Line Score Question - Pace and Cap

Ryan · 06-04-2022, 01:51 PM

Sort of an odd thing going on to me and I am curious for some input as to what everyone thinks.

My Churchill Downs dirt sprints model is showing 24 of 33 winners ranked 1 or 3 in the primary line score. Horses ranked 2 are just 3 for 33 in my rankings.

How would you navigate this? In the past year of using RDSS, I have never run into this sort of situation. My thought process is to essentially toss those that rank 2 on the primary line score. The only concern I would have is that more times than not, these runners fall into a dominant category in my model for dirt sprints.

MJS6916 · 06-04-2022, 02:49 PM

Quote:

Originally Posted by Ryan

Sort of an odd thing going on to me and I am curious for some input as to what everyone thinks.

My Churchill Downs dirt sprints model is showing 24 of 33 winners ranked 1 or 3 in the primary line score. Horses ranked 2 are just 3 for 33 in my rankings.

How would you navigate this? In the past year of using RDSS, I have never run into this sort of situation. My thought process is to essentially toss those that rank 2 on the primary line score. The only concern I would have is that more times than not, these runners fall into a dominant category in my model for dirt sprints.

without seeing all of the individual readouts,
I would treat the 2s and 3s the same until I found a dominant category like you said.

Still that's 81%+ in the top 3. If you're holding that rate, that's good!

I personally don't believe less than 150 events as a sizable enough sample
to zoom too much into. I know Sartin taught differently.

I guess I would search for the spreads between ranks 1-2, 2-3, and 1-3 .....
.....and compare it to your past models/results.
If you see the differences are more squeezed in this time around,
you might not have to make any more sense out of it.

Good luck and greater skill, Mike

Lt1 · 06-04-2022, 03:09 PM

What are the dominate factors[primary factors] where the winner ranks in your top2 the most. That's what I would depend on[and I do] regardless of tier rankings.
Tim

DanBoals · 06-04-2022, 08:36 PM

33 races is too small of a sample to not have oddities mathematically. The fact that you have 27 out of 33 winners in your top 3 is what you should be concentrating on, that is very good.

Howard told me that his original database that he based the methodology on was based on 10,000 races entered by hand. The primary line score that is used to create the BL/BL ranking is based mostly off what was learned in analyzing these 10,000 races on a DEC computer that Doc rented time on.

I would be willing to bet any amount of money, that if you keep handicapping exactly like you are now for another 167 races, you will find that you have slightly more winners in the second rank than you do in the third rank, and you will have slightly more winners in the first rank than you have in the second rank.

200 races is a large enough sample for something like primary line score or BL/BL to distribute normally under most circumstances. The software works. On a small sample size, you might see some funky stuff, but over any decent number of races, like 200, or a season at the track, the software does what it is expected to do as long as you do your part.

Ryan · 06-06-2022, 11:40 AM

Right on! I appreciate it. Just making sure I am on the right track.

I would agree too that the sample is much too small and it probably is just an odd thing going on. Wanted to make sure that I wasn't overthinking it.

Thanks for all the help guys!!

Lt1 · 06-06-2022, 11:54 AM

Don't be worried about sample. Doc also show using a proven math formula that what happens in 20 races will happen in a 100 races etc with only a 2 -3 % change in %. Not enough to be concerned with. FWIW I have proved this to myself many times. It's not necessary to carry the coals to Newcastle. It's more important to know what primary factors are dominate based on your handicapping.
Tim

DanBoals · 06-06-2022, 03:45 PM

Quote:

Originally Posted by Lt1

Don't be worried about sample. Doc also show using a proven math formula that what happens in 20 races will happen in a 100 races etc with only a 2 -3 % change in %. Not enough to be concerned with. FWIW I have proved this to myself many times. It's not necessary to carry the coals to Newcastle. It's more important to know what primary factors are dominate based on your handicapping.
Tim

Hey Tim,

When I was working with Howard back around 2000, he had 14 20 race printouts (using the Wager Decision Form) over a season at Santa Anita that showed a loss, so 20 races is not enough, even for Howard, to guarantee a winning outcome. As a math and computer guy, most of whose course work was in physics and whose real world experience is in writing technical software in digital signal processing, motion control and other math and physics intensive projects, I promise you, 20 races, or 33 races is not mathematically sound to prove anything. You can ask any statistics teacher, even at a high school level, and they will tell you that is too small a sample size. Since our software is based on a 10,000 race sample, we do not need sample sizes that large to recreate the wheel. Something along the lines of 200 races for someone new to the methodology should be fine. If someone has not bet 200 races on paper, IMO they should not be betting real money at the track unless they are pretty well off.

Dan

Lt1 · 06-06-2022, 04:25 PM

Hi Dan. I myself will do a max of 100 races and I know this is enough based on my work. Howard made the statement. I agree 20 races maybe too small but I've seen 50 races samples work well also. The trouble with large samples is they tend to give the user brain freeze. They get too confused when things aren't crystal clear. That's one of the reason[after hearing Tom Brohamer I now lean on the primary factor[s] where the winners in the top 2 the most times If I remember correctly the 10000 races were back in the day when they were using that[at the time] big college computer and we didn't have many computers on the market. I agree a new person should do as many 20 race cycles as needed to give them confidence and enough info for real betting. Trouble is most people folks won't do the work required.
Tim

DanBoals · 06-07-2022, 11:24 AM

Amen, brother.

