Making figures for new tracks
 

So with Arizona Downs and Century Mile opening as new tracks soon, how do you go about setting figures? Reliable par times would be essential whether you're a speed capper or pace capper but where do we begin? Do you wing it for a while, then at some point you decide you have a large enough sample size to set pars, variants, etc?

Seems to me the entire first meets are just a guessing game and there's a huge edge to be gained for anyone willing to do the work and learn these new tracks from their genesis. Just a thought.

Hey Deelo

Would you look at Trainers who had success at nearby tracks to see if maybe that would transfer over to the new tracks?

Given that you have no numbers to go by at the new locations maybe you take a look at what the Trainers are doing

Just a thought

Not so much looking at trainers, but how to form a pace figure out of scratch.

Perhaps some type of translation by comparison database? Using made-up numbers here, like something along the lines of TUP horses that run a 77 pace meeting SUN horses with the same 77. When they run at Az Downs, recording that and starting to form sort of a figure by comparison? Like maybe I set TUP as the par so a 77 there becomes a 77 at Az Downs, then with more info I start translating it to where a 82 from SUN is a 77 at AzD, a 68 from Zia is a 77 at AzDowns, etc...

Not sure if that makes sense, I just know there's tons of smart programming and figure people on here that could lend good ideas.

I'd consider posting this question in a different forum on the site. The Oracle has sort of taken over this forum as his own so most people don't look at it. There are a lot of smart figure makers here (including CJ, one of the experts in the field) but they're probably not seeing this post.

Hey Deelo

Please post on this site or any other that you choose I haven't taken over anything or called it my own

If you want to post here feel free your ideas are always welcome

Please don't feel that I am stopping you or anyone else to post

Thanks

I personally don't use pars at all. I use projections from the horse's previous races. Even for new tracks, the horses have to come from somewhere and I'll have numbers for them.

The tougher part is establishing a baseline for a speed chart. I believe Century Downs is going to be a mile oval and pretty standard. I'll make figures and see if any distances look quirky, check that run ups don't very too much at different distances, etc. Arizona Downs is the old Yavapai and I think the track is the same. Not necessarily speed wise, but dimensions.

I wait until they have run for three weeks so there is some data to work with.

For the conditions that have been run I copy pars from a track that I would expect to be similar (same state usually).

Those pars and the race times are used to make crude par times. With par times and the three weeks results I can make variants for the races run; my variants use pars but 95% of the information used is projection from previous races.

With all of my variants I recalculate every few weeks to take account of more recent events so with these crude variants I would immediately recalculate all of the pars, par times and variants a few times to converge them to more accurate values.


Since I have no interest in betting Az tracks (interest level in race betting not what it used to be; there are not enough playable races and edge to make it worthwhile) I will probably not bother with the new tracks this time.

Had to do this quite a bit over the years for on-going tracks when they changed the layout or made major changes to the surface and it has generally worked out well.

thanks guys, those were both helpful insights.

very well may not be playable but never hurts to be prepared just in case. if i happen to come up with any kind of meaningful model once they start i will post.

The additional challenge with Century Mile - the track it is replacing - Northlands, was a 5 furlong bullring. Some of the jockeys go elsewhere during the winter - Turf Paradise, etc, but many jocks sit idle. Some have little experience riding a mile track. The perennial leading rider -Rico Walcott is out indefinitely due to some serious health issues.

It could take a while to get a feel for Century Mile

Also, the weather has still been dicey in Edmonton - snowed there on Saturday so the track could be slow...but maybe not..haha

You could compare some of the early season sprint times to the early season sprint times from last year (remembering that 6f would have been 2 turns at Northlands) and try to gauge how much quicker these 1 turn races are...start with that as a benchmark, and go from there....

Good luck!!

I couldn't agree more about the value of using projections over pars. Abilities and times can vary considerably within classifications and pars pay no attention to the form cycle of the horses in the race. When doing projections form is vital in projecting what figure a horse will run that day in addition to the individual abilities of the top finishers in that race

I would start out using the figures I have for the horses that have been running at other tracks to make projection figures. I'd use some kind of standard distance chart.

Once I got going, I'd pay close attention to whether certain distances are running faster or slower than expected based on that chart. If so, I'd start tweaking the chart.

I would also create class PARs based on similar tracks in the area and the ongoing data as I accumulated it.

I think class PARs are useful. We are all going to have some biases in our thinking that will tend to create figure drift when we use projections alone. Looking at PARs will help anchor that so your figures won't start getting progressively faster or slower.

I honestly wouldn't even know where to start with making pars any more. There are so many conditions and so many class levels it would take literally a decade or more to get any kind of real sample size.

I just received Cynthia par book. They have pars for AZD, CTM, and Cnl.

I assume the rating for AZD and Cnl are from Yavapai and the previous meeting at Cnl.

I will use until proven they are unreliable for the current track surfaces.

The major tracks aren't that hard. I look at the most commonly run classes for older horses and makes those. For some classes you may not have enough of a sample, but after awhile you'll know where the class fits relative to some that you do have. So you can make an estimate for those.

For 2yos and 3yos you can break it into 3 month periods. You'll usually get enough data to make some reasonable estimates if you have multiple years of races.

The really oddball classes you aren't going to have data for and it's probably not worth worrying about, but if you read the conditions and look at the purse relative to other classes on the circuit, you'll usually know approximately where it fits.

The problem is not doing it for the circuit you play on all the time where you sort of already know the pecking order of statebreds, claimers, open and how they move back and forth. The problem is the dozens of circuits you are starting fresh with where you don't already know what's going on and there are a lots of odd ball classes. That's a monumental task.

are there official ratings for horses in your country?
most countries have them these days.

if there is, then you would no longer need to worry about class types/conditions, because you could solve the issue with simple regression analysis.