Robot II

[Corey]

I'm getting some inconsistencies with the results in Robot II when I change the surface to wet dirt.

If I run "Learn X" after making the surface change under "What's Winning" it is missing the majority of the higher percentage factors vs what I'm seeing when I run "Learn All" with the same wet dirt surface.

[Corey]

Just to add to the above post.

I've noticed the issue for Wox 1st race on Wed 11/11/09 as well (where no surface change was made)

When I run Learn X - I get the following three factors under "What's Winning"


What's Winning?
------------------------------------------------------------------------------------------
2TS 00015 00004 27% 33% 53% 1.50 0.78 1.01 $11.2 00000 $16 2.56
L2 00018 00004 22% 33% 56% 1.25 1.04 1.07 $11.2 00000 $16 2.06
RES=2 00012 00002 17% 50% 58% 0.65 1.12 0.78 $7.9 00000 $9 1.54


If I run Learn All - some of the factors show higher winning percentages.

Favorites 00009 00002 22% 44% 78% 0.62 0.90 1.07 $5.6 $7 2.10
Fr1=2 00008 00004 50% 63% 100 2.95 1.60 1.66 $11.8 $16 4.63
F/X=1 00008 00002 25% 50% 75% 1.56 1.25 1.12 $12.5 $16 2.31
F/X=2 00008 00002 25% 50% 63% 0.69 1.04 0.89 $5.6 $7 2.31

My question is why aren't the factors of Favorites, Fr1=2, F/X=1, F/X=2 appearing under What's Winning when I run Learn X?

[km]

Corey,

LearnX has minimum sample (Plays) size that need to be met before it displays the item.

If you find a high win% in Learn All, check the number of plays to see if it is light play count (15 or less), those won't show up in LearnX. Also, LearnX has I.V. and ROI minimums in some categories.

LearnX is trying to find the factors that are most likely to produce going forward rather than just a raw stat list. Use Learn ALL for pure hit rate regardless of play count.

If you find a discrepency between the statistics in Learn All and LearnX with wet tracks though, let me know.

[Corey]

Ken - thanks for the reply. That sample size limitation would explain why I'm seeing a factor appear in one and not the other.