"Trouble is most people folks won't do the work required."

That is one of the truest things anyone has ever said about handicapping.

There is a lot of work required to rise above the crowd in this sport. When you consider that the takeout is 15 to 30% depending on the track and kind of bet, you have to really know what you are looking at and really choose your spots. I have had some days lately where I handicap 6 races and then when they go off, none of them had the odds I need to make a bet. It is a hard, but that is also why it is so rewarding when you beat the horses and show a profit at the end of the month. Gotta put in the work, gotta pay your dues.

06-04-2022, 01:51 PM	#1
Ryan Maiden Join Date: Feb 2021 Posts: 2	Line Score Question Sort of an odd thing going on to me and I am curious for some input as to what everyone thinks. My Churchill Downs dirt sprints model is showing 24 of 33 winners ranked 1 or 3 in the primary line score. Horses ranked 2 are just 3 for 33 in my rankings. How would you navigate this? In the past year of using RDSS, I have never run into this sort of situation. My thought process is to essentially toss those that rank 2 on the primary line score. The only concern I would have is that more times than not, these runners fall into a dominant category in my model for dirt sprints.

06-04-2022, 03:09 PM	#3
Lt1 Grade 1 Join Date: Apr 2012 Location: Valley Stream NY Posts: 11,385	What are the dominate factors[primary factors] where the winner ranks in your top2 the most. That's what I would depend on[and I do] regardless of tier rankings. Tim __________________ Trust but verify

06-06-2022, 11:54 AM	#6
Lt1 Grade 1 Join Date: Apr 2012 Location: Valley Stream NY Posts: 11,385	Don't be worried about sample. Doc also show using a proven math formula that what happens in 20 races will happen in a 100 races etc with only a 2 -3 % change in %. Not enough to be concerned with. FWIW I have proved this to myself many times. It's not necessary to carry the coals to Newcastle. It's more important to know what primary factors are dominate based on your handicapping. Tim __________________ Trust but verify

06-06-2022, 04:25 PM	#8
Lt1 Grade 1 Join Date: Apr 2012 Location: Valley Stream NY Posts: 11,385	Hi Dan. I myself will do a max of 100 races and I know this is enough based on my work. Howard made the statement. I agree 20 races maybe too small but I've seen 50 races samples work well also. The trouble with large samples is they tend to give the user brain freeze. They get too confused when things aren't crystal clear. That's one of the reason[after hearing Tom Brohamer I now lean on the primary factor[s] where the winners in the top 2 the most times If I remember correctly the 10000 races were back in the day when they were using that[at the time] big college computer and we didn't have many computers on the market. I agree a new person should do as many 20 race cycles as needed to give them confidence and enough info for real betting. Trouble is most people folks won't do the work required. Tim __________________ Trust but verify

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
Gulfstream Race 1 - 2/7/18	Jeebs	Races of Interest	13	02-08-2018 10:30 PM
PRX 5th - 3/28/17	Jeebs	Selections	23	03-28-2017 11:53 PM
Pace line question	Bill V.	Sartin Methodology Handicapping 101 (102 ...)	15	06-12-2015 09:53 AM
Using RDSS with a sample race in my learning (Apr 13, 2013 - OP Race 2)	NeedForSpeed	RDSS	5	06-13-2013 02:28 PM

06-04-2022, 08:36 PM	#4
DanBoals Grade 1 Join Date: Apr 2009 Location: Reno, NV Posts: 445	33 races is too small of a sample to not have oddities mathematically. The fact that you have 27 out of 33 winners in your top 3 is what you should be concentrating on, that is very good. Howard told me that his original database that he based the methodology on was based on 10,000 races entered by hand. The primary line score that is used to create the BL/BL ranking is based mostly off what was learned in analyzing these 10,000 races on a DEC computer that Doc rented time on. I would be willing to bet any amount of money, that if you keep handicapping exactly like you are now for another 167 races, you will find that you have slightly more winners in the second rank than you do in the third rank, and you will have slightly more winners in the first rank than you have in the second rank. 200 races is a large enough sample for something like primary line score or BL/BL to distribute normally under most circumstances. The software works. On a small sample size, you might see some funky stuff, but over any decent number of races, like 200, or a season at the track, the software does what it is expected to do as long as you do your part.

06-06-2022, 11:40 AM	#5
Ryan Maiden Join Date: Feb 2021 Posts: 2	Right on! I appreciate it. Just making sure I am on the right track. I would agree too that the sample is much too small and it probably is just an odd thing going on. Wanted to make sure that I wasn't overthinking it. Thanks for all the help guys!!

06-07-2022, 11:24 AM	#9
DanBoals Grade 1 Join Date: Apr 2009 Location: Reno, NV Posts: 445	Amen, brother. "Trouble is most people folks won't do the work required." That is one of the truest things anyone has ever said about handicapping. There is a lot of work required to rise above the crowd in this sport. When you consider that the takeout is 15 to 30% depending on the track and kind of bet, you have to really know what you are looking at and really choose your spots. I have had some days lately where I handicap 6 races and then when they go off, none of them had the odds I need to make a bet. It is a hard, but that is also why it is so rewarding when you beat the horses and show a profit at the end of the month. Gotta put in the work, gotta pay your dues.