Energy Distribution and the speed figure

[classhandicapper]

Originally Posted by elysiantraveller

You said this earlier and now you are drawing up "conditions" that would enable her to excel...

LOL

Hagler and Leonard were both great boxers, but in a 12 round fight instead of 15 round fight, using the largest possible ring to give SRL room to move, with gloves that take away the advantage of the bigger puncher, Hagler was asking for trouble.

I think the idea is to think in terms of neutral conditions.

[elysiantraveller]

Except thats not the picture you paint. Zenyatta is the greatest mare of all time was the statement and then you make the case to best be able to beat her contemporaries (Azeri, RA) she would needs faster males on the front to soften them up... sounds hypocritical... just sayin.

[PICSIX]

Originally Posted by Cratos

I haven't ignored anything; being it is Derby week and I will be out of town my time is very limited, but eventually I will give an example.

Can we please see the example?

Thanks,

Mike

[Cratos]

Originally Posted by PICSIX

Can we please see the example?

Thanks,

Mike

In the “Triple Crown Trail” discussion on this forum I published a ranking of the pre-race Derby projections.

However if you are waiting for a detail explanation of how the model works from an algorithmic point of view that “ain’t happening” and the reason being is that the algorithms that was developed to make the model work is proprietary and the research and work that went into the development has been long and is still ongoing. Also there is some discussion to turn this into a retail software program.

If you have questions and they are not invasive to the inner workings of the model I will answer them; you can send me a PM

[cj]

Originally Posted by Cratos

In the “Triple Crown Trail” discussion on this forum I published a ranking of the pre-race Derby projections.

However if you are waiting for a detail explanation of how the model works from an algorithmic point of view that “ain’t happening” and the reason being is that the algorithms that was developed to make the model work is proprietary and the research and work that went into the development has been long and is still ongoing. Also there is some discussion to turn this into a retail software program.

If you have questions and they are not invasive to the inner workings of the model I will answer them; you can send me a PM

It is pretty clear what I asked for from the beginning and you agreed to give one. It has nothing to do with pre-race Derby projections. If you don't want to give one, just say so and we can delete this thread now.

[Tom]

If I had known it was going to take 170 posts to get to here, I would have bought Snap-capper Pro and been done with it!

[yak merchant]

Originally Posted by Tom

If I had known it was going to take 170 posts to get to here, I would have bought Snap-capper Pro and been done with it!

Yes for a thread I checked almost daily to see if anything new was posted this thread might have ended up being the biggest wait of my time in all my years of reading PA.

[lsosa54]

Originally Posted by Cratos

In the “Triple Crown Trail” discussion on this forum I published a ranking of the pre-race Derby projections.

However if you are waiting for a detail explanation of how the model works from an algorithmic point of view that “ain’t happening” and the reason being is that the algorithms that was developed to make the model work is proprietary and the research and work that went into the development has been long and is still ongoing. Also there is some discussion to turn this into a retail software program.

If you have questions and they are not invasive to the inner workings of the model I will answer them; you can send me a PM

Can you give us some basic insight as to how you came up with your pre-race choices?:

Grits, the following are my choices:
1. Midnight Interlude
2. Derby Kitten
3. Archarcharch

[classhandicapper]

Originally Posted by elysiantraveller

Except thats not the picture you paint. Zenyatta is the greatest mare of all time was the statement and then you make the case to best be able to beat her contemporaries (Azeri, RA) she would needs faster males on the front to soften them up... sounds hypocritical... just sayin.

That's not what I said.

I said she proved she was among the greatest mares of all time on two surfaces and possibly the greatest of all time. I was then defining the conditions I think would be in her favor, neutral, and against her vs. some of the other great mares. IMO, you have to define neutral. All the great mares had distance and pace preferences.

To give a an extreme example, she might not beat a lot of best sprinting fillies of all time going 6F on a speed favoring good rail day on the sharp turned Aqu Inner.

[classhandicapper]

Cratos,

IMO one of the biggest issues facing "numeric handicappers" is that all horses are unique in terms of the relationship between their short term speed/brilliance and stamina even when their overall ability tends to be similar.

Everyone knows that. They classify horses as speed horses, stalkers, closers, even paced grinders, assign letters to define them, create numbers to measure them etc... even when they of the same class.

But then almost everyone uses either speed figures alone or performance figures than combine pace and final time that treat all these horses the same way in their formulas.

