The other two losses, not really much to say, Jets blew a big lead in a really shitty game, and the Jags lost their quarterback although they were getting beat down already at that point, plus you could argue that had the Jags O-line been better they would have protected Foles there, so that should be part of the bet and it's not really a bad beat. As the Chiefs D sucks and much of the performance was essentially in garbage time, I would not assume Minshew is going to be good based on his ridiculous stats in that game. I would also avoid over-reacting to the Dallas offense as well although I think they will be better as they finally figured out they should copy what the good teams are all doing, and they have always had good personnel.
As for this week we have the Panthers who are basically just a model play against the TB defense, which is not very good anyway and is missing several players. The one thing I don't really like about this game is that Cam Newton might be slightly hurt, plus I front-ran this game at -6, so at -7, this is a worse bet. However, Newton is always hurt, meaning that is probably reflected in his rating.
The MIA-NE game is obviously the most interesting contest here. The problem with this game is that most NFL statistical models are going to assume that teams are trying to win that season, which means that especially early in the season, all teams are going to be regressed to the mean a little, and big blowouts under-weighted. The nature of the NFL rules with the salary cap, coach turnover, training camps and limited rosters, copycat league, etc, ensures that teams really can't ever be that bad, assuming they don't want to be, until later in the year. However, the Dolphins are obviously not trying to win this season as they are spending only about half of their cap allocation on their active roster (the rest is dead money due to trades/cuts and cap space), and even of that half, about 25 million is on bad contracts that they probably tried to front-load knowing they were doing nothing this year. Other than the Browns when they were tanking, and not covering games either, there isn't much precedent for this, as we have a team that is basically running out a roster of replacement level players and rookies with a few okay players still around by accident, mostly on defense.
With our assumption of mean-reverting "trying to win" thrown aside, which is valid in almost all early-season NFL games as probably 95% of such games played since 2012 were played by teams that were realistically trying to at least make the playoffs, we know that the number of 12.5 is only a lower bound on how many points the Dolphins should be getting. At 14 or 14.5 I would certainly never touch this game but going across the key numbers of 17 and 18 I still think this game is worth the gamble, even against the Patriots. I personally never bet against the Patriots if it is close, but I don't think they are as bad to fade in games with large spreads, as they tend to do some weird stuff at the end of games with the lead (I recall getting screwed by this last year).
The other three games are pretty standard "fade teams that looked good or bet on teams that looked bad in week 1" plays. There is also a pretty good angle around Denver in the early season at home against teams not yet in shape, although they looked like garbage last week, so really fun to have to bet on them.
]]>The model for the early season is similar to last year except I added in regular season wins data to the model which led to a modest improvement.
Historically the "worse" teams by last year's record and underdogs have done well in week 1. Also, the home advantage is slightly larger early season, which is probably because there are no divisional rematch games, where the home advantage is known to be smaller, and maybe also because all teams have more hope early on so the fans are more involved.
In JAC-KC, the play is mostly based on the talent level of the two defenses. Of course we use Madden to measure this defense but even if you think Madden sucks, just going by cap hit, KC has the second lowest cap hit on defense in the league (OAK is lowest) while JAC is in the top third. The difference between KC and OAK is that they spent their cap dollars on good offensive players plus they won the draft lottery at QB, so KC is favored on the road at an okay team, while OAK is getting points at home to DEN.
The other three plays seem kind of gross but honestly I just punch the stuff in to the model and only veto plays when there are super good players hurt. There is not much difference between the Jets and Buffalo, Avery Williamson out will hurt the Jets, but I still have them just slightly better on D and flat with Buffalo elsewhere so we lay -2.5 there. Chargers are a "obvious public" play that I did not like before the Luck injury. Finally, the Giants suck and the Cowboys paid Zeke his money so obviously they are covering -7.5 in that one.
Good luck to all this season.
]]>As one would expect the market is pricing the games pretty well by this point in the season and we only have a 2 game card this week (I am laying off the GB game no matter what):
The Pats looked like shit last week so we are of course on them, while we will continue the 49ers fade. To be honest I almost passed the Pats as I think I am over-rating Brady at this point which would push the play down into the 13s where the line is, but with the motivation advantage and coming off a loss with Belicheck as coach I think this is still good enough for a small bet.
As for the other play it does appear like this Mullens can play football but the 49ers still have the worst running game in the league as well as a terrible defense. Kind of hard to believe the Bears could be 11 points better on a neutral field but I just go with what the model says (and it loves defensive talent which the Bears have). Good luck to all this week.
]]>We are now moving into the silly season where many teams have little to play for other than pride and next year's contracts. While motivation does not mean all that much it still is of some importance. As it turns out, we aren't fading it in any of the three picks this week.
