Our Handicappers week eight free NFL picks are now ready. We've also posted our Expectation Value Probability Distribution for week eight.
Thanks,
- Happy
Thursday, October 29, 2009
Friday, October 23, 2009
New Page Format
We reformatted our week seven free NFL picks page. We're thinking of doing our entire NFL Oddsmakers Free Picks site with the new format. Currently most of the pages (look at past week's picks) are still do the old format. Which do you prefer?
Other ideas for how to make our site more attractive / user-friendly?
Thanks,
Happy
Other ideas for how to make our site more attractive / user-friendly?
Thanks,
Happy
Monday, October 19, 2009
The Advantage of the Over Under Bet
Before I created my NFL Predictive Model for forecasting the outcomes of NFL games, I always thought betting the Over/Under was just not all that interesting. What really matters is who is going to win the game, right? Coming off NFL week six where my over/under bet units were 5/5 (and my line bet units were only 3/6) I thought it might be a good time to sing the praises of the over-under bet.
The over bet is one bet that can be put in the bag before the game is over. In fact, during the Saints vs Giants game I’d already won one of my best bets before half-time! The under can start looking pretty sure by midway through the fourth quarter. Recently examples include under in the Bills vs Browns and under in the Chiefs vs Redskins (both games in which our model best bet the under).
By contrast, the soonest you can count your un-hatched chickens in a spread bet is at the end of regulation when you’ve taken a huge underdog (we notched the Bills vs Jets as a victory at the begging of overtime because we had 9.5 points).
On a more philosophical note, the OU is also more a reflection of the character of the game. We were glad our model took the over in the NO vs NYG game because we thought the game pitted great offenses against decent defenses and that it would be a shootout. As far as who would win, who knew? The model has NO by 2.7 points but with zero confidence.
On a more practical note, adding the Over Under to our offerings has increased our bet-worthy events from 3-5 per week to 6 – 10 per week. The increase in bet-worthy events narrows the probability distribution of our expectation values, increasing our odds of a winning week. Note that even though our individual weeks range from 63% - 70% odds of winning, YTD we have a 76.5% chance of winning (and each event is only 55% - 60%). The more events we can add, the faster we narrow the distribution around our 110% expectation value – making winning a more certain long-term outcome.
The over bet is one bet that can be put in the bag before the game is over. In fact, during the Saints vs Giants game I’d already won one of my best bets before half-time! The under can start looking pretty sure by midway through the fourth quarter. Recently examples include under in the Bills vs Browns and under in the Chiefs vs Redskins (both games in which our model best bet the under).
By contrast, the soonest you can count your un-hatched chickens in a spread bet is at the end of regulation when you’ve taken a huge underdog (we notched the Bills vs Jets as a victory at the begging of overtime because we had 9.5 points).
On a more philosophical note, the OU is also more a reflection of the character of the game. We were glad our model took the over in the NO vs NYG game because we thought the game pitted great offenses against decent defenses and that it would be a shootout. As far as who would win, who knew? The model has NO by 2.7 points but with zero confidence.
On a more practical note, adding the Over Under to our offerings has increased our bet-worthy events from 3-5 per week to 6 – 10 per week. The increase in bet-worthy events narrows the probability distribution of our expectation values, increasing our odds of a winning week. Note that even though our individual weeks range from 63% - 70% odds of winning, YTD we have a 76.5% chance of winning (and each event is only 55% - 60%). The more events we can add, the faster we narrow the distribution around our 110% expectation value – making winning a more certain long-term outcome.
Friday, October 16, 2009
Converting from a money-line to a point-spread
I've seen many people attempt to answer the question, "How do I convert a point-spread to a money-line?" (or conversely, "How do I convert a money-line to a point-spread?" The answers I've seen given have been unequivocally bad; but that’s because the correct answer is not simple, and people like simple answers.
To understand that a simple plug-and-play formula won’t do the trick, just check out the lines at your favorite sportsbook. I checked betus.com at 8:00 PM on the 16th of October and found the following lines:
St. Louis +9.5 -110 (points) or +400 (moneyline)
Jacksonville -9.5 -110 (points) or -500 (moneyline)
Buffalo +9.5 -110 (points) or +350 (moneyline)
NY Jets -9.5 -110 (points) or -450 (moneyline)
From this it is clear that there is more involved than simply knowing the point-spread to determine the moneyline (or more than just knowing the odds to determine the pointspread).
The function that is used to convert from a pointspread to a moneyline is the normal distribution function. This function must be integrated from negative infinity to the point-spread to determine the probability of an outcome. In all cases the mean is zero. If the pointspread is -9.5 that is the value used for x. The value of the integral from –∞ to -9.5 is also a function of the standard deviation; and this is what varies from game to game.
