdom of the crowd experiment, members of the public were asked to guess the weight of a slaughtered ox. The average of all the guesses was apparently within 1% of the actual weight). Not sure if I can post external websites, but basically it's has various competitions where you predict the outcome of an event (and you get points based on how accurate your predictions are). Once you've entered your predictions, you can then see the averages of everyone elses predictions. This is how I'm measuring the wisdom of the crowd. The favourite longshot bias is a psychological phenomenon where the chance of a big favourite coming in is underestimated, and the chance of long odds outsiders overestimated. This effect becomes more pronounced the shorter the odds. The idea of this system then is to try and take advantage of both of these by using a crowd's collective opinion of a particular outcome, and the selections will only be made on short odds favourites where the crowd's estimate of the chance of the bet winning represents value over the bookies odds.

[Godders]

Selections for tomorrow: NatWest T20 Blast - Northamptonshire (vs Derbyshire) @ 1.73. 27% value Super league - Hull (vs Wakefield) @ 1.14. 10% value Super league - Huddersfield (vs Hull KR) @ 1.25. 13% value Super league - Castleford (vs Widnes) @ 1.17. 11% value MLS - Portland Timbers (vs San Jose) @ 1.57. 10% value

A great start to the system with all 5 coming in. Bank now £109.30, so next bets will be £5.46 Bets made: 5 Bets won: 5 SR: 100% Yield: 37.2%

[Paparazzo]

Not sure if I can post external websites, but basically it's has various competitions where you predict the outcome of an event (and you get points based on how accurate your predictions are). Once you've entered your predictions, you can then see the averages of everyone elses predictions. This is how I'm measuring the wisdom of the crowd.

Let me see if I've got it: You take for instance 20 predictions of a match, Let's say that 18 out of 20 forecast a home win. Thus 18/20=90% or a 1,11 is your fair odd (or wisdom of the crowd odd). The market says 1,30 (still a short odd in the range of the longshoot bias) then you expect a 13% advantage... Is that correct?

[MPLouis]

I'll attempt to explain the system a little more. Wisdom of the crowd refers to the collective opinion of a group of individuals, rather than the judgement of just one person. The idea is that using a collective opinion reduces the 'noise' in individual judgements. As an example - I may think team A has an 80% chance of winning against team B. Someone else may think they only have a 60% chance. Who's right? Probably neither, but if you take the judgements of many people, the wisdom of the crowd theory says that the average of all the guesses will be very close to the actual chance of team A beating team B (in a particular wisdom of the crowd experiment, members of the public were asked to guess the weight of a slaughtered ox. The average of all the guesses was apparently within 1% of the actual weight). Not sure if I can post external websites, but basically it's has various competitions where you predict the outcome of an event (and you get points based on how accurate your predictions are). Once you've entered your predictions, you can then see the averages of everyone elses predictions. This is how I'm measuring the wisdom of the crowd. The favourite longshot bias is a psychological phenomenon where the chance of a big favourite coming in is underestimated, and the chance of long odds outsiders overestimated. This effect becomes more pronounced the shorter the odds. The idea of this system then is to try and take advantage of both of these by using a crowd's collective opinion of a particular outcome, and the selections will only be made on short odds favourites where the crowd's estimate of the chance of the bet winning represents value over the bookies odds.

Thanks for the explanation Godders :)

[Godders]

Let me see if I've got it: You take for instance 20 predictions of a match, Let's say that 18 out of 20 forecast a home win. Thus 18/20=90% or a 1,11 is your fair odd (or wisdom of the crowd odd). The market says 1,30 (still a short odd in the range of the longshoot bias) then you expect a 13% advantage... Is that correct?

That's exactly right

[Godders]

Picks for Friday T20 Blast - Essex (vs Middlesex) @ 1.62. 30% value. T20 Blast - Birmingham (vs Worcestershire) @ 1.8. 12% value T20 Blast - Nottinghamshire (vs Derbyshire) @ 1.44. 18% value T20 Blast - Kent (vs Somerset) @ 1.67. 36% value T20 Blast - Sussex (vs Glamorgan) @ 1.67. 11% value NRL - Raiders (vs Knights) @ 1.44. 21% value Super League - St Helens (vs Huddersfield) @ 1.57. 16% value

[thecurlyone1]

Nice idea. Good luck

[Godders]

After yesterday's results: Bank: £129.06 Total bets: 11 Bets won: 10 SR: 90.91% Total staked: £57.60 Total profit: £29.06 Yield: 50.45%

[Godders]

Picks for Sunday NRL - Warriors (vs Storm) @ 1.57. 16% value. Super League - Warrington (vs Wakefield) @ 1.29. 18% value T20 Blast - Kent (vs Gloucestershire) @ 1.62. 19% value T20 Blast - Northamptonshire (vs Leicestershire) @ 1.75. 32% value. Bets £6.45

[thecurlyone1]

Godders are you using another forums app to help you pick these ?

[MPLouis]

Godders are you using another forums app to help you pick these ?

I think Godders explained his sources in post #7 above

[WWWYKI!]

