2008 NFL Week 17 Picks

In order to keep this experiment running, I will post this week's computer generated picks. This week may be one of the hardest weeks for prediction since there is much more volatility than usual caused by coaches sitting down their key players. Will Eli Manning play the whole game? Will the wide receivers from Arizona play their hardest? True, these player's salary depends on statistics they accumulate the whole year, but risking an injury will cost them more than having a couple more catches.

For us making point spread predictions, we'll stay away from games whose team have already clinched the playoffs. This week, there are two games that really stood out to me as I thought the point spread was a bit off. The statistical model also confirmed this and made me like these two picks even more, these are: Chicago @ Houston and Denver @ San Diego. San Diego favorite by 8? This is going to be a close game my friends. Chicago playing for a playoff hope should not be an underdog. So those are my first two picks. The rest to follow:

Pick 1: Visiting Underdog Denver +8 incorrect point spread pick
Pick 2: Visiting Underdog Chicago +3 incorrect point spread pick
Pick 3: Visiting Underdog Miami +2.5 correct NFL point spread pick
Pick 4: Home Favorite Baltimore -12.5 correct NFL point spread pick
Pick 5: Visiting Underdog Washington +3 PUSH

GameVegas LineEstimatePrediction-Vegas





Afterthoughts Week 16

This model does not work. I have to go back to the drawing board and come up with a better way to predict sides against the point spread. I give a lot of credit to the people at Vegas putting the point spread, they know how to make the outcome almost random. I have always thought and still believe that there are games out there that are not well aligned, that the point spread is off. It is easy to see those games after the fact, but not right before the game starts. I do also think that 3 or 4 picks a week is a lot, if a system were to find these opportunities, they would be less than two games a week, if that.

What am I going to do now? I do not know. Hopefully I will have a good run in the playoffs (like I did last year) and push up the percentage to at least 55. In the off-season is back to crunching the numbers and searching for data and different methodologies to test. If by back-testing a strategy I get above 57% say, then I will introduce it next year. If I find a couple of strategies, then I might post picks from all 3 strategies and we shall all witness the best one. Still, I have a lot of work to do. For now, I will finish this season with the same methodology I described before.

2008 NFL Week 16 Picks

Regardless of how awkward the picks might seem, last week showed me there is no sure pick. Who would have thought that Cincinnati could beat Washington? How many think Jacksonville can beat the Colts? Or Seattle beating the Jets? It can happen. When you look closely at the Colts scores, a lot of their games have been close. Yes, they have 7 in a row which makes it a scary bet.

I will not try to make the case for each pick this week. In this exciting NFL week, it will be strictly the model making the picks. We'll stick to the theory I proposed in my previous blog because I do believe this model is of good value. The theory was to pick the games with the highest difference between the point spread prediction and the Vegas spread. Games with teams that have clinched playoff spots are out. The question is how many games; this week lets pick those games with difference of 5 points or more (except for the Tampa Bay pick). Without further ado, here are week 16 NFL picks of the week:

NO Pick 1: HOMEDOG Seattle +5 injury
Pick 2: HOMEDOG Jacksonville +6 incorrect point spread pick
Pick 3: Visiting Favorite San Francisco -5.5 incorrect point spread pick
Pick 4: Visiting Favorite Houston -7 incorrect point spread pick
Pick 5: Home Favorite Tampa Bay -3.5 incorrect point spread pick

Good luck!

As always, if you read more you will find the point spread predictions for each game of the week. Notice I have remove the confidence measure, it was about time.

GameVegas LineEstimatePrediction-Vegas
BUFFALO @ DENVER-7-11.7-4.7
SAN DIEGO @ TAMPA BAY-3.5-8.0-4.5

Afterthoughts NFL Week 15 - Point Spread Predictions

Another bad week for picks not so bad for point spread predictions. Let me explain. When I first started making building a model that could predict point spreads I, I wanted it to be as precise as possible given team-level data. That is, using team records, game results, home field advantage, and offensive/defensive stats I build a bunch of models in order to pick the most accurate. That was step 1.

The next step was to determine which of these predictions was best to use in making picks. The obvious candidate was those predictions that were most different from the point spread. This theory was thrown out because predictions were way off when there were key injuries or teams had clinched playoff spots at the end of the season.

I knew it would be a waste of time to try to pick every game. So the goal was to find a way to use the predictions to make picks. That is when I built the now useless confidence levels you will not see posted on the blog again. These confidence levels were merely the percentage of times predictions picked correctly against the spread in the past 7 years given the point spread and the difference between the point spread and the prediction.

I have noticed the past 3 weeks that the original idea works well if one manually disregards games with key injuries (something I cannot do automatically given the data I have). Lets take a look at this past week point spread predictions posted on the previous blog. These are the 6 games with the biggest difference in point spread vs the Vegas spread. Disregarding the Giants/Dallas game due to injuries to Burress and Jacobs all picks won except the Washington/Cincinnati game and Seattle/St. Louis was a push. Unfortunately, it is hard for me to go back to historical data and test this theory, but until I have a better confidence measure, this is what I will use to make picks from now on. Last year, I made a huge comeback in the playoffs, only making one incorrect pick. There's still hope.

GameVegas LineEstimateDifference
SAN FRANCISCO @ MIAMI -6.5 2.9 8.4
NY GIANTS @ DALLAS -3 5.7 injury
SEATTLE @ ST LOUIS 3 8.6 5.6
BUFFALO @ NY JETS -9 -4.3 4.7

2008 NFL Week 15 Picks

Last week the picks did not go well and I do not have much explanation other than bad luck. Any betting system will have streaks of good and bad luck, the important thing is to stay calm in both states. Keep your pick and betting strategy constant. For amount allocation stay with Kelly's formula and bet proportional to your bankroll. Not only will your chances of ruin are almost zero, but also when the winning streaks come, your winnings will increase exponentially.

