The article is primarily about how much money from various betting channels finds its way into the purse account, but I was more interested in the discussion of Tony Zhou\'s computer-centric approach to betting. Two things that struck me:
He does not know why his computer likes a horse, just that, at the odds, it does.
He is gleaning signal from the time at which money is bet on specific horses.
https://www.nyra.com/aqueduct/news/between-the-hedges-where-do-purses-come-from
You can\'t knock the guy\'s success so far, but the big question is: how will he handle a losing streak? We all have them. We\'ve all made adjustments that made things worse. How we respond at that point is critical to how long we stay in the game.
Really bad losing streaks are more common for those using human based skills (humans go on tilt) or if the game changes and he doesn\'t figure it out in time (not likely)
Fair point, but there is probably more variance and volatility in his day-job than at the track. I have followed him on twitter for awhile - he has written some very interesting posts regarding the management of risk/reward and the search for +EV vis-a-vis betting and the markets during that time.
Tony Zhou was on the most recent Sport of Kings podcast. It\'s a sobering listen with respect to how tough the competition is. It also offers some insights relevant to the \"Odds Change\" thread. Part of the edge possessed by computer players (as robotic teams or otherwise) is their insight into how races will be bet. Zhou talks mostly about win odds on the podcast, but I have to guess that his database also gives him insight into other pools, including those not displayed by the tote.
https://www.sportofkings.net/podcast
Thanks for the link. That was a good show and I loved the topic.
Be nice if he didnt sound like he was 200 feet underwater
Aw, the first 20 minutes of the podcast are the best thing under water since Led Zeppelin\'s \"No Quarter.\"
\"Walking side by side with death, the devil mocks their every step.\"
Only thing I noticed was that Scott should probably switch to a sativa, but the indica is hard to put down. :-)
Bit,
Thanks for the link. I respect Tony and know a friend a good while back that spoke with him at length about what he does. Much respect.
That said, if you go back and look at the charts and the PP\'s as they discuss at least the first two races they covered (I didn\'t have time to continue when I listened earlier this week), you can see where his computer thoughts are not accurate on odds either for every race. In fact, some major errors. He did have a good opinion or two where I may not have but his computer misses some things that the human eye and mind might pick up quicker or process better. For example, one horse that he discussed what his computer would say involved a horse who has done well with lasix and not as well without it. He didn\'t account for this difference if I had to bet in his algo. This is an angle that still has some use as these rules continue to be inconsistent as lasix is phased out. I recall discussing this a touch with JB and Sean at Saratoga this summer.
There were other places where I saw that his comments and projections maybe indicate his algo misses a mark or two. I like he didn\'t want to mention the weakness of the whale computer players but did say where less info is a place they are weak basically. It makes the mind work on where there are still a few spots available.
I do agree over time as he says a computer can be more accurate in some ways. But they are not infallible by any means. Opportunities are far, far less as we all know esp at larger odds which makes the game less appealing in many ways. This fits with the idea that the lowest goal for them is to grind out a profit after the rebates are sent their way for their large handle.