Quote:
Originally Posted by RichieP
Friday 10/30 was the day the music died for "auto" killing virtually a whole months profit in one session.
There is one big problem with this and it's been coming. I cant continually be buried day after day where I need a 20-30 buck horse to get even. Yesterday it never came.
I recognize this pattern from the 90's when a friend and me analyzed for days the patterns of higher paying winners relative to our reads.
I will stop the betting here on this for now, take a small profit, bet enough to watch videos a few months.
Now it is time to see whats what with a couple of things. Now the hard work begins 
Ted if you would update the table thing you have that would be great!
Screen shots of yesterdays debacle  and final overall totals for the month |
Richie,
I see you had a bad day, but I still see you with an .11 ROI. Am I just reading this wrong? Its not my place to interfere with anything you're doing, but I am certain of one thing - if you're looking at this as a research project, a sample of 277 races is far too small to have any validity. This doesn't mean that your autopaceline model also couldn't benefit from tweaking, but with this small a sample, the jury is still out. If you start tweaking now, you'll never know the true value of this particular autopaceline method. If you do want to make some changes and start over, try to do a sample of at least 1000 races before drawing any conclusions about the strength of your model. If you're familiar with the poster 'sjk' on the PA site, he told me that his profitable model didn't stabilize until he had a sample of 2,500 races, but hasn't deviated thereafter for the past decade.
There's another page out of Benter's book that also seems worth consideration, and tfm alludes to this in his reference to the homogeneity of the horse population Benter was handicapping - as part of the test of your model (or simply handicapping) make sure that you're playing at tracks you can beat. I'm convinced that both the economics of racing, the decline of interest in the sport, and the growth of whale betting has killed the profit potential of certain tracks. The more races you do, and the more you analyze the data, the more these tracks will reveal themselves, and allow you to eliminate them from play. Any place that could be renamed Chalkdust Downs, and there are a number of them, is not going to be a profitable venue for this kind of autoselection method, or even for normal Sartin handicapping.
Another thought I had is that publicly posting your results by the day may cause stress that you don't need.
Best of luck,
B Jennet