It's as if changing the pace from 46 to 44 will have the same impact on a lightning bolt of a horse and a more even paced horse just because they are very similar in overall ability (the same is true of all large variances).

It's as if a deep tiring track will impact the speedy faint hearted horse as much as the speedy horse loaded with stamina or that a closer without deep reserves of stamina won't be impacted at all.

Aside from accuracy issues, surface differences etc... I think this is where all the existing numeric models fall apart.

If you are working towards addressing that, it's likely you will have a leg up.

To date, the only approach that works for me is more class oriented comparative trip style handicapping and an analysis of the horse's overall record.

[Tom]

But then almost everyone uses either speed figures alone or performance figures than combine pace and final time that treat all these horses the same way in their formulas.

Where did you get that crazy idea????

[Cratos]

Originally Posted by yak merchant

Yes for a thread I checked almost daily to see if anything new was posted this thread might have ended up being the biggest wait of my time in all my years of reading PA.

Yak, I apologize and I didn’t realize that there was urgency for me to present an example. Therefore I have chosen the Schuylerville Stakes at Saratoga from last year to give an example; the example should be posted before the end of today. The reason for my choosing that race is that it is a 6f race where none of the horses have ever run 6f and I hope I will be able to illustrate the projected time at the greater distance.

[classhandicapper]

Originally Posted by Tom

Where did you get that crazy idea????

I don't think I was clear.

I don't know much about energy ratings/Sartin methodology etc... I know many people make pace figures for individual fractions, categorize horses, and look at different ratings depending on running style, but I've never seen a method that tries to incorporate each fraction into a performance figure that varies by horse even when the fractions are the same.

22 45 is treated like 22 45 no matter which horse did it.

But IMO 22 45 may be a jog in the park for one horse and a ticket to being eased for another even though they may finish in a head bob if the fractions were 22.3 46.1.

Many people try to handle this with pace figures and match up analysis before the race and by looking at charts and race development after the race, but I haven't seen a formula that says the impact of 22 varies by horse and calculates the performance figure or total rating that way.

If one exists, I'd like to see it so I can see if it's getting to the crux of the problem that I think exists. It seems like an incomprehensible problem to me.

[thaskalos]

Originally Posted by Cratos

Yak, I apologize and I didn’t realize that there was urgency for me to present an example. Therefore I have chosen the Schuylerville Stakes at Saratoga from last year to give an example; the example should be posted before the end of today. The reason for my choosing that race is that it is a 6f race where none of the horses have ever run 6f and I hope I will be able to illustrate the projected time at the greater distance.

No disrespect Cratos - I am always fascinated by your posts, eventhough most of what you say is "over my head" - but, IMO, a race which is YET to be run would be much more instructive than one whose winner is already known.

Based on my own experience, it gets considerably easier to predict and explain the chaotic nature of this game...after the race is run.

As Tom Ainslie so brilliantly remarked:

"After every race, another system is born."

[Cratos]

Originally Posted by classhandicapper

Cratos,

IMO one of the biggest issues facing "numeric handicappers" is that all horses are unique in terms of the relationship between their short term speed/brilliance and stamina even when their overall ability tends to be similar.

Everyone knows that. They classify horses as speed horses, stalkers, closers, even paced grinders, assign letters to define them, create numbers to measure them etc... even when they of the same class.

But then almost everyone uses either speed figures alone or performance figures than combine pace and final time that treat all these horses the same way in their formulas.

It's as if changing the pace from 46 to 44 will have the same impact on a lightning bolt of a horse and a more even paced horse just because they are very similar in overall ability (the same is true of all large variances).

It's as if a deep tiring track will impact the speedy faint hearted horse as much as the speedy horse loaded with stamina or that a closer without deep reserves of stamina won't be impacted at all.

Aside from accuracy issues, surface differences etc... I think this is where all the existing numeric models fall apart.

If you are working towards addressing that, it's likely you will have a leg up.

To date, the only approach that works for me is more class oriented comparative trip style handicapping and an analysis of the horse's overall record.

Indirectly my model addresses the issue that you have pointed out. The model takes 3 curves into consideration for pace: F-curve (front-runner), P-curve (presser), and C-curve (closer). Those designations are only for the model and are not distinct to any horse because pace can dramatically change style. The model solves this issue with exponential curve fitting. The trick is getting the negative constant, β correct.