I thought this was going to be a one-pick week on Monday but some heavy action has come in throughout the week that we will be fading. Certainly there are angles that would lead one to support Cleveland as they have been playing better since firing Hue, they have a rookie QB that looks good and has picked up experience throughout the year, and Denver is really banged up on offense. But at some point there is a price where Denver is worth supporting and especially getting the motivation advantage on the other side of 3 I think they are an easy choice here.
The Seattle play is really the same story. The 49ers are also a team that has been playing better since Mullens was brought in, and have a slight injury advantage against the Seahawks who are also coming off a short week. So we are taking Seattle as the price is just too low.
As for Atlanta, the pick is a little worrying as they have under-performed most of the year, with us on them many of the games although to be fair a couple of the games were bad-beat type losses. I am not sure how this team, that has been fairly healthy with Matt Ryan at QB, is only 4-9. Well, I guess I do know, because they have managed to have the worst defense in the league. Probably a good bounce-back candidate team next year if they fire their coaching staff. Both of their coordinators are total clowns and since the head coach chooses these guys he probably should go too. I would only take this at less than 10.
]]>Three plays this week. One fading the McCarthy firing, one betting against SF because they suck, and finally, we will have to sack up and take Oakland as there is just too much value to be had with them this week.
I might go ahead and enter this Punt Analytics contest. Unfortunately I don't think there is much they can do other than ban punts, which they can't and won't do because it would make games less competitive plus the players union will hate it. But it's a very interesting data set.
]]>This week we have three plays, none of which have any narrative other than "the model thinks the line is wrong". But in general it looks like we are fading last week's results, as well as "motivation", which can't be the worst thing. Good luck to all in your NFL betting and remember to avoid any pesky political discussion during Thanksgiving; save that stuff for gambling Twitter.
]]>Anyway, thanks to those two easy locks offset by the bad beat on the Panthers we are now up a couple of units on the season. For those of you backing these plays with max bets you should be up a quarter milly at least with your 2.15 units of profit:
This week we have an action-packed two game card:
The Atlanta pick is mostly just an injury-based play as Atlanta has been getting healthier while Dallas has a lot of guys either out or questionable, especially on defense where I tend to value injuries more highly. Atlanta has been almost historically bad on defense and if you take the numbers at face value the current line of a little more than 3 seems about right, but defense tends to regress to the mean much harder than offense, even this late into the season.
As for the Jags this is more of a motivation-angle fade as their season is basically over at this point while the Steelers seem to be picking up momentum after their strong win last week. The injury list for the Jags is a mile long although it is inflated a little by injuries to backups and special teams players, who hopefully don't matter much. In other words, every reasonable narrative in this game favors Pittsburgh so we are taking the Jags.
As for the Le'Veon Bell situation, basically irrelevant and the Steelers are smart not to pay him. Unsurprisingly James Conner has had a monster season although hopefully he doesn't play this week. While RB doesn't matter, at some point there has to be an impact, due to lack of familiarity with the starting role, if nothing else.
Good luck to all with your NFL plays and remember that this late in the season, the line is probably right, so don't bet much money.
]]>A small three game card again this week, all plays that could be described as marginal at best, other than the Carolina game which unfortunately has already moved a point and a half off the open of +5. Don't really see any reason why I wouldn't back this team, they keep covering usually in games they dominate yet are priced as an around league-average team. Very odd we like them every week.
The other two games are basically injury/small-sample size fades. Washington is very banged up and while the model prices that in to some extent you could argue that the compound injuries on their offensive line may suggest a larger-than-normal adjustment for injury which I did not make here. As for the Giants, the rating here assumes that Mullens is of a similar talent level to Beathard; it would favor the Giants more otherwise. One game against a terrible team that has quit on their season shouldn't change one's opinion of a player that much, he is probably still a third-string level player, but as always we will see next Monday. Both of these plays I would play for a below-average bet if at all, especially the Skins.
]]>As promised I went back and compiled the record for all these picks:
So through 6 weeks we are up a happy meal or so at 0.65 units, through about a good morning's worth of college basketball plays. If you shopped for lines, which for the love of god you should do in the NFL, I would guess you would be doing a little better, maybe up an upmarket happy meal at say Five Guys or something.
This week we again have what is frankly a pretty thin card with only two good plays:
I didn't forget the Chargers are playing on a neutral field here, and this is a pretty interesting game for one reason: the Chargers, coming from the west coast, are playing a 9:30 AM EST game, against an east coast team. This could be a pretty significant Circadian advantage situation favoring the Titans probably even more so than the well-known west coast team traveling to the east coast for a 1 PM start situation. In my book Picking NFL Winners: A Statistical Handicapping Primer we study the circadian advantage, and find that over the past 15 years, teams have had an advantage of about 0.29 points per time zone of circadian advantage plus an additional advantage from travel distance. However, since 2012, statistically there has been no advantage at all from either. It appears that teams have adjusted their practice and travel schedules to adjust the body clocks of their players to better prepare for these games.