As of this week, some teams have already clinched their spot on the playoffs. My stats cannot account for benching star players and for players not giving their 100%. That is, my model assumes that each player and coach will play with the same intensity and desire to win as throughout the year. For that reason, we will stay away from such games.

This week, the model predicted 4 visiting team favorites. For these games it might be that the home-field advantage is over estimated. Some of these games have also over shadowed teams that have been performing not so well recently, but still have a decent team. For example, Green Bay @ Jacksonville. Green Bay, although a very unstable team right now which I thought deeply to stay away from, has demonstrated a this year a few times that it can play. Rogers is good, well at least when facing bad defenses. The Jaguars on the other hand have probably thrown the towel. The market has devalued the Packers almost to the Jaguars level, buy low, Green Bay.

Kansas City has definitely improved but they are still a struggling team. Although the Chargers beat them by a point at home, they are coming off a strong win. Their mere hopes of playoffs is still alive and will pick up their play Sunday, hopefully.not enough to get swept off by the Chargers. Sell high, buy Chargers.

The next two picks are hitting on other two really struggling teams. Not the Lions, their spread of 17 might be too high, but picking the Lions is a risky move. I am talking about the Bengals and Raiders. They are both going down, by a lot. Well there they are this weeks "computer made" NFL picks. Noticed all 4 picks are VISITING FAVORITES, not my style, but that is just they way it is this week.

Pick 1: VISITING FAVORITE Washington -7 incorrect point spread pick
Pick 2: VISITING FAVORITE Green Bay -2 incorrect point spread pick
Pick 3: VISITING FAVORITE New England -7correct NFL point spread pick
Pick 4: VISITING FAVORITE San Diego -5.5 incorrect point spread pick

And below, the table with point spread predictions and confidence levels:

GameVegas LineEstimateConfidence
DENVER @ CAROLINA-7.5-5.846%
BUFFALO @ NY JETS-9-4.343%

2008 NFL Week 14 Picks

NFL picks is very much like marketing, half science and half art. You use statistics to uncover opportunities and patterns out of all the overwhelming amount of information. With this information in hand, you use 'art' or experience you have had with the numbers to make wise decisions.

Injuries in major positions bias my point spread predictions. It is better to stay away from volatile teams with recent roster changes in RB, QB, or LB positions. This is why this week, although I personally like Tennessee to roll over Cleveland, and the model also predicts they will do so, we stay away and wait for calm water. Of the other 3 NFL games with double-digit spreads I talked about in my previous post, I will stay away from all of them. Although I think St. Louis might be a good bet with Steven Jackson back in the lineup, my model cannot pick up his presence well and predicts Arizona will barely cover. Enough about what my model does not pick. Here are the opportunities it found and that I as an experienced handicapper, agree with its pick.

RESULT: 1-3 No Good

Week 14 NFL pick-les of the week:
Pick 1: VISITING FAVORITES Jets -4 incorrect point spread pick
Jets lose one game and suddenly their spread drops to 4 points? They were favorites to the Rams on the road by 9 and at home with Denver by 8. This team is thinking Super Bowl and San Francisco is thinking next year. Favre, pass it to your teammates and you shall cover.

Pick 3: HOME FAVORITES Da Bears -6 correct NFL point spread pick
Wake up and smell the coffee Orton. You have a decent defense, an awesome running back, and a great wide receiver/punt returner. If you have to go to your old ways of handing off the ball all game, do it!

Pick 4: VISITING UNDERDOGS Atlanta +3 incorrect point spread pick
They beat them once and they'll beat them again. Brees, they have you all figured out.

Pick 5: HOME FAVORITES Green Bay -6 incorrect point spread pick
So the Texans beat the Jaguars, Browns, Bengals, and Lions, big deal. Packers have had a tougher schedule and have crushed the Bears, Colts, and Seahawks.

I forgot to post the week's prediction table. I will do so this week if you read more. Good luck!

GameVegas LineEstimateConfidence


Big Spreads in Week 14

Last week, 12 visiting teams covered the point spread on NFL games. The two biggest spreads were covered this week. At the beginning of the season, spreads over 10 were covered about 20% of the time, but in the past two weeks this number changed to just under 60. Underdog teams included in these big spreads have been Detroit, St. Louis, Cincinnati, and Seattle.

This week there are four games with big spreads in favor of the home team. Will the visiting teams cover these spreads? Or will the heavily favorites run by these unfavorable teams?

As of today, four games have double-digit spreads. First, Cincinnati at Indianapolis. The Colts have won their last five games, but none above 7 points. Cincinnati is partly throwing the towel already and the Colts might use this game to blow them out and lift their spirits for the playoffs.

Also receiving double-digits point spread for Week 14 is Cleveland at Tennessee. Las week, Derek Anderson injured himself bringing in the backup's backup for week 14. Ken Dorsey is in, who? One might think that Tennessee should easily cover, but Cleveland this year being so unpredictable, no one knows how this new comer can come out throwing.

The Oakland @ San Diego game opened at 10, and today is was at 9.5. San Diego has lost the last 3 games by less than 6 points. At 4-8, the Chargers do not look that well off than Oakland. The Raiders have had surprising wins against Denver in week 12 and the Jets in week 7. The have kept Carolina and Miami within 9 points in the past 3 weeks. Was this spread estimated based only on their previous game this year when Oakland lost by 10? I'm leaning Oakland, but we'll see what the predictions say.

