Haven't done the figures for the race yet, so these are general comments.

Jockeys say stuff like that all the time. On rare occasions it might even be true. But having dealt with trainers (and jockeys second hand) for many years, I know better than to just accept a statement like that at face value. When we look for dead rails (and we do for dirt races) we look for data to back it up, not anecdotal evidence.

We have some trackmen who are very sophisticated handicappers (most notably in Chicago), and flag horses on days they believe there is a dead rail-- but then I look at them when I do the day to see how their figures are stacking up with their previous figures. We also have a program in the computer that flags horses that ran on the rail (in this case on turns only, because that is what the computer gets, for ground loss) on days that at least 2/3 of the horses that ran on the rail finish in the bottom halfs of their fields-- and again, I then go and look at the figures those horses ran relative to their histories, and decide whether to give out "x's". In fact, I just did a dead rail day yesterday at Churchill (second turn only), first one I can remember there, but I might be wrong about that.

I have never heard of a dead rail on turf before, but that doesn't mean it doesn't happen, and I'll take a look at this day when I do it. When I worked with Michael Dickinson he was a nut about this kind of stuff, and on occasion walked the course before Da Hoss ran (and in at least one instance had his girlfriend Joan walk the course in high heels). When I was at Ascot this Spring an amazing sight was a 30 horse field split into 15 horse groups along each rail going down the straightaways-- the jocks feel the going is better along the fence. No idea if they are right, but certainly agree that conditions can affect one part of a track differently than others-- 1 and 2 turn split variants being an example.

There are a lot of practical problems with the idea of making different variants for different paths. First, horses seldom stay in one path all the way around, which by itself would make any relative speeds hard to quantify. Second, the number of horses running in any path is relatively small, decreasing the sample size you would be using to make a path variant, and the track often changes speed during the day, which could in theory affect different paths differently. Third, in dealing with dead rails I have never found that anything resembling a common correction exists-- horses just quit, don't fire, etc.

The good news is, if there is a difference in speeds for the different paths other than the dead rails, it is relatively minor. The tightness of the data base itself indicates that-- the number of horses who pair up, run in a tight range, etc.



TGJB