endless requests.....

[del_dog]

Ken

I was sitting around with nothing to do on a slow day and wondering if you thought it feasible to have a "last race on surface" selection under the angles section?

A way to set a last race surface of poly or turf or dirt.

Then we could use the robot to check for poly to dirt - turf to poly etc.

In combination with the connections section , I think it would be handy.

This can be gotten from the export files now of course.

If there is a way to do it now inside the robot I am too dense to see it?

thx

[Sheets Guy]

Ken

I was sitting around with nothing to do on a slow day and wondering if you thought it feasible to have a "last race on surface" selection under the angles section?

A way to set a last race surface of poly or turf or dirt.

Then we could use the robot to check for poly to dirt - turf to poly etc.

In combination with the connections section , I think it would be handy.

This can be gotten from the export files now of course.

If there is a way to do it now inside the robot I am too dense to see it?

thx


Tremendous idea. I think many bad faves can be eliminated this way.

[km]

dog

We had "Surface Switch" in an older version of the HTR2 Robot. It was not effective at all and was replaced by the far more potent Razor Sharp workout.

I see that you really want to use it as negative though. And that i would consider: NO Surface Changes". I'll test it and let you know if it is valuable in the contest of the Robot.

[del_dog]

Ken



I had in mind to be able to check a specific surface in the race section and then check last race on - dirt or poly or turf - in the angles and then run the robot to look at all the various outputs for anything that popped up before going to a more in depth analysis.



Thanks for checking it out.

[km]

gotcha Dog

that would require 12 individual new filters using the 4 distinct surface types in HTR

Dirt to Turf
Poly to Wet
Turf to Poly
etc.

It takes me about 4 hours to implement a single filter and another 2 hours or so for follow up testing to make sure it working correctly. In this case, all 12 are necessary, so the labor time is a large investment for filters that have not shown high impact returns.

In fact, most of these would have a negative outcome. So we would then need 12 additional filters such as =

NOT Turf to Dirt
NOT Wet to (fast) Dirt
etc.

to capture the opposite for eliminations.

Any ideas for one or two filters would be considered, but a dozen or more new ones is out of my time frame right now.

[del_dog]

Ken

I think I am with you now. I think anyway.

Would JUST a box to select the last race surface to be tested suffice?

Then the existing various generic surface categories would contain the results of the tested for type of surface from the angle check box.

The stats accumulated in the robot results would just contain the races that matched the "from surface" of the last race.

So, if I checked last race turf and ran the robot ,I would have only races with a last surface of turf going into the robot run.

The same counters and statistics as now would be shown.

Does that require the extra set of buckets to hold the results?


I would not want to request something that would require the effort to get all the specific buckets set up.

Or, is it that to hold the results would require the buckets even if the only races selected for inclusion in the robot run were for just the last race button that was selected?

[km]

Yes, you are correct with that last statement 'dog
It's all or nothing with this surface-to-surface change items. It would be quite a chore to code for all possibilities.

The Robot is a work in progress and has no completion date, it's something i'll consider in the future for sure.

[del_dog]

Yes, it doesn't sound feasible at all.
I can pull it out of the export files.

thx

[Gollygeewiz]

Hi,

As to the turf to dirt races - my db shows that they would model nicely, if only those Races were incl in the modeling.

I feel it is essential to separate OCL from Clm, my (non HTR) db shows the following diff in hit% and ROI between these 2 type races only looking at 3u Male Spr, under 7f, Field Size 5-7, for one trk:

Fin Pos
1 2 3 4 >4th Public Choice
OCL H% 29 14 14 14 29
ROI -31 -56 -24 +13

CLM H% 41 27 18 9 5
ROI -20 -13 -16 -53

I dont believe that there is a workaround in the system to do this without a mod - IS THERE - but the results for all tracks would be very very nice. There are such a large # of these races, I am surprised that you dont have this in place.

I was very happy to see the Field Size flexibility.

Another suggestion - instead of check boxes turning factors on - would it be practical to give us a choice - on/elim with those ck boxes?


THANKS, golly

[km]

Gene, yes, you can use HX4 Export to query (model) virtually all data in HTR including surface changes and Class Types in your database software.

It has text of all Class Types and Last Surface raced to handle all those items.

Read HX4.TXT in the HTR LIbrary.