Finishing off the double-digit spreads on the Week 14 Bodog's NFL point spread is St. Louis at Arizona. The Rams have lost their last six games, four of which were blowouts. Last week, they kept it close with Miami, why? I think Steven Jackson's presence in this team is huge. I think they will give the Cards a hard time (if Jackson plays) this week.

Afterthoughts Week 13

Why were the picks so good this week? Was it because in this odd week visiting teams covered the spread in 12 of 16 games and all my picks were for visiting teams? Was it because the model is earning back some of its losses from that past two weeks to jump back to 60% levels? Or was it because my picks did not involve games whose teams did not have serious roster changes? I do not know the answer to these questions, but I do know that we went 4-0 and we are back above 60% ATS!

After two weeks of disappointing results, the predictions really stepped up to the plate this week. Now I am confident again that there is something in this complex statistical model. Many stopped visiting the site because of these two disappointing weeks, but some encouraged me to keep going forward.

Like many statistical models, this one predicts well in the long-run, but is volatile week by week. That is why I encouraged, in a previous blog, to use betting size wisely. In the stock market, it is hard to predict exactly when stocks will hit bottom, but with fundamental analysis, one can spot stocks that will pay back well in the long run. Not all stock picks will pay off, but one makes sure that those that do, do well. My picks and model behave the same, not every week is going to be a winning week, but in the long run and with careful handling of the money, the payoff will be profitable. I cannot be those handicappers that advertise every week that they went 4-0 or 3-1 last week. Nobody can do that, they have to be lying! I can only say that I have been at 60% in my first two years of giving free handicapping picks.

Looking forward to week 14.


2008 NFL Week 13 Picks

Thank you all for your encouragements. Some of you have given me great ideas to develop on and some have just given me the encouragement to keep this experiment going. This week my picks are based on the model's predictions, as always, but the confidence measure is being ignored. I am posting picks on the predictions that are farthest from the spread without any teams suffering from recent injuries or changes.

This week, all I had time was to upload last week's data and run the model. Comments for each game have been avoided as well as the point spread predictions. I feel confident this week's picks will come through as we all munch on turkey legs. Enjoy your Thanksgiving. Ignoring misconceptions, team discrimination, and going back to my old self of picking teams based on statistical predictions, here are week 13's NFL point spread picks:

Week 13 RESULT: 4-0 Perfect!

Pick 1: Kansas City +3correct NFL point spread pick
Pick 2: New Orleans +3.5correct NFL point spread pick
Pick 3: Pittsburgh +1correct NFL point spread pick
Pick 4: Tennessee -11correct NFL point spread pick

Afterthoughts from Week 12

It is easy to say, this blog or these picks suck. You know why? Because I am holding myself accountable by making my picks public and keeping track of them. Many "Football Experts" from ESPN and other stations change their opinions on teams every week (sometimes even during a game), but they never play back the videos when they said New Orleans was going to win their division or the Jets were not going to make the playoffs.

Is it time to go back to the drawing board? Probably, but I have to stick with what I have this year. Although I am not completely dissatisfied with the point spread predictions, it is time to send the confidence measure out the window. For example, last week the prediction for the Dallas game was a point over the spread. The confidence was thirty-something percent, so I chose SF as my pick, wrong.

I have had issues with this confidence measure from the beginning. It is not game specific but situation specific. Also, situations with big sample sizes (> 80), the confidence tends to be close to 50%. Meaning, if all situations were as likely, all confidence measure would be close to 50%. The confidence should be a function of the teams involved, the off/def stats, and the spread currently only the spread and the prediction is considered.

Lesson 1: Throw the confidence measure out the window and do not chose the reverse picks.

The other issue is that I only have access to game/team level data. So if I player is absent or has been absent, the model will not pick up on it directly. One of you pointed this week at the Dallas game and if it was considering Romo's presence and health. Well, no, I cannot statistically tell the model "Romo is expected to play better today, please account for it", if I had player level data, I could probably do something like that. So what lessons did I learn from this?

Lesson 2: If the team has changed one of its key players in the past 5 weeks, stay away from that game.

I am not throwing the towel yet. At 56%, I hope to bounce back to 60 before the end of the season. Every week is a learning experience and with every experience we become better decision makers. I may be equipped with great statistical tools, but when the pressure is on and we have to decide the NFL picks of the week, we have to keep our plan and stay cool. If we let emotions make decisions for us, or if we become biased towards certain teams, we will not be successful. Stay calm and act rationally and remember that as bad as a pick might sound, Oakland covered this week.

Betting Size and Kelly's Formula

After several weeks of highly volatile results, I decided to go out and search for a mathematical formula to compute betting size for each game. This is a topic that I would like to dedicate more time to and even come up with a more sophisticated method to decrease my chances of ruin and increase my profitability in the long run. For what is left of this year at least, I will use Kelly's formula.

Kelly's formula determines the percentage of your bankroll to use on a bet. In our case of the weekly NFL spread, it will determine the amount of bankroll to use each week. The formula is simple and takes into consideration the odds received on the wager and the probability of winning.