We can back-calculate the standard deviation the casino used by iterating (such as tools goal-seek from MS Excel). In the game where Jacksonville is -500 you have to lay $500 to win $100. But remember the 10% juice. Without it you’d only have to lay $450 to win $100. This means that Jacksonville should win 4.5 times more often than it loses. Neglecting ties this means that Jacksonville should win 4.5 / 5.5 of the contests or 81.8%. Taking the inverse of the integrated normal distribution function (=norminv(0.818, 0, guess) in MS Excel) you can change your guess (using tools goal-seek) to get a point-spread of 9.5. In the case of the Jacksonville vs St. Louis game you’ll find that the sportsbook used a standard deviation of 10.46. You can now verify this by using the same standard deviation but a negative spread; you’ll find St. Louis has a 18.2% chance of winning. One divided by 18.2% = 5.5. This means that the casino without juice would give you $450 winnings for every $100 bet (you get $550 off $100 18.2% of the time). Since they take their juice off your winnings, the line should be 90% of 450 or 405. Note that they’re actually offering +400. I have a feeling rounding errors tend to favor the house!
A similar exercise will reveal that the standard deviation used in the Bills vs Jets game is 11.2. I would have simplified to assume that the standard deviation tracks the over/under, but that is clearly not the case here. The casino most likely has team specific standard deviations (due to a more complex model than mine).
This brings up a separate, but related issue. If the casino’s model is more complex than mine, how can I expect to win in the long run? The key is that I don’t have to be smarter than the sportsbook. The book is balancing their desire for profit (holding their line) with risk mitigation (splitting the betting public evenly). If I have a decent model (I’m right greater than 52.4% of the time) I can win in the long run. Currently we’re running 58% season to date.
To understand that a simple plug-and-play formula won’t do the trick, just check out the lines at your favorite sportsbook. I checked betus.com at 8:00 PM on the 16th of October and found the following lines:
St. Louis +9.5 -110 (points) or +400 (moneyline)
Jacksonville -9.5 -110 (points) or -500 (moneyline)
Buffalo +9.5 -110 (points) or +350 (moneyline)
NY Jets -9.5 -110 (points) or -450 (moneyline)
From this it is clear that there is more involved than simply knowing the point-spread to determine the moneyline (or more than just knowing the odds to determine the pointspread).
The function that is used to convert from a pointspread to a moneyline is the normal distribution function. This function must be integrated from negative infinity to the point-spread to determine the probability of an outcome. In all cases the mean is zero. If the pointspread is -9.5 that is the value used for x. The value of the integral from –∞ to -9.5 is also a function of the standard deviation; and this is what varies from game to game.
We can back-calculate the standard deviation the casino used by iterating (such as tools goal-seek from MS Excel). In the game where Jacksonville is -500 you have to lay $500 to win $100. But remember the 10% juice. Without it you’d only have to lay $450 to win $100. This means that Jacksonville should win 4.5 times more often than it loses. Neglecting ties this means that Jacksonville should win 4.5 / 5.5 of the contests or 81.8%. Taking the inverse of the integrated normal distribution function (=norminv(0.818, 0, guess) in MS Excel) you can change your guess (using tools goal-seek) to get a point-spread of 9.5. In the case of the Jacksonville vs St. Louis game you’ll find that the sportsbook used a standard deviation of 10.46. You can now verify this by using the same standard deviation but a negative spread; you’ll find St. Louis has a 18.2% chance of winning. One divided by 18.2% = 5.5. This means that the casino without juice would give you $450 winnings for every $100 bet (you get $550 off $100 18.2% of the time). Since they take their juice off your winnings, the line should be 90% of 450 or 405. Note that they’re actually offering +400. I have a feeling rounding errors tend to favor the house!
A similar exercise will reveal that the standard deviation used in the Bills vs Jets game is 11.2. I would have simplified to assume that the standard deviation tracks the over/under, but that is clearly not the case here. The casino most likely has team specific standard deviations (due to a more complex model than mine).
This brings up a separate, but related issue. If the casino’s model is more complex than mine, how can I expect to win in the long run? The key is that I don’t have to be smarter than the sportsbook. The book is balancing their desire for profit (holding their line) with risk mitigation (splitting the betting public evenly). If I have a decent model (I’m right greater than 52.4% of the time) I can win in the long run. Currently we’re running 58% season to date.
Wednesday, October 14, 2009
Happy's Picks for NFL Week Six - Statistical Model Based and Expectation Values Explained
We've updated our NFL Free Pick Results vs Expectation Values to include Week five results.
We’ve also now posted our NFL week six free picks which include five (5) bet worthy Week 6 NFL lines and three (3) bet worthy NFL Week Six over unders. There are three best bets (Ten +9.5, over 47 in NO vs NYG, and Under 37 in WAS vs KC). The expectation value distribution has returned to a near normal distribution but the predicted ROI of 106.6% is the lowest of the season. Don't be too discouraged though; this is in part just because we've become more conservative in using an adjusted cumulative distribution function as our odds of winning each event.
Our Free NFL picks require no registration or membership, are not accompanied by annoying popup windows, and are concisely located on a single page for each NFL week.
During NFL Week 5 we were 64% on bet units (58% season to date). We are 11/16 (69%) on best bets for the season.
