I think it would be handy to be able to select more than one line for an individual horse-that way, you can see how it performs using several lines and take the best. This is an old trick I used to use in Sartinb's Kgen - run 3-4 lines, look for an aberrent line and throw it outm then pick the best one of the remaining.

If I wasn't doing this, I'd probably be doing something else.

I like your idea. On any horse use two or three, or none. Pace make the race, just like Ralph said on another thread.

If you wanted to use, say, 3 pace lines for a horse, wouldn't it go something like this? OK, for the first pace line, which pacelines on all the other horses in the race do I want to compare it to? Then, for the second paceline, which pacelines on all the other horses in the race do I want to compare it to? And on and on ... ?

The simulator would do all this after your pacelines have been picked. And spit out the top four.

The old Sartin programs were hand entry, so I did not put that many individual lines over and over again. What I did was to match up the horse against himselve, looking to see if one of the lines was totally out of wack with the others - the output was a graph, so it was easy to see-similar to the speed graph. Secondly, i would use the best, non-wacky line. Using the simulator would be a big improvement over this because you could load 3-4 lines for several horses. If the 2 horse finishes 1,2,4,5 in the race, you have a good contender-right there on multiple lines. Likewise, if a horse finished nowhaere off of 4 lines, he is a throw out.

If I wasn't doing this, I'd probably be doing something else.

Now to get it implemented. sure would beat a single paceline. almost throws out false favorites.

I have to admit that I'm not 100% sure about what folks are exactly talking about, but I haven't read things too carefully.

I've been wanting to run such a test to see if it could improve certain aspects (sim winner) over the various pace line selection processes we have now. But I haven't had the time.


The "architecture" of the code currently prohibits a quick solution to allowing some sort of multi-paceline selection. That is, it will be a pain in the ass to alter. Here we are not (I don't think) talking about the ability to select _A_ pace line by looking at multiple pace lines, but to use the values, whether median or average, from the combination of multiple pace lines.

I'd bet against the efficacy of using multiple pace lines. Certainly the current auto-paceline (and our human pace line) selection mechanism is sometimes deficient, landing on a pace line that is just way different than what the horse will show today. However, and I'm sure you've all heard this before, horses don't run on "averages" - they run in "peaks and valleys". Still, it would be interesting to see how the "averages" play out (or the medians for that matter).

In many races, esp. turf and more often dirt route than turf, pace truly does make the race. Here I don't mean that pace is more important in this type of race vs. that type of race. I mean that most sprint races are run the same way by each horse every time. That E-Horse just isn't going to "control" the pace the way a router can sometimes control the fractions of a race. He's either got it in the tank, he doesn't, or there's a major amount of early speed that's going to do him in.

So, pace doesn't as often make the winner as it "makes the race"... It "makes" the way the race is going to shape up. How fast is everyone going to be running at each point in the race. Given this, I'd like to see improved paceline selection procedures that takes these factors into accout. But, as with everything horse racing, its a tough problem. The data is disparate, the runners seldom true to form, and the "right" data may not exist (you want a Route pace when all the entrants have only a couple of sprints in their pps).

I'll work on the multi-paceline selection stuff, but no promises as to a completion date.

-N

"I'd bet against the efficacy of using multiple pace lines. Certainly the current auto-paceline (and our human pace line) selection mechanism is sometimes deficient, landing on a pace line that is just way different than what the horse will show today. However, and I'm sure you've all heard this before, horses don't run on "averages" - they run in "peaks and valleys". Still, it would be interesting to see how the "averages" play out (or the medians for that matter)."

Nathan,

One of the reasons I don't use your excellent program is the inability to use multiple pacelines. Multi-paceline handicapping is certainly different than single paceline handicapping, but I wouldn't bet against its efficacy. IMHO, it's a losing bet.

quote:

---

Originally posted by Jake:
"One of the reasons I don't use your excellent program is the inability to use multiple pacelines. Multi-paceline handicapping is certainly different than single paceline handicapping, but I wouldn't bet against its efficacy. IMHO, it's a losing bet.

---

ok....why can't you use multiple pacelines in ES? I do it all the time. Unless you are talking about the program doing it all for you? You just go into the form view and choose the paceline you want to use. I do have questions about the auto selection by ES of some pacelines.....today at Tampa the program chose a paceline that I was not happy with, and I don't know why it did what it did. But it did make a difference. I put the card up at this site if anybody wants to look at it. http://home.columbus.rr.com/justralph

Race 2 horse number 5. Soul of diablo.
Todays race was a 5.5 sprint. The program chose a paceline 3 back, prior to a layoff. All three of the last 3 races, were 6F affairs. I don't know why it chose the 3rd back race. 3 back was the best race this horse had run in a while. It probably has to do with some filter I have set or something I am doing wrong. If I had gone with the default program selected paceline, then the 5 horse shows up in the sim 5 lgths ahead of the eventual winner of the actual race. Not sure why it chose that race, but I would be curious to know why. On the eventual winner (4horse) it chose the last race for that horse. It was also a 6F race? Like I said, its probably something I set.

Just Another Horse player

"ok....why can't you use multiple pacelines in ES? I do it all the time."

I'm not talking about swapping out different single pacelines and trying them on for size. See Nathan's "average play out" comment. For example, no current way to do average of last 3 or best 2 of 3 in Equisim. I'm not saying those are better paceline approaches, but they do have certain efficacies...largely odd-related and surface-related functions.