Fraction of Bankroll = [(odds)*(prob) - (1 - prob)]/odds,

where prob=probability of covering the spread and odds=odds given for the wager

I will make the following assumptions and then compute the fraction of the bankroll to use each week:

  • Odds = 10 to 11. The typical -110, for every $11 you bet you receive 10. Most bookies change these odds depending on the game and which team you choose. To simplify the computation, we assume the odds are always -110.
  • Prob = 60%. The probability that my system makes correct picks is assumed to be 60%. This is where we stand today and this is where we finished last year. We assume, regardless of the confidence measure, that every game has equal chance of success.
  • Picks are independent of each other. Kelly's formula assumes independent events. The fact is that NFL picks are not like rolling a ball on a roulette, but it is not that far off to assume they are independent. Here we are assuming that last week's New England pick is not correlated with this week's New England pick because the Vegas spread adjusts to avoid trends.

Therefore, the formula simplifies to: [(.909)(.6) - (.4)]/.909 = 16%

Now, this does not mean that for every game you will spend 16%. Kelly's formula applies to sequential gambles. Because on Sundays, bets are mainly made simultaneously, I would use 15% (a bit less to go on the safe side) of your bankroll every week (including Thursday, Sunday, and Monday night). Although Thursday, Sunday, and Monday night games you are able to make bets sequentially, I do not want to give more bankroll to these games just because they are the night games.

For example, if you have a bankroll of say $1,000 then you can risk around $150 per week. If there are 3 games you choose to bet one, then (for now) bet $50 on each game. Next week, if your bankroll increases or decreases, you adjust the total per week accordingly. At the end of the year you will not be ruined and hopefully you have a decent profit.

To put this in perspective, suppose you start investing in week 4 with a bankroll of $1,000. At $150 a week and 60% success rate you would average about $35 profit a week. Even if you do not increase your weekly bet as your bankroll increases, which you should, you would end up (13 weeks plus 2 playoff weeks) with a total profit of $35*15 = $525 or about a 53% return on investment. Not bad considering there seems to be no bottom in the stock market. Kelly's formula is only valid as you make more and more gambles, that is, it is not a short-term solution it is an optimal solution for the long run (think 10-20 years).

I know it is tempting to bet it all or larger amounts, especially when you want to make up ground. Betting it all or deciding on a size without a strategy will make you act like a gambler and not like an investor in the NFL point spread market. Stay calm, optimize betting size, play it safe, and use 15% of your bankroll every week. Even if you think Buffalo is a given cover this week, remember that in any given Sunday anything can happen.

In the future, I would like to relax the second assumption, the one that states that every game has equal chance of winning. I have been publishing a confidence measure that although is detecting opportunities well, withing those opportunistic games, it is not clear-cut that one game is better than the other. If this measure becomes more accurate, then of the $150 for the week, the distribution of these $150 can be done more optimally according to the refined confident measure.

2008 NFL Week 12 Picks

Enough crying about last week. We are still at 60% and learning more about becoming a better handicapper every day. This week, the model's point spread predictions follow the same methodology, but the way we make the pick is tweaked a bit.

As we can see from the table below, the following picks have a pretty good confidence and a comfortable difference between the point spread and the Vegas spread: Green Bay +2.5 against New Orleans, Buffalo -3 against KC, Tennessee -5 against the Jets, San Francisco +10.5 against Dallas, and Cleveland -3 against Houston.

Of these picks, the consensus has Cleveland at 65%, Green Bay at 59%, San Francisco at 59%, Buffalo at 55%, and Tennessee at 40%. Now, I am convinced that going with the consensus all the time is not a good strategy because the house will always win. We need a better edge and that is where advance modeling predictions hopefully step in. Combining these two information sources will hopefully provide us with solid picks.

You have to take those consensus picks carefully. For example, Indianapolis, Washington, and New England are the top three this week and my model is not detecting a solid edge on neither of them. Some of the games that I mentioned are backed up by the consensus, except for one of them. This one is the Jets/Titans game. As interesting as this game might be, I will stay out of it. Bettors are favoring the Jets and my model predicts the Titans will win by more than 5. The other games consist of Week 12 NFL Picks:

Week 12 RESULT: 1-3 %$#@&*%$#%^?!

Pick 1: VISITING UNDERDOGS Green Bay +2.5 incorrect point spread pick
Pick 2: VISITING UNDERDOGS Buffalo -3 correct NFL point spread pick
Pick 3: VISITING UNDERDOGS San Francisco +10.5 incorrect point spread pick
Pick 4: HOME FAVORITES Cleveland -3 incorrect point spread pick

If there is any significant injuries or situation I should be aware of for any of the above picks, please let me know. As always, read more to see the table with point spread predictions for each NFL game.

GameVegas SpreadPredictionConfidence
OAKLAND @ DENVER-10-10.550.1%

Afterthoughts Week 11

Every week is a learning experience. In order to become a better handicapper I have to avoid weeks with no wins. Week 11 turned out to be one of these. If I would have stayed within game plan, I could have done slightly better. Let me explain. We are hanging by a thread on 60% and need to make solid picks in the coming weeks in order to beat my goal and finish the season above 60.

In the table from the previous post, Oakland was as good a pick as Atlanta, so why was it not included in one of the picks. One word: Emotions.

Lesson #1: No matter how bad a team is, there is always a possibility that the team will cover the spread. If model indicates a pick for this underdog, go for it.

I do not regret the Pittsburgh pick because we got cheated out of it. Although Pittsburgh covered the spread with a last second touchdown, after various reviews, it was determined that the last TD did not count and the final score was 11-10. Referees admitted it was an error, no forward pass was don, but still that did not change the score.

The Atlanta pick I still think was a good one. They did not play as well as they could and lost. Fine I'll take that one.