Still, there is an argument that while one week is enough to move the body clocks of the players 3 hours, it might not be long enough to move the body clocks 6.5 hours, as is the case here. In this case, you might lean towards the Titans by an additional point. I still would play the game due to several questionable players on the Titans and the fact that Mariota seems to be injured, which would make his performance level lower than the model suggests, but it is closer. Another point favoring the Chargers is that they played an early game on the east coast last week, meaning the differential is not as large as it seems.
In the other game we like the Bears at home to the Patriots, although this assumes Mack plays. I never thought we would bet the Bears this season but since they got beat last week at Miami against Miami's backup I guess that sets us up for a bet. Usually I just avoid all games where a key player is questionable but it sounds like Mack will be good to go this week and is just sitting out practice for precautionary reasons. The Patriots haven't really been all that good this year, the Bears have, and if it wasn't for priors the Bears would be an even stronger pick. There is always the argument that the Patriots always start slow and could be rounding into form, but it is important not to overthink things too much.
As for other games if you like/need action WAS is another okay pick, while TB is favored by the model, but I think should be passed because they fired their defensive coordinator this week. While I have not done the study on coordinators, the impact of in-season head coach firings is HUGE - these teams play absolutely terribly even compared to what are typically very low expectations.
]]>I also managed to have MIA last week. At least I didn't end up watching the game, and in fact didn't even realize I had the play until I checked my sportsbook balances. Oof.
Not much action this week as there are mostly lines I agree with, then one play that is off by almost a full TD which is weird. The Browns are certainly playing a little better with Mayfield and I do think the number I have here is a couple points too high but playing against shitty teams mostly at home is always going to pump up the stats of any new QB. Their defense has been fairly impressive but that is not being ignored in the number here. A stiffer test this week to be sure although the Chargers are a little banged up. As for the Cards, basically just a math play there.
Notable this week is OAK who have under-performed my preseason projections by more than any other team. I guess bad coaching and old players who already have contracts really doesn't win in the NFL. Who would have thought.
For the teaser players out there there are three solid Wong plays out there right now in TEN, PIT, and IND all in the 2.5/ 3 -120 range that should cover more than enough even if you have to lay -120 on a two-teamer. I don't really like TEN so I laid off that but if you are one of those "efficient markets" guys you certainly should fire and I am sure most of you have pounded those already by now.
Good luck to all this week.
]]>A four pick card; right now we are still in the point of the season where we are taking the same teams week after week, and fading some others. Sometimes that is horrible, and sometimes it is not. In any event we are coming to the part of the season where stats are starting to make a difference to the projections, meaning there will likely be fewer plays in the weeks to come as we begin to use much the same approach that the market and sportsbooks seem to use.
The Arizona pick is probably the most interesting; obviously they are not a good team at all but they are extremely healthy while the 49ers are banged up. They also at least have a semi-ok defense, and the Madden raters think Josh Rosen, who fell in the draft apparently because he is politically left-leaning (amazing), is pretty good compared to Beathard. How can we go wrong?!
]]>An attractive slate this week. We will hope that OAK can finally play a second half of football, or at least three quarters. This is a team with a lot of talent, and one might wonder if they are being outcoached in the second half. Of course, this would mean that their coaches who have been very successful in the first half somehow lose all competence in the second. For now we will chalk it up to variance.
Finally, we won't be counting this toward any records but a JAC-HOU teaser is a nice choice this week. Unless you have to lay a crazy number like -130 or more, Wong teasers are still very viable, as long as you stick to "real" 7.5-8.5 favorites and 1.5-2.5 dogs, and not games with heavily shaded lines. If you need a third leg, for example if your local offers 3-teamers at +180, teasing up TB to 9 is not that bad given they are a strong +3.
]]>Amazingly I won money on my own betting having taken NO in a Wong teaser. With CLE kicking the extra point to go up 1 with a minute left I had to be drawing dead with NO -2.5 but the next three kicks went miss-make-miss, probably a 1 in 10 shot meaning they instead won by 3 and the other two legs (MIA +9, Jac +8) covered as well. I had some other winners that I didn't tout here due to QB injury questions or line moves.
Anyway I don't have any interesting write-ups on these plays and instead I am just going to post the plays and lines here. Teams I pick are in bold.
