RPM The Good And Not So good

[Saratoga]

Originally Posted by Topcat

headhawg,

First, I encourage everyone to do their own testing and to come their own fact based conclusions.

Secondly, as Speed Figure explains above, both BRIS and Handicappers Data Warehouse get their data from the same place, that is Equibase, it is not like one gets their data from Equibase and another from Timeform. The value comes from what the HDW programs can do with the data.

Third, I did not test HDW data in RPM programs, but tested the data or similar factors. One that did show a gap in favor of HDW was last race speed rating in BRIS vs. Cramer Speed rating, and you can argue that this isn't "apples to apples" as Cramer uses a different projection method in calculating track variant. While they correlated over 90% of the time, Cramer was a little better.

In over 10,000 races the BRIS best last race speed rating , if over 1 point of next horse, the loss was in my test 14% or $1.74 for every two dollars, HDW Cramer speed rating same criterion loss of 10% or $1.80. Both losers and this is one of the factors that would give an edge to HDW but this is one of the few that showed a difference, now as you get into PSR, projected speed ratings and different weighted pace calculations there is added value but those, IMHO is driven by the software partners of HDW and not necessarily the raw data.

I'm still waiting for verification on this statement....

In my years of handicapping , I have NEVER saw a speed based rating

at .90


As for you Dave.....your number of starts and wins are way off for 2023....Can you verify that

please....I am meticulous in my data records , so I am confident in my numbers

[Dave Schwartz]

Originally Posted by Saratoga

I'm still waiting for verification on this statement....

In my years of handicapping , I have NEVER saw a speed based rating

at .90


As for you Dave.....your number of starts and wins are way off for 2023....Can you verify that

please....I am meticulous in my data records , so I am confident in my numbers

Nothing I post is ever WAY OFF.
LOL

This was a subset.

Total sample was 27,446.

Look

Race filters
1-1-23 to 12-31-23
Dirt, fast tracks, 5f to 9f
Removed races with entries
*Removed young horses 2yr olds & 3yr olds before May)
*Forgot to indicate that in the original run.

Horse Filters
No FTS

So, shorter than 5f and longer than 9f were removed.
Off tracks were removed.
Races with entries were removed.

The purpose was to remove things that obfuscate the importance of the core data.

Show how many races you found and I'll run the entire sample.

[Dave Schwartz]

I DID have a wrong filter in one of the runs:
Don't recall which one.

The full year.

But still no TINY TRACKS like Salem or Brown Cnty Fair.

2023.
168-rcRtgWIN BETS
Field1 Field2  Starts   Pays    Pct   $Net     IV    PIV     DV     BV  AvPay AvOdds AvPubC
1st            37,409 10,975   29.3  $1.66   2.03   1.04   1.07   2.21  $5.64   1.89   2.07
2nd            34,231  6,627   19.4  $1.64   1.34   1.02   1.06   1.49  $8.47   3.29   2.98
3rd            32,414  4,551   14.0  $1.58   0.98   0.98   1.02   1.12 $11.23   4.68   3.69
FH             44,261  4,163    9.4  $1.51   0.77   0.97   0.97   0.76 $16.03   7.35   4.94
RH             75,571  4,587    6.1  $1.34   0.45   0.91   0.86   0.52 $22.09  11.17   5.93

Total         223,886 30,903   13.8  $1.51   1.01   0.98

$NetR=  1.24   IVR=  4.51   PIVR=  1.14 PubChR=  2.86

2023.
174-rPSRWIN BETS
Field1 Field2  Starts   Pays    Pct   $Net     IV    PIV     DV     BV  AvPay AvOdds AvPubC
1st            34,864 11,013   31.6  $1.69   2.19   1.05   1.09   2.35  $5.34   1.71   1.87
2nd            33,233  6,684   20.1  $1.65   1.39   1.02   1.06   1.54  $8.21   3.13   2.78
3rd            32,373  4,575   14.1  $1.55   0.98   0.97   1.00   1.15 $10.94   4.57   3.57
FH             44,986  4,235    9.4  $1.53   0.77   0.97   0.99   0.76 $16.27   7.37   4.88
RH             78,431  4,396    5.6  $1.33   0.42   0.89   0.86   0.49 $23.81  11.93   6.03

Total         223,887 30,903   13.8  $1.51   1.01   0.98

$NetR=  1.27   IVR=  5.21   PIVR=  1.18 PubChR=  3.22

2023.
175-rCPwWIN BETS
Field1 Field2  Starts   Pays    Pct   $Net     IV    PIV     DV     BV  AvPay AvOdds AvPubC
1st            34,402 10,363   30.1  $1.68   2.08   1.05   1.08   2.25  $5.58   1.83   2.03
2nd            33,189  6,539   19.7  $1.61   1.37   1.01   1.04   1.53  $8.19   3.17   2.89
3rd            32,218  4,713   14.6  $1.58   1.02   0.99   1.02   1.16 $10.83   4.51   3.62
FH             44,833  4,425    9.9  $1.56   0.81   0.99   1.01   0.78 $15.82   7.15   4.87
RH             79,245  4,862    6.1  $1.32   0.46   0.89   0.85   0.54 $21.55  10.85   5.87

Total         223,887 30,902   13.8  $1.51   1.01   0.98

$NetR=  1.27   IVR=  4.52   PIVR=  1.18 PubChR=  2.89



     FxNum     Factor   $NetR    IVR   PIVR PubChR  Score
      1    174 rPSR      1.27   5.21   1.18   3.22  52.94
      2    175 rCPw      1.27   4.52   1.18   2.89  48.16
      3    168 rcRtg     1.24   4.51   1.14   2.86  45.52

[Saratoga]

I don't want you to think that I'm attacking you on this but I am puzzled on

why there's such a discrepancy.....

heres my stats and what I use to get it...

I didn't take out entries or the 2 & 3 yo before May or small tracks , cause I don't know how

to do those ...but I'm looking to get up to your numbers , not increasing the

difference...

Anyway thx for the info

[Dave Schwartz]

Difference in what?

[Topcat]

Originally Posted by Saratoga

Let me say that .90 roi with Cramer speed is quite good in this day and age....



Its ROI for Rank 1 is .85 for 2023


thx

It's my wife's birthday and in the morning I have to start travelling for work and will be out of the loop for a few weeks. So I'm tempted to take the easy way out and just say, well maybe I'm wrong.

The point of my post was that the data ,which both comes from Equibase is not that much different. It’s the programs output and weighting and yes I do believe at least 3 of those programs that I am familiar with on HDW are far superior because of what they do.

BTW: I would say I'm impressed in today's current very small fields that you can get .85 on a single factor. Field size has dropped a lot even in the last ten years.

I was looking at summary results in Excel. My data is admittedly dated, lots of it before 2018 , definitely not "this day and age" , and by way of asking it for a 1 point gap advantage, it is already filtering out ties and I remember very few maiden races and I think I may have eliminated any who failed to meet pars.

I'm pretty good with Excel, Access and some other relational databases but I can only see smart on a clear day, and it is often far away. Lol

So my secret weapon with data analytics is my adult son and his software wizards that can code, dissect reverse engineer files etc.

Currenly he, and his business partners run their own software business , Surf Node, that helps buyers/traders arbitrage Solana crypto, and their clients are making million dollar bets (my word “bets” not theirs) so while they helped me in the past with some programming and data analysis request right now they are a pretty busy to help even for dear old Dad . They started recommending buying Sol back in $8 to $12 range after that FTX fraud and Solana just hit $180.

So see if I was smart I would have been buying Sol at the bottom.