Tester and Access

[LOU M.]

I would imagine this would be to cumbersome for the tester but could it work in Access? I would like to have the same output as the tester with an additional filter "by Trainer".It might be interesting to see how trainers perform in these catagories at various tracks,surface etc.BTW Ken, any news from Ron Tiller on the new trainer stats?

Thanks, Lou M.

[km]

Lou -

I'll send Ron an email and ask him to reply about the trainer stats formation.

[LOU M.]

Thanks, I'm getting ready to reup my subscription now that I can utilize HTR a little better. Any suggestions on how I can accomplish the above post?

Thanks, Lou M.

[Ron Tiller]

Lou,

No news. I can't implement the new stats (the programming for which is substantially complete) until I finish 3 other big projects and I can't get those projects done because computers keep crashing, software keeps breaking and daily responsibilities and a multitude of other smaller tasks always stand in the way of finishing those projects. So they have been in limbo for 9 months and continue to be.

I'm sorry I said anything here that raised everybody's expectations of imminent new data. I won't be opening by big yap again till finished.

Ron Tiller
HDW

[hurrikane]

Lou,
you can do this..I'm afraid the easiest way would be to program it but maybe with queries.

I'm afraid the sample size will end up to small to be of any use though.

[km]

Lou - re: using the Tester for trainer stats

It's possible, but cumbersome. We would need a way for the user to insert the name of of the trainer as a filter before running the test. There are two ways to do this =

1) manually type in the name of the trainer
2) choose the trainer from a list

If I place an input box on the Tester interface right now, you could type in the name of a trainer and it would filter the output for that one trainer's horses only. The problem is the exactness required with the typing. Although I could write code to remove the necessity of upper and lower case letters, the user would still have to get the spelling and spacing exact.
For example: "Von Hemel, Don" ; "Von Hemel, Donnie K." Father and son would have to be typed in perfectly, including commas and spaces and periods to match the Equibase exact spelling. We could try testing on the basis of fewer letters -- but if we set it for 6 letters, and we type in "Von He" to get Von Hemel, we would end up combining all the Von Hemels.

The 2nd possibility is to create a list to choose from. This would be a massive alphabetical selection box and would be impractical with thousands of trainers to choose from. But this idea would overcome the spelling and typing problems.

With either approach, the output from the Tester may not be that useful. How much are we going to learn from Von Hemel with a few horses that were Fr1 = 1 or K=3 for instance. The workout rating would seem to be the most valuable to test with trainers from those stats, but most of the other items are not that significant or would have very small samples.

You can tell I have thought this out before and want to help but none of these approaches appeals to me for trainer stats inclusion in the Tester. It is much easier to run this sort of data from Access using the trainer name from the export.

[LOU M.]

Actually I was going to use this as an elimination tool.Say a trainers ROI is close to break even and if you could slice off 10% of his losses it would hit positive territory and possibly become a play for Pinnacle.Maybe a trainer never wins when he is sixth or lower in K. If the educated members don't feel it's worth while I will defer. On the other hand if it has merit any way to implement it is OK with me.A note on sample size. I agree when you are testing random factors in an uncontrolled setting a large sample size is prudent.In the case of trainer patterns consistency is the key. Trainers are specialists just like us,speedcappers pacecappers etc.when they find a successful formula they repeat it.The trainer has done all the testing and now you just go with it. Even trainers with large barns have limited opportuities for their patterns to succeed. A trainer may only go 2 for 7 in a year but does it year after year with $20 mutuals. They're there and I enjoy the research.Thanks for everyones input,Lou M.

[tomcat]

I think it might be more import to know where a trainer is winning....Lake is tough on East Coast tracks, but not so good on mid west tracks. Some trainers do well in the mid west but not so well in the far west, and visa versa.

[km]

Lou, when you are subscribing to HTR again, please email me, we'll set you up with some way to filter a specific trainer from the Tester.