Our picks are based on a statistical model that generates predicted outcomes using multiple systems and compares them to multiple sportsbooks. We use the delta (large is good) between the model spread and the casino spreads along with the population standard deviations (small is good) to arrive at confidence factors.
We also compute the Cumulative Distribution Function (Z-statistic) for each pick to predict a likelihood of success. You need 52.4% to overcome 10% juice and 51.2% to overcome 5% juice. Our bet units are based on both the Z-statistic and the confidence.
For more details, be sure to visit Happy’s Free NFL Picks.
We’ve also now posted our NFL week six free picks which include five (5) bet worthy Week 6 NFL lines and three (3) bet worthy NFL Week Six over unders. There are three best bets (Ten +9.5, over 47 in NO vs NYG, and Under 37 in WAS vs KC). The expectation value distribution has returned to a near normal distribution but the predicted ROI of 106.6% is the lowest of the season. Don't be too discouraged though; this is in part just because we've become more conservative in using an adjusted cumulative distribution function as our odds of winning each event.
Our Free NFL picks require no registration or membership, are not accompanied by annoying popup windows, and are concisely located on a single page for each NFL week.
During NFL Week 5 we were 64% on bet units (58% season to date). We are 11/16 (69%) on best bets for the season.
Our picks are based on a statistical model that generates predicted outcomes using multiple systems and compares them to multiple sportsbooks. We use the delta (large is good) between the model spread and the casino spreads along with the population standard deviations (small is good) to arrive at confidence factors.
We also compute the Cumulative Distribution Function (Z-statistic) for each pick to predict a likelihood of success. You need 52.4% to overcome 10% juice and 51.2% to overcome 5% juice. Our bet units are based on both the Z-statistic and the confidence.
For more details, be sure to visit Happy’s Free NFL Picks.
Sunday, October 11, 2009
Week Five Another Profitable Week - Free NFL Picks from bestofblog.net
NFL Week Five bet units are in the bag; we finished 7/11 on our bet units for the week. It looked like Houston would pull it off and make us 10/11 (we had three units on them), but still a respectable week for us. It brings us to a 109.5% ROI for the 2009 NFL Season so far and completes our fourth winning week out of five.
We want to give props to the first other statistically based pick system we've seen on the net, advancednflstats.com. I guess they're technically a competitor, but its just refreshing to find someone other than us posting useful statistically based probabilistic data on NFL outcomes.
We are able to and would like to convert their probabilities to ATS predictions; we're awaiting their permission to do so (or they may just choose to do it themselves). Anyhow, we'll post week six picks on Wednesday or Thursday.
- Happy
We want to give props to the first other statistically based pick system we've seen on the net, advancednflstats.com. I guess they're technically a competitor, but its just refreshing to find someone other than us posting useful statistically based probabilistic data on NFL outcomes.
We are able to and would like to convert their probabilities to ATS predictions; we're awaiting their permission to do so (or they may just choose to do it themselves). Anyhow, we'll post week six picks on Wednesday or Thursday.
- Happy
Thursday, October 8, 2009
Our NFL Week 5 Free Picks Are Now Available
We’ve now posted our NFL week five free picks which include three (3) bet worthy Week 3 NFL lines and two (2) bet worthy over unders. There are three strong best bets (Minnesota -9.5, Houston +5.5, Under 40.5 in the Buf vs Cle) which has resulted in a less smooth than normal distribution but a good predicted ROI of 113.7%.
Our Free NFL picks require no registration or membership, are not accompanied by annoying popup windows, and are concisely located on a single page for each NFL week.
During NFL Week 4 we were 55% on bet units (56% season to date). We are 9/13 (69%) on best bets for the season.
Our picks are based on a statistical model that generates predictes outcomes using multiple systems and compares them to multiple casino lines. We use the delta (large is good) between the model spread and the casino spreads along with the population standard deviations (small is good) to arrive at confidence factors.
We also compute the Cumulative Distribution Function (Z-statistic) for each pick to predict a likelihood of success. You need 52.4% to overcome 10% juice and 51.2% to overcome 5% juice. Our bet units are based on both the Z-statistic and the confidence.
For more details, be sure to visit Happy’s Free NFL Picks.
Our Free NFL picks require no registration or membership, are not accompanied by annoying popup windows, and are concisely located on a single page for each NFL week.
During NFL Week 4 we were 55% on bet units (56% season to date). We are 9/13 (69%) on best bets for the season.
Our picks are based on a statistical model that generates predictes outcomes using multiple systems and compares them to multiple casino lines. We use the delta (large is good) between the model spread and the casino spreads along with the population standard deviations (small is good) to arrive at confidence factors.
We also compute the Cumulative Distribution Function (Z-statistic) for each pick to predict a likelihood of success. You need 52.4% to overcome 10% juice and 51.2% to overcome 5% juice. Our bet units are based on both the Z-statistic and the confidence.
For more details, be sure to visit Happy’s Free NFL Picks.
Subscribe to:
Posts (Atom)
