2008 Week 5 NFL Picks

Be patient my friends, the model will be running with enough data after week 5. I suggest not using these picks, but after this week you will see how well the computer picks up trends happening with teams regarding the spread. I do not believe in looking at recent stats like this team has been 4-1 ATS in the past 5 games or this team at home has X ATS record. Those stats like that are interesting but useless.

An NFL spread statistical model can pick up overall trends in previous years and applying that knowledge to the current team trends. That's the secret and that is why I have been successful the past two years. Unfortunately, before week 5, there is not enough data to account for this year's performance.

This week there are some tempting spreads if you are willing to take the -125 odds. These include taking the Colts, Patriots, Chiefs, Seahawks, Buccaneers, or Dolphins Colts and Patriots are favorite by only 3 on the road against Houston and New England, tempting. Cincinnati is underdog by 17 against Dallas, also tempting. The rest I am not sure are worth the odds. But don't jump your boat. That is why I will give for the last week, the DO NOT USE picks of the week:

RESULT: 5-1 We're Back in Business!

Pick 1: Buffalo 0
Pick 2: Washington +5.5
(will Washington's defense shortage hurt them more than Brian Westbrook's abscence? Confidence decreased.)
Pick 3: Tampa Bay +3.5
Pick 4: New England -3
Pick 5: Cincinnati +17
Pick 6: Indianapolis -3

2008 Week 4 NFL Picks

As the model picks up its third week of data, I am starting to feel more comfortable with the computer picks. Still, it is too early to rely completely. I have mentioned that the statistical model will not be in full force until week 5. Nevertheless, it is more fun to see some picks and see how the predictions fare in the early weeks.

Lots of free picks this weeks! Bet wisely, diversify you portfolio with these picks, and enjoy the returns.

Here are the OFFICIAL Week 4 NFL Free Picks:

RESULT: 2-3

Pick 1: Green Bay +1
After two wins against Minnesota and Detroit, Packers lost to the unstoppable Cowboys. They did not get blown out though. Tampa Bay on the other hand should have lost against the Bears, but due to the Bears lack of skill, they somehow managed to win in OT. This only tells me that Tampa is not as good as people think they are. My model also thinks so. It's predicting Green Bay by 6 points and a 65% confidence of beating the spread.

Pick 2: Arizona +1.5
How long will the line give so many points for a famous QB? What about the other players? Favre unfortunately can only play QB, not everything else. The Jets roster has not changed much from last year, i.e they still are pretty bad. Their weak offensive line will have Favre running around and under pressure. The Cardinals on the other hand, have played well passing and rushing. Despite their loss to Washington last week, they still managed to play a good game, run for 116 yards and Warner's 192 yards kept them in the game. In brief, Arizona stops the pass (Favre gets 1 TD 1 Int), Warner gets 3 TD, close game, but Jets won't win, Arizona by 5.

Pick 3: Philadelphia -3
You do not need a Ph.D. in statistics to be all over this game. Orton and the Bears are going nowhere. Hester is back, but so what, just do not kick to him. The Eagles are playing great this year with a close loss only to Dallas. Brian Westbrook will probably be out Sunday night, but McNabb will play. This spread should have been around 6 or 7, is Westbrook worth the 3 or 4 point difference? We'll see. The Bears defense will hold off the Eagles within 10 points, but 3 points is just too low.

Pick 4: Buffalo -8
The model is all over this game predicting Buffalo to win by 15 points. That comes as no surprise since the Rams are probably have the worst team ranking and Buffalo is in the top 10. This spread is not 10 or 11 because in the past two games, Buffalo was not able to completely dominate its opponent and only winning by 1 and 4. The other reason is the unpredictable performance of 38 year old Trent Green who starts his first game of the year. I am keeping the pick, and manually lowering the confidence 3 points.

Pick 5: Tennessee -3 Confidence = 53%
Love the home team favorite by 3 when the model predicts higher. I also like the fact that Tennessee is 3-0 and Minnesota 1-2. Too bad the confidence measure did not turn out to be higher. The Titans' key to its wins has been its defense that ranks third in the NFL overall and blending that with an offense that suddenly looks competent with veteran Kerry Collins operating the controls. Minnesota has Gus Frerotte connecting with ex-Bear Bernard Berrian and Mr. Peterson running more than 100 yards in every game. Makes you think twice about this be, doesn't it? I'm going for it, Tennessee at home wins by 10.