Interesting system Godders,is there any chance you could pm me the site you're using?Currently reading the Wisdom of crowds so this has caught my eye.I tried to pm you but it's not working right now for some reason.

[Godders]

Stats after Sunday's picks as follows: Bank: £143.44 Total bets: 15 Bets won: 14 SR: 93.33% Total staked: £83.40 Total profit: £43.44 Yield: 52.09%

[harry_rag]

Wisdom of the crowd refers to the collective opinion of a group of individuals' date=' rather than the judgement of just one person. The idea is that using a collective opinion reduces the 'noise' in individual judgements. As an example - I may think team A has an 80% chance of winning against team B. Someone else may think they only have a 60% chance. Who's right? Probably neither, but if you take the judgements of many people, the wisdom of the crowd theory says that the average of all the guesses will be very close to the actual chance of team A beating team B (in a particular wisdom of the crowd experiment, members of the public were asked to guess the weight of a slaughtered ox. The average of all the guesses was apparently within 1% of the actual weight).[/quote'] Hard to quibble with results thus far but, just to be clear, I believe you've clarified that you're basing your edge on % chance represented by actual odds v % of people predicting the outcome in question, e.g. 18/20 think something will happen equates to a 90% chance? Based on the examples given above, a more accurate measure would be a large number of people estimating the % chance of something happening then taking the average as being the true odds? (So team B would have a 70% chance of winning based on the 2 man "crowd" in your example.) Have I got that right, i.e. it's not based on the average of the crowd's estimate of the true probability, just their predicted outcome?

[Godders]

Hard to quibble with results thus far but, just to be clear, I believe you've clarified that you're basing your edge on % chance represented by actual odds v % of people predicting the outcome in question, e.g. 18/20 think something will happen equates to a 90% chance? Based on the examples given above, a more accurate measure would be a large number of people estimating the % chance of something happening then taking the average as being the true odds? (So team B would have a 70% chance of winning based on the 2 man "crowd" in your example.) Have I got that right, i.e. it's not based on the average of the crowd's estimate of the true probability, just their predicted outcome?

I'm not expecting these results to continue!! I don't understand your question though?

[Godders]

One for tomorrow- Super League - St Helens (vs Warrington) @ 1.91. 17% value. Bet is £7.17

[christos83]

Nice point of view but how you choose the games and from where you get your crowd opinion? Do you give some value based on the number of the crowd ? I mean do you rate the same a game that has 8/10 and the same a game that has 79/100 ? Can be used for football ?

[rockymog]

Hi Godders, how do you work out the % advantage, I understand the first bit about the wisdom versus actual odds, how do you arrive at the % advantage from there. thanks rockymog

[allen29]

Hey I was thinking about using the wisdom of the crowds principle as well for sports betting. However, I know only a few (2-3) sites that publish predictions of football events. What is the site that you use that you say aggregates these predictions?

[Godders]

Picks for tomorrow (Friday) NRL- Storm (vs Panthers) @ 1.44. 14% value. Super League - Leeds (vs Salford) @ 1.17. 12% value. T20 - Kent (vs Glamorgan) @ 1.67. 36% value T20 - Northamptonshire (vs Yorkshire) @ 1.73. 35% value T20 - Birmingham (vs Lancashire) @ 1.73. 39% value T20 - Surrey (vs Somerset) @ 1.75. 24% value T20 - Worcestershire (vs Derbyshire) @ 1.73. 33% value T20 - Essex (vs Middlesex) @ 1.67. 33% value T20 - Sussex (vs Hampshire) @ 1.8. 32% value AFL - North Melbourne (vs Essendon) @ 1.33. 19% value Bet will depend on outcome of the St Helens bet in play at the moment. I'll post the bet size later this evening

[harry_rag]

I'm not expecting these results to continue!! I don't understand your question though?

Sorry, could have been more concise! I suppose I'm saying there's 2 ways you could use the "wisdom of the crowd". I think what you're doing is based on asking people what they think the outcome would be, so if 18 people out of 20 think a team will win then that equates to a 90% chance. The other way would be to ask those 20 people to price up the event and average their opinions to get the "true" odds, e.g they might estimate that the team has anywhere bewteen a 60% and 80% chance of winning with an average of 70%, which you'd then compare to the odds on offer. I think the latter is closer to the principle of the theory and seems more relevant to the question "what are the true odds of this happening" but it would be harder to do and hey, what does it matter when the results are positive doing it the other way! Good luck, will carry on following with interest.

[Godders]

Yes Harry, I agree with you. That's primarily the reason why I'm tracking the 'value' - my thinking is that the greater the confidence in a particular result (ie the shorter the true odds would likely be if that were the question asked), the more people would pick that result, meaning a greater value percentage. Will be interesting to see after a large number of bets whether there's a difference in the yield across different value sizes. Heading towards the 2nd loss at the moment with St Helens trailing Warrington 12-8 at the moment.

[Godders]

Christos & Rocky - I've explained how I select picks and get the values in previous posts. Let me know if you're still having trouble understanding though. Allen- don't think I can post to external sites, but Google sports predictions games and you can get a myriad of hits

[Godders]

A last minute try gave St Helens the win, so tomorrow's bets will all be for £7.49