The confidence measure is not providing reliable results. I can't remember a week where the highest confidence game covered. Also, when the confidence is far below 50%, doing the opposite is also not being as consistent as it should. I should consider doing each confidence dependent on the teams in the game. Currently, the measure takes into account the spread and the prediction only.

The other possibility is considering the wisdom of the crowds. Crowds are not always good predictors of games, but their knowledge tied with my statistical analysis might work well. Wagerline provides more than 3,000 people's picks of for each game. For this week for example, picks above 59% went 5-2 (something you do not see every week). These picks coincided with my point spread predictions for all games except Miami which my model correctly predicted would not cover. Their top picks were: ARI, PIT, TB, MIA, SF, NO, and TEN.

For today's game, Monday Night Football, between Cleveland and Buffalo, 52% is picking Buffalo, a no pick in my new hypothesis. The model predicts Cleveland will win the game by a few points. So the best thing to do for this game is to stay away from it.

I will not change my philosophy, what I will do is tabulate historical consensus picks and see if my hypothesis is true. Hypothesis: My statistical analysis complemented by consensus picks does fairly better than just my picks or just consensus picks.

The Other Week 11 NFL Picks

Unfortunately, yesterday's game did not result as we expected but we have 2 picks to make up for it. The Patriots made a great run at the end but fell short of the game and the spread. The good thing is that the loss was just a slight dent on the bankroll. For Thursday's game I had warned about the Patriots' linebacker injury and that it should be considered in the percentage of bankroll to be used for the week. Hopefully, there is plenty left for these Sunday's picks:

Week 11 RESULT: 0-3 What happened?!

Pick 2: Pittsburgh -5 incorrect point spread pick
Last week the Steelers had the same spread against a stronger team and without their starting RB. This week Parker is back, Ben will play well to not lose the starting role and the defense will also help. The Chargers have been playing well lately, but Rivers will have to worry against the number one defense against the pass.

Pick 3: Atlanta -6 incorrect point spread pick
Denver is hurt both on defense and offense. Atlanta is unbeaten at home and those wins have been by wide margins (except against the Bears). I think Vegas has not yet realize that the young guys in Atlanta have learned a lot this season and can play with the big boys.

You can also read more and see the point spread predictions for each game. These were the two games that stood out to me as great picks. Other picks I considered posting were Philadelphia at -9 and Oakland -10.5. Feel free to comment and give your opinions on these games or others.

NFL Point Spread Predictions Week 11:
GameVegas SpreadPredictionConfidence
OAKLAND @ MIAMI-10.5-6.456%

2008 Week 11 NFL Picks: Thurday Night Game

I have 3 picks for you this week. The first pick is Thursday night's game. The other two picks will be posted Friday. This is a pick I believe you will not like. I have my doubts too, I couldn't believe it. But take those opinions and emotions and throw them in the garbage and listen.

The New England Patriots and the New York Jets are considered to be of equal rank by Vegas. Some sites have the Patriots favorites by -3 and some by -3.5. They are tied at the top of the AFC East and both have a 5-4 ATS record this year. Offensively, New York and New England are not as far apart as one would think. Besides points, which the Jets are averaging 28 and the Patriots 20, they are pretty much even in every other category. Defensively, the Patriots have problems stopping the run and the Jets the pass and it all it evens out to both teams allowing around 300yds/game.

The Jets have won 3 in a row and 5 of their last 6, although of those 5 wins, 3 include STL, KC, and CIN. Both teams have defeated Buffalo (by more than 9pts.) and lost against San Diego (by more than 19pts.). This two teams seem to be in fact even, so a spread of 3.5 seems quite right, right? Not for me.

This spread is brought down from what it should be -5 or -6 for the Patriots. This is due to the Jets' games against St. Louis and Arizona where they beat them by 44 and 21 points respectively. See, some guys in Vegas are running similar models I am and some are using a pool of opinions and bets, I assume, to come up with the point spread. The problem lies in weighing this 44 and 21 point spread the same as the other games. My model does not suffer from "fat point spreads", it recognizes abnormal scores like 47-3 and minimizes its impact. Isn't that cool?

When a game is predicted to be between 2 and 4 points in favor of the home team when the spread is near 3, the model is correct 60% of the time in the past 10 years, hence my first pick of the week.

Pick 1: New England -3.5 incorrect point spread pick

One big concern is last game's injury to linebacker Adalius Thomas. He is probably out for the season. This is a guy that averages 7 sacks per season and already has one on Favre. For this reason, I will save a bit more of my weekly money pot for the other two games.

Afterthoughts for Week 10

After debating a long time whether I should have picked Pittsburgh at home or Arizona on Monday night, I am glad I didn't. Following my lessons learned from week 9, I kept my cool, stuck to the strategy, and did not get greedy.

Although the algorithm predicted these two teams to cover the spread there were some issues. Pittsburgh had the uncertainty of Ben playing and RB Parker for sure sitting out. The Cardinals were favored, but the confidence was low (54%). San Francisco was using a new QB and Singletary was coaching them for the 2nd time. Both picks would have lost.

The strategy is to bet for games whose situation is seen previously (throughout the years) more than 20 or 30 times (big sample size) and whose confidence is at least 57%. With a confidence of 54%, I do not have enough proof to say that statistically I have found an edge. When a game is sitting at above 57% confidence of covering the point spread, then I look for injuries and other extraneous situations. If everything looks normal, as it did for the 3 picks we ended up making, then we go for the pick.