A good excercise to start analyzing the point spreads

Here is an idea that will help you get started thinking about which games are over/under valued in the coming week. Look at each game, but do not look at the point spreads! See each team's record, who they have played against, the defensive, and offensive rankings and make your own judgement of what the spread should be. If you follow the NFL weekly spreads, your estimate will probably be really close in most games and not more than 5 points in most. Make a note of those games whose spread are the farthest from what you think they should be. Then, watch my picks and see if one of them coincides. These games will probably be your best bets.

2008 Week 3 NFL Picks

NFLpickles.com is in a pickle. So far, my NFL statistical picks methodology have not given you any correct picks. I am disappointed as much as you are, but do not lose hope. If things do not start to turn around after week 5, that is, if my record from week 5 to week 16 is not above 60% then NFLpickles should cease to exist or at lease reinvent itself. Be patient, weeks will pass by, data will come in, and the record should improve. I will separate my ATS record as prior week 5 and post week 5 and you will see the power that a robust NFL statistical model can do.

This week I am using the model very cautiously. Also, I am using the "pick confidence" measure my model produces more freely, meaning, that even at 55% confidence, I would consider the pick. Also, I am weighting recent games much more heavily than I usually do in later weeks. Prior knowledge of team roster changes is also used for this week to determine how accurate predictions are. I am confident in this week's picks, I feel a comeback. Here are the Statistics Beats the Point Spread Week 3 picks.

RESULT: 3-1

Pick 1: Cleveland +2.5
Unbelievable that Cleveland is not even favorite in this game. Do you remember last year's Baltimore team? With rookie QB Flacco starting for Baltimore is hard to tell, but he wasn't so impressive in his first game (1 TD, 51% completion, 64 QB rating). Baltimore has played only one game this year, against Cincinnati, and won by just a TD. Cleveland on the other hand, has had a much tougher schedule so far; losing against Dallas (10-28) and barely to Pittsburgh (6-10). Baltimore has good defense and Cleveland has yet to show they can score, so I do not see Cleveland blowing Baltimore away, but probably winning the game. This might be a better Under bet, but I have not statistically explored Over/Under bets just yet.

Pick 2: Atlanta -5Kansas City is a mess. They lost by 15 against Oakland at home, do not have a quarterback, and their running game is not producing. This spread would have been a lot higher had Atlanta not lost by more than 10 to Tampa Bay last week. Tampa really shut the Falcon's running game, something Kansas City will not be able to do. Talking about Tampa shutting the running game, see my next pick. Anyways, model predicts Atlanta by 7, I am still confident, I believe the young guys in Atlanta will come out swinging.

Pick 3: Tampa Bay +3
Kyle Orton is just not going to cut it. How many games does the guy have to suck so the Bears realize he is not NFL quality? Bears field position will deteriorate since Tampa will be kicking the ball to the punt retuner, not Hester. Bears should not even be favorites in this game. Model predicts Tampa by 3, but they are underdogs by 3. This is what's going to happen: Tampa stops the run (Mr. Forte is good but cannot do everything, teams needs to have a passing game), Orton can't throw, defense gets tired, Tampa wins.

Pick 4: Dallas -3
I probably should not publish this pick. First, the confidence is low and secondly, Green Bay has a new quarterback. If one assumes that Rogers is as good or less than Favre, the pick is good. If Rogers keeps playing the way he has, better than Favre last year, then the pick is of no use. This is a big roster change that my model does not take into account, just yet. The only reason I am sticking with the pick is that Rogers has yet to face a playoff contending team. Minnesota and Detroit are not quite the challenge that the Packers will face in playoff contention. Dallas had a great match up last week and they looked great out there. I like Dallas this year, it might be their year, we'll see.

2008 Week 2 NFL Picks

Before I give you this week's computer model picks, here are a couple of reasons we should wait at least until week 4 or 5 before using my picks:

• Predictions follow team specific trends
• Last year's data, although weighted less than this year's, has significant impact on results
• Roster changes are not taken into account completely until week 6

Having said this, I have to get myself out of the whole. I am going to make a couple of picks this weeks for games where team rosters have not changed much AND the model finds a good opportunity in playing. Still, I encourage everyone to just follow them and take no action until week 5. These are:

RESULT: 0-2

Jacksonville -5.5
The model is catching this game as an over-reaction of the market to last week's Buffalo's win. Buffalo's run over the Seahawks was unexpected, but is it that Buffalo is that good or that Seattle is overrated? The Jaguars had a rough week 1, but they have a solid team and are still favorites in their division. They play well at home and came out last year with a 3-2 ATS, but one was a 2-point win against Pittsburgh which they were favorites by 2.5. They lost against Miami in the pre-season, so if this is accounted for in this point spread it shouldn't be. The last time these two teams met was in November 25, 2007 when the Jaguars beat the Bills at home 36-14.

