With the year's edition of the Football Outsiders book being released in various forms recently, now is as good a time as any to compare FO's win projections from previous yeras to the Vegas lines. Here are the average misses of both over the last three years (see last year's post on this for a bit more context):
Despite missing badly on their riskiest projection ever (FO had the Packers winning 11.4 games; Vegas said 8.2), 2008 was a great year for their win projections, as they improved their accuracy against the O/Us for the third straight season, this time beating the lines by a good margin. This was possible because, while they may have been wrong about Green Bay, they were on the right side of the Vegas line for the next six biggest discrepancies:
It's worth noting that, even if their average miss is the same as the Vegas lines, the FO projections should hit better than 50% on these, since they just need to be on the right side, even if the Vegas projection is closer (the Eagles are a good example of this).
Six for seven is a bit extreme, but it's far from the first time FO has done well against the O/Us. Overall over the last three years FO is 55-41 (.573) against Vegas, and if you weight the larger discrepancies more heavily, it jumps over 60% (depending on how you weight them).
In the coming weeks we'll take a look at what this means for various 2009 odds. Just looking at the divisional odds, one team really jumped out.
6 comments:
I may be missing something, but how was PFP "RIGHT" about PHI and TB? They missed by 2.7 for Philly (Vegas missed by 0.2) and they missed TB by 1.3 (Vegas missed by 1.2). Isn't the "right" or "wrong" determined by how far from the actual value their prediction was?
Any plan to update the standings of the baseball contest where we picked the Over/Under on win totals with confidence levels? Thanks!
"I may be missing something, but how was PFP "RIGHT" about PHI and TB? They missed by 2.7 for Philly (Vegas missed by 0.2) and they missed TB by 1.3 (Vegas missed by 1.2). Isn't the "right" or "wrong" determined by how far from the actual value their prediction was?"
If I say the Browns are going to win 16 games, and the O/U is 5.5, I still win the bet if they go 6-10, even though the line was way closer than my guess.
I see, so you were treating those lines as though they had bet on the Vegas line.
Still, it seems a little confusing to talk about the average misses, then say that PFP was "right" when their prediction was very far from the actual result, and Vegas was almost spot on.
Yeah, well I tried to explain that in the paragraph following the "RIGHT/WRONG" table...apparently I didn't do a very good job.
Maybe if you dumbed it down to "WIN/LOSS" our friend mark might understand a little better.
Or he could just read the whole post...
Post a Comment