The Rams/Chargers game is a rare intracity matchup. Travel time and time zone / circadian effects used to have a big impact on NFL HFA but my studies have shown that teams have figured out how to travel better and these no longer matter. Still, there is reason to believe that there should be a lower home-field effect for teams playing in the same city.
Having bet the AFL (Aussie Rules) in the past, intracity matchups are common in that league as something like half the teams play in Melbourne. In these games the HFA is 2 points compared to 8-10 in intercity games (closer to 10 when a team from Western Australia travels to the east coast of that country and closer to 8 in other games). To me the AFL is close enough to NFL football that I don't mind going with this, meaning we will go with 1 point of HFA here instead of the usual 2.8. If you don't agree, I would pass on the Chargers.
]]>Week 2 of the NFL season is all about not over-reacting to week 1. We should still be using around 85% of our preseason projections at this point, as one game just doesn't mean that much for sides. With totals, one game is a little more important because it gives us a better idea of how certain teams, particularly those with new coaches or schemes, are likely to play on offense when it comes to pace of play and run/pass mix. But we still hedge our bets heavily with the totals as well, because while we adjust play style for context there is still plenty of variance even in statistics such as seconds elapsed per play.
On to the plays (worst line to take in parenthesis):
Buffalo +7 -105 vs Chargers (+6) -
Obviously Buffalo is coming off maybe the worst game anyone has had in the NFL since ever and there's every chance we could see another 2016/2017 Browns scenario. However, other than at quarterback this roster is not really all that bad, with league-average talent on defense. Getting 7 at home against the Chargers who are missing two key players on defense is just a little too much to pass up. Another week of worst-ever offensive play and the model will probably stop picking Buffalo as it wasn't exactly high on them last week, but we stick to them for now. Let's hope that they can turn things around, or play terribly and get some turnovers; either will work.
Dallas -3 (-4.5) vs New York Giants
This is not really a wager on the Cowboys so much as against the Giants who appear to be one of the worst teams in the league particularly on defense. The Cowboys actually played an okay game last week and have a slight talent edge across the board here which combined with the home advantage gives us enough value for a wager.
Dallas / Giants over 42 (over 44)
Neither of these offenses are likely to light up the scoreboard and Dallas is particularly awful for totals as they run the ball more than most. However both of these defenses are terrible, and the total is just too low for a dome game.
Chicago / Seattle over 43.5 (43.5 only)
The Bears have switched to a pass-heavy offense while Seattle is very weak on defense this year and may also be favoring the pass more than in the past with a depleted offensive line and running back. The loss of Baldwin will hurt a little but there is still enough value here for a small play, although pass if the winds kick up above 15 mph.
Steelers / Chiefs under 53 (under 49)
These are two poor defenses, particularly the Chiefs who are missing Berry, and obviously the Chiefs put up a ton of points last week. However, both of these teams actually play a slow-paced style, and I have the Chiefs as the slowest-paced team in the league, running down the clock by two seconds per play more than the average team, a trend that continued last week. Although it sounds like he will play the knowledge that Big Ben is banged up further helps the under here. This is just a huge total and you have to really buy the hype that the Chiefs are one of the best offensive teams ever to take the over, in which case you probably have to bet the Chiefs as well.
]]>As for why anyone should bet money on something totally ridiculous like Madden ratings, I'm not sure I can answer that question well other than to say that Madden ratings would have crushed the market early-season in past years, so maybe they will win again this year. What won't win is anything using last season's stats at all - every model I have tested that included Madden ratings had no use for past stats, suggesting the ratings are pretty good.
There aren't many games in the NFL each week and the lines are usually pretty sharp so we could easily get crushed but here are the picks for Week 1. All of these are posted on Pickmonitor as well. The number in parenthesis is the worst number I would take on the game.
CLE +4.5 (+3.5)
IND / CIN o47.5 (o49)
NE - 6 (-7)
CAR / DAL o42 (o44.5)
DEN / SEA o42.5 (o46)
DET / NYJ o45 (o46)
OAK +4 (+3)
Although all four totals picks are overs, I liked the under in the PHI game tonight and had a small wager on that, and may consider the OAK and LAC unders depending on wind conditions as well, so I think the adjustment I have made is more or less in line with the market.
Also, below are my season wins projections. Note that these are based on week 1 rosters only, so for some teams like PHI/TB who should get their quarterback back later in the season, as well as teams who play PHI/TB twice, they will change a little. Although I have wagered on a few of these, I would not recommend doing so as you would be taking the worst of the number on most of the teams I like given I moved the line already with some of these moving further after I bet. They are provided for educational purposes only. In any case, sloppy seconds are never a good thing in sports betting, or in life.
]]>