Another week of football is in the books and we’ve reached the ‘official’ quarter mark as every team has played at least four games now. There were some good games this weekend, although I didn’t watch as many as usual since the Lions were on the bye (took the opportunity to run some errands); I was particularly pleased to get home and see that virtually every early afternoon game was within a touchdown at the end. The afternoon games were a bit more lopsided, but that’s ok when Brady and Peyton are there to bring some entertainment. Let’s see how the models did at picking the games.
The models continued to spit the bit on the over/under. Luigi was 5-9 for a season record of 24-35-2, Yoshi 1 was 6-8 for a season record of 26-33-2, and Yoshi 2 was also 6-8 for a season record of 26-33-2. It’s about this point where someone will say, why not pick the opposite of what the models say? That’s because there’s really no reason to believe that the models are purposefully picking the wrong way, and so what we’re seeing is chance guessing that, so far, happens to be below chance. Luigi, which is doing the worst of the bunch, is not significantly different from 50% and has a confidence interval of about 30 to 56%. Essentially, even at this point there’s a chance that Luigi is actually above chance but has been on a crummy streak. I wouldn’t go that far, but I would expect the models to go about 50/50 the rest of the way. For the morbidly curious, Luigi would be statistically below chance if he goes 2-12 next week. 3-11 or better and 50% is still in the 95% confidence interval.
I pointed out before the games started that the Colts should hang with the Packers, and look how that turned out? Luigi did pretty well picking outright winners overall, going 11-3 to bring his season record to 32-29. Yoshi 1 was 8-6 for a season record of 33-28, and Yoshi 2 was 11-3 for a season record of 33-28. If you’ve been paying attention, you may have noticed that Yoshi 1 virtually always picks the home team to win. That obviously can’t be the best strategy long-term, so I expected Luigi and maybe Yoshi 2 to start moving past. Perhaps the replacement officials did have an impact in those first couple weeks, or maybe it’s just dumb luck that Luigi went 20-9 the last two weeks after starting 12-20.
The moneyline is a little different; each model decides if a team has a better chance of winning than the odds suggested by the moneyline. So you aren’t trying to decide if a team will actually win, per se, but if a team has enough of a chance to win to make the bet worthwhile. Those did not go so well this week. Luigi was only 4-10, moving his disappointing season record to 16-40. Tending to pick underdogs will definitely lead to slow times when they don’t come through, but things need to turn around soon for Luigi. Yoshi 1 was 3-10 for a season record of 23-36 and is still a little in the black for the season, and Yoshi 2 was 3-11 for a season record of 18-39 and moved into the red this week.
My favorite part of the week is the spread. Picking every game, Luigi was 6-8 for a season record of 25-33-3. But looking only at the games Luigi had a good opinion for, he went 5-4 for a season record of 21-22-3. Yoshi 1 was 7-7 overall for a season record of 31-27-3, and in pickable games was 5-5 for a season record of 27-20-2. Yoshi 2 was 6-8 in all games for a season record of 26-30-3 and 4-5 in pickable games for a season record of 21-23-3. After a rough start, things are starting to pick up for Luigi. 5-4 isn’t anything to write home about, but it’s a positive.
Last time I started looking at how the picks go when all three models agree about which way to go. The over/under is still abysmal, and the moneyline picks were only 3-9, which isn’t too surprising since each of the models had an off week. In terms of the spread, the models only agreed on 4 games and went 2-2 in them (the Rams and Colts covered while Washington and Buffalo did not).
Looking at ‘pickable’ games is one way to be more confident about which games will go my way, and restricting things even more to the top five is another. That’s what I do for the Hilton SuperContest. Luigi picked Indy (win), Denver (loss), Buffalo (loss), San Diego (loss) and Minnesota (win). So a relatively disappointing week at 2-3 moves the overall record to 9-9-2. Yoshi 1 picked Jacksonville (loss), Indy (win), St. Louis (win), Buffalo (loss), and Cleveland (loss). That 2-3 mark moves the season record to 9-10-1. Yoshi 2 picked the same as Luigi but Washington instead of Minnesota, which was a loss. That 1-4 mark moves the season record to 9-9-2. It’s been rough going so far, but these picks typically do better. I expect them to catch up over the next 12 weeks.
As is tradition, we finish up with Bill Simmons. Bill went 8-6 against his lines for a season record of 33-26-2. Luigi went 7-7 for a season record of 27-32-2, Yoshi 1 was 7-7 for a season record of 32-27-2, and Yoshi 2 was only 5-9 for a season record of 25-34-2. So Bill is leading all the models so far.
Power rankings tomorrow. Have a good Tuesday!