Looking back at the win lines
I wrote a couple of posts this season where I compared the pre-season betting lines on how many wins a team would get in 2006 to where I thought they would. Short version: I went 7/8 for a 66% return on a hypothetical betting investment.
Here are the eight picks, with commentary, a review of the pre-season line predictions and results, and some criticism of the original piece.
* Angels, take the under at 89.5 (-130), I think the lineâ€™s off by nine games, you have to win 56% of the time to make money
Ding! Winner, but just, just barely.
* Blue Jays, take the under at 87.5 (-115), I think the lineâ€™s off by seven games, you have to win 54% of the time to make money
Ding! Another scraper.
* Cardinals, take the under 93.5 (-110), I think the lineâ€™s off by five games, you have to win 53% of the time to make money
* Giants, take the under 83.5 (+105), I think the lineâ€™s off by three games, you have to win 49% of the time to make money
Ding! Even worse than I thought.
* Mariners, take the over 75 (+115), I think the lineâ€™s off by six games, you have to win 54% of the time to make money
Ding! But they didn’t end up at 81.
* Mets, take the under 90.5 (+110), I think the lineâ€™s off by five games, you have to win 48% of the time to make money
Bzzt. So I thought they were a 85-win team, and they won way, way more than that. Ow.
* Reds, take the over 73.5 (-115), I think the lineâ€™s off by five games, you have to win 54% of the time to make money
Ding! They were even better than I thought.
* White Sox, take the under 92 (-105), I think the lineâ€™s off by ten games, you have to win 52% of the time to make money.
Ding! They were a lot closer than I thought, though.
If you lived in Vegas and hoofed it down to a sportsbook immediately, putting 100 on each of those eight, you’d have won 7 of 8. Payouts:
Blue Jays +87
White Sox +95
You would have made $529 (or so) in profit on your $800 investment.
If any of you did this, and I hope you did so legally, feel free to cut USSM in on your windfall by hitting that little paypal button on the upper right. I’ll wait.
So here’s the really interesting thing, from my perspective: how well did the line I quoted at the end of February predict the standings? And I’m just going to look at the over/under, and not the odds on each side.
So Angels spot on, Athletics were +4, the Rangers -2, the Mariners +3.
White Sox 92
White Sox -2, Indians -12, Twins +13, Tigers +18, Royals +1. Wow, did the line blow it here. What’s interesting, as I look at the odds, though, is that there was huge betting on the Indians under, the line moved to take the *over* on the White Sox, with an equally heavy tilt to the over on the Tigers, and… I don’t have data for the Twins. So it looks like the money was moving towards a more realistic view of the standings… sometimes.
Red Sox 92
Blue Jays 87.5
Devil Rays 68
Yankees 0, Red Sox -6, Toronto 0, Orioles -3, D-Rays +7.
Dodgers +4, Giants -7, Padres +10, D-Backs +3, Rockies +7
Cards -10, Cubs -18, Astros 0, Brewers +6, Reds +7, Pirates -8
Mets +6, Braves -9, Phillies +3, Nationals -5, Marlins +13
The line was way, way off on a lot of these, but so were my guesses, sometimes in opposite directions (the Mets, for instance, finished about as many games over the line as I thought they would finish below).
I based my guesses on 2005 RS/RA, the BP projections, the Diamond Mind work, and thought I had a really good line on what their performance would be. Once I had that, I’d look at the Jays, compare them to my list, and say “there’s no way they’re an 87-win team, pah!” If you think they’re an 80-win team, like I did, then the chances you’re going to lose your money are pretty slim – I said 4%.
This turns out to be one of the stupider things I’ve ever written, for a couple of reasons:
– no pre-season projection is great, and there are wide variations in a season’s outcome
– the teams that finish the season aren’t the teams that start it (particularly, even Diamond Mind’s simulations, which include injuries, can’t include trades)
Sure, if I knew in advance how well everyone would perform, who would get injured, who teams would call up and trade and bench and whatever – if I knew in fine enough detail a team’s path through the 2006 season, I could project their finish and their chance of exceeding it. But if I had all that information, doing projections would be pointless.
Or I could get MLB to play the same season 100 times with the same players at the same age with everyone’s injuries perfectly healed, that kind of broad generalized projection might more closely match the grouped results of those hundred seasons. This is equally stupid.
In any season, the results aren’t going to match up with any kind of projection. Even while I might be able to see things in a general sense, like “the Tigers are stronger than they’re being given credit” that still doesn’t make my prediction of x wins a certain outcome. In retrospect, of course, this seems obvious, and my “the line is seven games off” is embarassing to read now.
I’ll be more cautious next time.