09 Aug 2006 18:48:31
Sample Data?

Hello, I've reading around the web and at my library at different
racing systems, and I would like to try these out with some past data
and see how I would do if I were to try these out at the track.

However, I don't want to pay the 200 bucks to Equibase to get access to
the database results there for a month. Anyone have some sample data I
could get to play around with some of these systems?

Thanks,

Dan
nospam_yoblin_nospam@gmail.com



24 Aug 2006 21:31:19
DrHuang.com
Re: Sample Data?

You can save data from the software RacePredictor.com, and download the
software from www.racePredictor.com
.
<yoblin@yahoo.com > wrote in message
news:1155174511.571525.209460@m79g2000cwm.googlegroups.com...
> Hello, I've reading around the web and at my library at different
> racing systems, and I would like to try these out with some past data
> and see how I would do if I were to try these out at the track.
>
> However, I don't want to pay the 200 bucks to Equibase to get access to
> the database results there for a month. Anyone have some sample data I
> could get to play around with some of these systems?
>
> Thanks,
>
> Dan
> nospam_yoblin_nospam@gmail.com
>




05 Sep 2006 15:19:33
Uber Micro
Re: Sample Data?

"DrHuang.com" <DrHuang@com.DrHuang > writes:

> You can save data from the software RacePredictor.com, and download
> the software from www.racePredictor.com .


You should put AD, if you are advertising your own paid software


> <yoblin@yahoo.com> wrote in message
> news:1155174511.571525.209460@m79g2000cwm.googlegroups.com...
> > Hello, I've reading around the web and at my library at different
> > racing systems, and I would like to try these out with some past data
> > and see how I would do if I were to try these out at the track.
> >
> > However, I don't want to pay the 200 bucks to Equibase to get access to
> > the database results there for a month. Anyone have some sample data I
> > could get to play around with some of these systems?
> >
> > Thanks,
> >
> > Dan
> > nospam_yoblin_nospam@gmail.com
> >


Of course building a system that works on past performance isn't hard
to do. Check out the law of non-falseafiablity.

To check properly you would have to do some form of crossvalidation to
see that the system tends to generalize. The best way to check a
system is to just keep track of races that will happen from tomorrow
and beyond, get enough results to give you a satisfactorily small
confidence interval, and off you go.


--

Always Interested in finding other math-minded players.
Feel free to email uber_micro at husmail.com. If you are
selling something please indicate by using the Prefix AD: