Correlation vs. causation

[TempletonPeck]

As I succumb to Derby fever, I have been taking a trip through the archives and am pondering this:

In some seminars, TGJB points out these stats: horses with more rest into the Derby (specifically 4 or more weeks vs. 3 or fewer weeks) are more likely to run a new top, and horses with fewer preps are more likely to run a new top. These two things are pretty obviously correlated.

But isn\'t it entirely possible that the reason a horse with only 2 preps, and 4+ weeks into the Derby is more likely to have a top in the Derby is not because of the rest, but rather because that development and maturation was happening and we didn\'t know about it? Said another way, because his current top isn\'t as good an indication of his present ability as is the current top of a horse who has raced more recently, because by virtue of its age, the data isn\'t as good?

A hypothetical: 2 theoretical horses, both are entered in the Derby. Assume everything else about them is identical.

Horse A has run twice this year, 2/2 and 3/3. His race 2/2 earned a 6, and his race 3/3 earned a 4.5.

Horse B has run three times this year, 2/2, 3/3, and 4/4. He earned a 6, a 4.5, and a 3.

The Derby is on 5/5. Who ya got, and why?

Add Horse C, who has run 3 times this year, 2/2, 3/10, and 4/16, a 6, a 4, and a 2.

Now?

[RICH]

all things being equal, the horse with the highest odds