We have also learned that of the games that we should make bets for, those with higher confidence are not necessarily doing better. That is, a 58% confidence is not beating a 64% so the strategy (for now, until I figure out a better confidence measure) is to spread the wealth evenly among picks and diverse the portfolio. There are many gambling strategies in terms of the amount to put in, I will not get into that, but I will say this: do not put all the eggs in one week.

2008 Week 10 NFL Picks

RESULT: 3-0 Perfect!

Tough week for picks. My statistical model did not find many opportunities and those that did are somewhat sketchy. For example, it found Pittsburgh to win at home, but will Ben play? There is too much uncertainty if he doesn't and I would not recommend the bet. Also found the Jets at home against St. Louis. This game is off the table in many sites so if you can get -8 go for it, anything above that would be a scratch. There were 3 other games whose confidence was not as high as the ones I just mentioned, but that still are great plays.

Without further ado, here are Week's 10 NFL Spread Picks, back with comments:

Pick 1: VISITING UNDERDOGS New York Giants +3 correct NFL point spread pick
If you think these two teams should have the same rankings, then the spread is correctly priced. If not, I would assume you think the Giants are slightly better, so does my computer prediction. Tthis spread should be either a pick or the Giants slightly favored.

Pick 2: VISITING UNDERDOGS Kansas City +15 correct NFL point spread pick
This young team is getting tired of being slapped around. They are not the Rams or the Raiders, they will come out swinging. San Diego has scored well at home this year, but another +15 spread at home is highly unlikely.

Pick 3: VISITING FAVORITES Tennessee -3 correct NFL point spread pick
Last week was Tennessee's first ATS loss, nothing to worry about. The Bears are heavily favored at home because of some 4 or 5 game streak. Pay no attention to that, Tennessee will crush them regardless of who throws the ball for the Bears.

Read more to see the table with the game spread predictions and its confidence (probability of the prediction being on the right side of the spread). Remember, a negative sign implies home team is favorite.

GameVegas SpreadPredictionConfidence
ST LOUIS @ NY JETS-8.5-1158%
SEATTLE @ MIAMI-8.5-8.150%

Afterthoughts from Week 9

Thank you to all of you that encouraged me to keep going after week 8's losing streak. We came back with a solid 4-1 ATS in week 9. Your input and encouragement motivates me to improve and put the time to make this work.

One of the biggest benefits I get from doing this week in and week out is the lessons I learn every week. Each lesson I write down in this blog, I carry it with me, and I apply it to my methodology of making picks. As I learn and get experience, I will become a better handicapper.

Here are this week's lessons:

Lesson #1: Don't get greedy

If we've had a perfect or profitable week and there is only Monday Night Football left, why gamble? Even if the confidence measure is the highest for MNF, take your profits and buy a gift for your kid. Wait for the next set of games, there is no rush.

Lesson #2: Check for Key Players Coming Back From Injuries
I need an assistant, help, comments. Someone that tells me, hey Jaime I notice you picked Washington this week, did you notice that Parker is back at the RB position for Pittsburgh this week? Is your NFL model accounting for this? NO. We have a community of readers in this blog that can contribute their knowledge by adding comments to the NFL point spread picks. I will do my part and check more thoroughly for key injuries.

Lesson #3: Ignore Incoherent Trends
Although picking Pittsburgh on MNF based on the election trend would have given me a win, it is better to ignore these coincidences. As most of you have heard, if the Redskins win, the party who won popular vote the previous election wins(used to be party who won the election the previous year until 2000, that is when they changed to popular vote). Looking at the polls and seeing a comfortable Obama lead would have led to a Steelers pick. But come on, this is nuts! Finding a correlation between two completely independent events is not what smart NFL pickers do. Stick to your game plan, ignore incoherent trends, and find patterns that make sense, have significant meaning, and a big sample size.

See you next week!

NFL 4-0 Sweep this Week, get Monday Night's Pick HERE

I get the "NFL 4-0 sweep this week" emails every week from various sites selling picks. I have always wondered how truthful they are, but never had the extra money to pay and find out. Has anyone tried? No one can be perfect or even close to perfect every week. This week, as you have witnessed, it is my turn to truthfully say it. Here it goes:

Dear Reader,

We pulled off the perfect 4-0 in the NFL this week! If you are interested, you can see this week's picks in my previous blog post.

This season we have completed a total of 26 picks, 17 of which have been correct for a total of 65% ATS. To finish off the week, we have one NFL pick left, Monday Night Football.

Remember, only our premium readers get access to our highest value picks. Purchase now to make sure you get access to the picks. Take advantage of these premium winning picks from statistical professionals who are crunching the numbers for you to create winning spread strategies. Order here. Just kidding, they're free. This week's Monday Night Football Pick is:

Pick 5: Washington -2 incorrect point spread pick

2008 Week 9 Picks

RESULT: 4-0 Perfect!

Last week I felt comfortable about most of my NFL picks, especially Carolina and San Francisco (both ATS losers). Except for last week, after I make my picks, I check how popular these picks are. I look at Yahoo's Spread Pick'em League as well as the NFL Consensus Picks at wagerline.com. Doing this for the past 2 years has given me the mental evidence that popular picks do not do well. I do not have the statistical evidence to back this up, but if I do have some time and need to prove it I will.