Keep in mind one of the key trade offs between the Robot and a database.

Simplicity of use with limited data access ( Robot )
vs.
Full access to all data, but an extensive learning curve (db software and exporting)

[DanG]

Keep in mind one of the key trade offs between the Robot and a database.

Simplicity of use with limited data access ( Robot )
vs.
Full access to all data, but an extensive learning curve (db software and exporting)
Those trade offs becoming less and less btw. In fact; when all aspects (time off our own clock) are factored in the Robot2 wins hands down imo.

I like knowing Access; but not an Access user I’m aware of can take the software out of the box and start effectively using it on day one. The Robot2 is so logical, inviting and intuitive you have to keep reminding yourself just how powerful it is.

It’s odd; the Robot2 is so straightforward it almost makes some Access users suspicious. We are so used to walking through mind fields to solve issues (those of us without Donnie to teach us) and a child could learn the Robot2 in one day. That is meant as a MAJOR compliment btw; the needless complexity to some windows products can be very frustrating…at least for myself.

[Bob]

Ken,

I am not sure how hard this would be to do on your end but I have always wanted to look at horses that dropped dramatically from their mlo to tote odds.

One way to do this would be to have another filter box on R2 where a user could choose ml odds of say 6-1 and then tote odds of <2-1. It might be interesting to see how they do plus we could do analysis on what factors may have caused the money to go to the horse.

Just a thought.

Bob G

[km]

Bob, that's one you can easily run a db query on - and lots of us have. No advantage at all to the ML bet in half. Part of the problem is the obvious though - the ML is distorted with scratches, and some ML are just terrible to start with, so the betting action is really more of a correction.

Maybe Dan or someone will run the stats for you. QUERY: nAODDS <= (nMLO * .5)

Check with Pizzaman's ATR program for gathering real betting action information live.

[njcurveball]

Ken,

I am not sure how hard this would be to do on your end but I have always wanted to look at horses that dropped dramatically from their mlo to tote odds.



I hope this formats well. I broke out the results by ML odds. My Db is 2+ years worth. My ROI is for a $2 bet.

No real benefit for checking the final odds against the ML. Losses are just about the track take any way your slice it. It does help place and show profits beat the track take though.

Query for all odds < ML


RMLO STS WPCT WROI PROI SROI
---------- ---------- ---------- ---------- ---------- ----------
1 68753 36 1.63 1.71 1.75
2 50110 26 1.65 1.7 1.75
3 46125 20 1.6 1.64 1.72
4 41944 16 1.63 1.63 1.69
5 33056 13 1.6 1.62 1.68
6 24418 10 1.59 1.6 1.64
7 16527 9 1.64 1.63 1.62
8 9456 8 1.7 1.59 1.59
9 7128 7 1.73 1.64 1.61


Query for all odds < (ML * .75)



RMLO STS WPCT WROI PROI SROI
---------- ---------- ---------- ---------- ---------- ----------
1 49419 40 1.64 1.74 1.77
2 29630 29 1.67 1.74 1.77
3 26575 23 1.63 1.67 1.74
4 24192 19 1.64 1.63 1.71
5 18769 15 1.64 1.67 1.72
6 13549 13 1.61 1.63 1.69
7 8849 11 1.66 1.63 1.64
8 5219 10 1.77 1.65 1.67
9 3667 8 1.56 1.64 1.65




Query for odds < (ML * .5)



RMLO STS WPCT WROI PROI SROI
---------- ---------- ---------- ---------- ---------- ----------
1 22631 46 1.66 1.77 1.79
2 9870 35 1.66 1.77 1.78
3 8784 30 1.7 1.73 1.76
4 8324 24 1.68 1.7 1.75
5 6680 20 1.63 1.67 1.75
6 4824 17 1.62 1.65 1.73
7 3101 15 1.72 1.74 1.75
8 1812 14 1.73 1.7 1.73
9 1251 10 1.46 1.54 1.58

[del_dog]

and I think it really varies by track.

At some of the smaller tracks , if you know the guy doing the ML then some of them can be helpful to drop out the obvious losers.

[Bob]

Thanks everyone for your input and DB runs...looks like it is a fruitless pursuit.

Ken, have you found any value to the HTR number size? In other words, do HTR horses greater than 70 win more than 65, etc.

Regards,
Bob G