I rather think it might be, when you start taking wildcards and so on into consideration: for example the Tigers might reasonably be expected to win fewer games than Texas but still be a shorter price for the WS.
You’re asking to reverse-engineer the process; what odds-makers should be doing is starting with W/L records and working it out from there.
I’ll have a go, though…
Comment by Aaron (UK) — December 18, 2012 @ 10:32 am
Yup, it’s too difficult to do in half an hour on Excel, that’s for sure.
Howver, some additional percentages do fall out of the data:
Sum (AL teams) = 52.41%
Sum (NL teams) = 47.59%
This doesn’t seem unreasonable.
Sum (AL winner from Central) = 23.8% – so we can already see the “equal chance in the postseason hypothesis” can’t possibly be consistent with these odds
Sum (AL winner from East) = 42.1%
Sum (AL winner from West) = 34.2%
Sum (NL winner from Central) = 27.7%
Sum (NL winner from East) = 38.2%
Sum (NL winner from West) = 34.2%
In short the odds are expecting one wild card team from the East & the West in both leagues. This is hardly a massive shock.
Comment by Aaron (UK) — December 18, 2012 @ 10:50 am
Also, the author’s method for generating true prices (which sum to 100%) tends to over-penalise the good teams.
My preference is to raise the raw probabilities to a power (which you find by trial and error). In this case it is c. 1.09645, and that, for example, leaves the Blue Jays at 8.99% and drops the Astros all the way to 0.30% [still not low enough, but…]
Comment by Aaron (UK) — December 18, 2012 @ 11:01 am
Is it possible to short-sell in sports markets? I can see the Dodgers’ odds being much lower in July.
Comment by Well-Beered Englishman — December 18, 2012 @ 11:12 am
Yes, on betfair, but not if you live in the States.