Last week the two most popular picks were: SF (67%) and CAR (65%), my two favorite picks for week 8, both did not cover. On week 7, Indianapolis (74%) and the NYJ (73%) were the two most popular, neither covered. Although in week 6, both the Eagles (68%) and the Saints (69%) were the heavily favorites and they did cover. As my previous blog on sample size says, we would need many more games to come to any conclusion on this. My hypothesis is that "popular" games are either 50-50 or slightly below 50%. If a game is so popular wouldn't it be more likely to be wrong? I think this for two reasons: 1) Vegas may be guiding bettors one way for more profits and 2) if many are selecting the game then there is no hidden treasure in finding that there is bias on the spread. This is the goal of a good handicapper, and in my case the purpose of using an advanced mathematical model; to find bias in pricing the spread for a game but if everyone is finding it, it is no secret and there is no need for mathematics.

Whatever the case may be, 2 of my 4 picks this week are the most popular ones, the other two are quite unpopular, I'll let you find out (if you care) which is which. I will stick to the model and disregard whether they are popular or not. If consider popularity, I am tramping with the model and we will never see how good it really is. Enough talk, now the NFL Spread Picks of the Week:

Pick 1: VISITING FAVORITES Philadelphia -7 correct NFL point spread pick
Pick 2: VISITING FAVORITES Atlanta -3 correct NFL point spread pick

I can't decide between two picks the model detects as opportunities, KC and MIN. The HOME DOGS KC +9 showed some signs of hope last week, but it was against the over-rated Jets. You can expect them to score about 10 points in this game against the #2 defense in the league. Tampa has struggled on the road losing to the Saints, Broncos, and Cowboys and barely beating the Bears. They might struggle again, probably win, but will the offense produce enough on the road?

The Texans face their easiest challenge on the road after losing the first 3 weeks to a very tough schedule. Now they have 3 wins in a row, of which 2 were against the 2 worst teams in the NFL. Minnesota on the other hand find themselves at having a shot at the division, if they can beat someone in it. By securing a win this week, they might have a chance at it. I do not like the fact that they have no passing game. Even with great WR Berrian and Wade, Mr. Frerotte cannot seem to find them so easily. This makes it hard for them to pull away on a game and win decisively.

I should probably not choose either, what do you think? But instead, I am choosing both and once again distributing wealth evenly to protect my portfolio.
Pick 3: HOME UNDERDOGS Kansas City +9 correct NFL point spread pick
Pick 4: HOME FAVORITES Minnesota -4.5 correct NFL point spread pick

Pick 5: Monday Night Football (PIT @ WAS) this pick will be posted on the next blog post...

Read more to see the table with the raw outputs for each game's point spread prediction and confidence measure.

GameVegas SpreadPredictionConfidence


Afterthoughts from Week 8 - A bigger Sample Size

Three years ago a friend told me that I should use my analytical skills to predict the point spread. So I decided to build 10 different models, test them, and start making predictions using those models that work best.

I experimented with simple offensive and defensive ratios, various linear regression models to predict point spread, a logistic regression (binary outcome) of the probability that a team would beat the spread, and a neural network to predict straight winners. Neither worked so well in predicting ATS winners.

Then I started thinking of a way to combine and categorize these models in search for games in which the spread is a bit "off". After combining the estimates of those models that turned out to be somewhat successful (>52%, the best only predicted 55%), splitting the point spread and the estimations into categories I started to see some good percentages.

Games for which models predicted the home team to beat the spread and the home team was favorite by 3-4 did the best, on average predicting above 65%. Also, heavily favored visiting teams for which the model predicted they would slightly not cover also did well (63%). But what happened on week 8? Didn't I choose those categories for which the games were predicted to be above 60% successful and only got 2 out of 5? Yes, the problem was the sample size.

I have categorized the point spread into 10 categories and the estimation into 8, this totals 80 different possible categories that a game could land on. Although I have 10 years of historical data, some of these categories have less than 10 games, a very small sample size.

This is why when you see stats like "this team is 5-0 ATS at home on Monday night", you should ignore it. 5 games?, that's it? A reasonable sample size would begin at 30, and once you reach 50 or 60 you can feel confident there is a significant trend.

Looking back at past week's point spread picks, I would have done much better had I sticked with bigger sample sizes. I felt into a gambler's trap; I thought there was a trend and went for it when there was not enough evidence to determine there was indeed a trend. I have already written 50 times,'do not pick a game that does not fall into a category with a sample size above 20 and a confidence level greater than 58%'. This is my 3rd year doing this so I guess it is the lessons that a beginning handicapper learns.

I still believe this strategy will pay off. Currently, we are sitting at 59% post week 4, not bad, but I want to be above 60% by the end of the year. You will keep receiving my analysis and picks of the week, at least this year, if I am not above 60% by the end of the year I will close the blog.

Any comments? Good luck.

Monday Night Football Pick Week 8

After obtaining this week's result data, I decided to run the model again to estimate the point spread on the Monday Night Football game. The result was that the prediction model estimated Tennessee to win by 10, making a 6-point difference between the estimate and the spread. I saw this difference in many games this week that would have won had I picked them. So although the confidence is not so high, a measure I still have many doubts on, I am making an extra pick this week.

MNF Pick: Tennessee -4 correct NFL point spread pick

2008 Week 8 NFL Picks

After sticking to my model's prediction and ignoring what the consensus and others were saying, we had a solid 4-1 last week. The model correctly predicted and probably "knew" that Carolina would be better at home against NO and that New England had not forgotten how to play defense. I regret picking the Colts last week, but only because it was their first week without RB Addai. In any case, I like the way you did not involve emotion in the decision making and I like how the model is picking up trends that people forget about when a team has one or two rough weeks. Again, the objective is then to minimize risk, make solid NFL picks, and in the long run it shall pay off.