Pittsburgh -6
Are the Steelers back to their Super Bowl shape? Maybe. Are the Browns over estimated this year? Could be. My statistical model is picking up this game, I think, because the Steelers have a good ATS record against the Browns and because the Browns are coming off a couple of losses, in their past games. The pick predicts Pittsburgh by at least 9 points, and historically when a home team is underdog by more than 6 points, 55% of the time in the past 3 years, they don't make the spread. The graph below is an interesting one. It shows Cleveland vs. Pittsburgh games for the past 7 years. Of 15 mathups, Cleveland has won 4 ATS (including the last one though). To read this, looking at the last game, you can see Pittsburgh was favorite by 9, but Cleveland cleared by 3 since the score was 31-28 Pitt.


Good luck!

Week 1 Results: Ouch!

You must be disappointed as much as I. I have not given up on my algorithm and I hope you have not given up either. It is too quick to use the computer picks, it needs a couple of weeks of data in order to take into account roster changes - LESSON LEARNED! Still, I made the picks and that puts me in a rough 0-3 ATS start for the 2008-2009 NFL season.

2008 Week 1 NFL Picks

The data is clean, the statistical model is ready, and the picks are out! In previous years, I would not make picks during the first few weeks to give time to gather data. This year, I am going to put my butt on the line and make picks from the beginning! The reason being that handicappers as well as I, have the same disadvantage, we do not have data for this year. We do not know how the off-season changes have affected the team's performance. So we have to put together a bit more error into the estimates, use last year's data carefully, rely on our knowledge of the teams, players, coaches, and the organization as a whole.

Our advantage is that we do not have to make a picks or point spread estimates for every game, we just choose the ones that look like a good opportunity. This week I have 3 free NFL picks for you based on my robust statistical predictor for NFL games.

RESULT: 0-3

Pick 1: Detroit -3
Again, it is probably wise to wait a couple of weeks, especially because both teams have new faces in key positions. Atlanta has their 3rd round pick QB Matt Ryan and RB Michael Turner, both excellent prospects, but will they come out swinging or will they ease into the league? Who knows, but I just do not see Atlanta starting the season with a win.

Detroit on the other hand has veteran Rudi Johnson who together with rookie Kevin Smith, can build up Detroit's running game (that is if Rudi gets his luggage back). Detroit's WR are back and QB Jon Kitna will be hitting quality targets in Calvin Johnson and Roy Williams. Detroit's offensive line and coaching staff is more stable.

Looking just to 2 years of data, my model found this game as a bargain. Unfortunately, 72% of bettors agree with me according to covers.com. I like to swim against the current but betting against Atlanta is tempting and even more so when a computer tells me to, you can't pass this opportunity.Both teams ended last year with a 0-4 ATS record, but I do believe Detroit is a much better team and 3-point spread against Atlanta in Atlanta is just no enough, 67% confidence.

Pick 2: Cincinnati -1
There is no Chad Johnson anymore, he is legally now Ocho Cinco. Que bueno! If 85 is focused and ready to play, I see Carson Palmer, who will play Sunday despite suffering a broken nose during the preseason, hitting hisfavorite targets Johnson and T.J. Houshmandzadeh. They have an untested Chris Perry starting at RB who, if healthy, can be hard to catch.

Baltimore did terribly last year against the spread. Will the market adjust for that this year? It does not seem to be doing that this week. A team that was 19% ATS last year playing at home should be a bit more underrated in my models opinion with 63% confidence. Go Bengals!

Pick 3: Houston +6.5
A dangerous pick, but I will stick my guns although the wager amount should not be as big as the previous 2 games. Pittsburgh is coming from a 3-1 preseason with Ben Roethlisberger showing signs that he could come back to his Super Bowl years performance. Houston on the other hand, has new coaches and an army full of young ambitious young players. Tough pick, but the model expects the Steelers to win, but not by more than a TD with 61% confidence.