Looking at the stock market down today 500 point and the Dow hitting 8500, NFL betting suddenly seems to provide a safe haven for investments.

The NFL Spread Picks of the Week (8) are:

RESULT: 1-4 Better weeks are around the corner

Pick 1: HOME FAVORITE Carolina -4.5 incorrect point spread pick

Pick 2: VISITING FAVORITE Washington -7.5 correct NFL point spread pick

Pick 3: HOME FAVORITE Pittsburgh -3 incorrect point spread pick

Pick 4: VISITING FAVORITE San Diego -3 incorrect point spread pick

Pick 5: HOME FAVORITE San Francisco -5.5 incorrect point spread pickHasselbeck is out, pick is in for this week.

As you can see from the table below (Read More), there could be other opportunities although they did not show all the signs I like to see from my stats. Last week, all the 'other good picks' I mentioned were not good at all. So you might want to disregard these or look at them a bit closer. These games include the 3 Home Favorites by a touchdown: Jacksonville, Baltimore, New England, and the visiting underdogs Cincinnati +10. Good Luck!

NFL predictions against the spread and confidence for all games in week 8:

GameVegas SpreadPredictionConfidence
KANSAS CITY @ NY JETS-13-13.553%
TAMPA BAY @ DALLAS-3.5-3.553%
BUFFALO @ MIAMI1.58.950%

NFL Point Spread Standings - Up to Week 7

Not surprisingly, Tennessee has been shocking everyone including bettors. After Vince Young left, nobody thought they would be 6-0 and less so 6-0 ATS this year. Has everyone caught up on this? Is the spread going to over react to them now? I do not think so only because they lack star power. I see the spread under-estimating this team throughout the year, including the coming week against the Colts at home.

I just wanted to put this up there. These numbers do not mean much (to me). If anything, I would be cautious of teams that have a very bad and very good record because the market will start adjusting on the other side. It is interesting to see so far which teams have been over-rated: CIN, DAL, MIN, SF, and DEN and which teams have been under-rated: TEN, TB, ARI, ATL, BAL, BUF. Look at my previous post to see how these teams performed against the point spread in previous years and see if you find any significant trends, if you do let me know.
2008 NFL Records vs. Point Spread

TAMPA BAY52071.4%
NY GIANTS42066.7%
GREEN BAY43057.1%
NEW RLEANS43057.1%
SAN DIEGO43057.1%
NY JETS33050.0%
KANSAS ITY24033.3%
ST LOUIS24033.3%

2008 Week 7 NFL Picks

After a disappointing week 6, I am ready to purely stick to the model's prediction. The Giants loss on Monday taught me a big lesson: forget how a team has been performing, judge the spread only, and remember anything can happen. The objective is then to minimize risk, make solid picks, and in the long run it shall pay off.

The NFL Spread Picks of the Week (7) are:


Pick 1: VISITING FAVORITE Tennessee -8 correct NFL point spread pick

Pick 2: VISITING FAVORITE Indianapolis -2 incorrect point spread pick

Pick 3: VISITING UNDERDOG Seattle +11 correct NFL point spread pick

Pick 4: HOME FAVORITE New England -3 correct NFL point spread pick

Pick 5: HOME FAVORITE Carolina -3 correct NFL point spread pick

I personally like the Colts rise to performance, Kansas City's face-off with the 5-0 Titan's and Seattle's back-up quarterback and high spread against Tampa Bay. I do not like going against the Saints in their current streak nor do I like the Patriots' increasing troubles, but I will stick to predictions and trends and hence pick CAR and NE. As you will see from the table below, other good picks could be: Jets +3 and San Diego 0. Dallas is off the board with Romo's absence.

As requested by various of my readers, below I have attached the raw output table of my model. The prediction and spread are based on the home team meaning if the sign is negative it favors the home team and vice-versa. The percent confidence is a calculation I make in order to determine which games seem like a good betting opportunity.
GameVegas SpreadPredictionConfidence
DETROIT @ HOUSTON-9-16.550.5%
SEATTLE @ TAMPA BAY-11-14.035%

2008 Week 6 NFL Picks

Last week, we found various visiting teams that were under-rated and came out to beat the spread. This week, I am finding a similar story. Usually, Vegas assigns a 3-point advantage to the home team. Sometimes bettors increase this home field advantage just because they are "good at home" without considering as much who are they playing against.

After going 5-1 last week, I feel comfortable with the amount of data we have for this year in order to account for this year's trends. I have been quite busy lately balancing work and a new baby. This means that I will keep my comments to a minimum. Last week I posted no comments on the games and look at how good we did, for superstition's sake, I will do the same this week. I was still able to run my model, analyze results, and give you the free nfl picks of the week.

In regards to the confidence measure I used to post that some of you have asked for, I have to work on it and measure its true validity. As for now, I suggest an even distribution of capital across the picks you decide to use.

Point Spread NFL Picks, Week 6:

RESULT: 2-3 It's ok, we'll get'em next week.

Pick 1: VISITING UNDERDOG Jacksonville +3 correct NFL point spread pick

Pick 2: VISITING UNDERDOG Oakland +7.5 incorrect point spread pick

Pick 3: VISITING UNDERDOG New England +5.5 incorrect point spread pick

Pick 4: VISITING FAVORITE Dallas -5 incorrect point spread pick

Pick 5: HOME FAVORITE New York Jets -6 correct NFL point spread pick

I have to admit that I was also tempted to pick Detroit, a home underdog by 13 against Minnesota. In order to keep the record for the computer picks only, I will ignore it, but personally sounds like a good pick.