Benchmark times - Ellerslie

[mardigras]

This is pretty funny. I'm not even sure what they think a punter can do with this information. Perhaps someone can enlighten me. I'm assuming you can get the detail for all the classes/distances, not just the ones below. These are the times supposedly expected to be achieved by the winner under the following scenarios.

I do find it a pretty bizarre coincidence that every single number for 3yos over 1400m has the exact same benchmark time as R65 over 1400m for each track condition. What are the chances of that?

I like the benchmark time for a Good 3 over 1100m for Open grade races. In my data going back also 20 years at Ellerslie, the fastest time I have at Ellerslie for any horse over 1100m is 103.36, yet their benchmark time for an open horse is 102.9. So it would have to be based on some form of made up interpolation/extrapolation since the data simply doesn't exist to support that. I guess that means that every horse in the last 20 years to run on a G3 over 1100m at Ellerslie hasn't run to the benchmark for that grade.

It is clear that the majority of these numbers are highly massaged due to the limited number of horses to have raced in the various categories (even over 20 years). And that should be stated. For where they have limited data, it looks like they have applied some form of generic adjustment from one track rating to another to smooth the curve and make it look plausible. 

There is possibly nothing wrong with applying some theories to the data, but those theories should be supplied so that a punter understands what the number means and how it was derived. Many of the combinations of distance/grade/track rating will have zero or very few races to have run in the last 20 years. I use various methods for assessing performance on tracks that will use a degree of interpolation. However I do overall ability based comparative benchmarking, not class based benchmarking. So I don't divide the information up into smaller groups relating to class which simply results in smaller amounts of data at each level. I try to assess the performance of a horse relative to all other performances on the track distance/condition.

I also wonder why the benchmark for a 2yo at 1400m compared to a 3yo at 1400m is between 0.3 and 0.4 seconds (2 - 2.5 lengths). The difference in the WFA scale between a 2yo and a 3yo at 1400 ranges between 8kg and 10.5kg for a 1400m race. I'd read that 8 - 10.5kg would be expected to deliver between 6 - 8 lengths impact. Even 0.3 seconds is high for the data I have.

[LookingForValue]

Meaningless when no analysis criteria shown. As you said, small sample data makes a mockery of the process. I did notice that they have applied a set sequemce of times across each track condition level that is suspicious in itself. 

1200m R74 0.7 secs between track conditions. G3 = 1:09.7 D4 = 1:10.4 D5 = 1:11.1 D6 = 1:11.7 etc.
1100m Open 0.6 secs 
1400m 2yo 0.9 secs
1400m 3yo 0.9 secs
1500m Open 0.9 secs
1600m R74 1 sec. 
etc

[mardigras]

Just now, LookingForValue said:

Meaningless when no analysis criteria shown. As you said, small sample data makes a mockery of the process. I did notice that they have applied a set sequemce of times across each track condition level that is suspicious in itself. 

1200m R74 0.7 secs between track conditions. G3 = 1:09.7 D4 = 1:10.4 D5 = 1:11.1 D6 = 1:11.7 etc.
1100m Open 0.6 secs 
1400m 2yo 0.9 secs
1400m 3yo 0.9 secs
1500m Open 0.9 secs
1600m R74 1 sec. 
etc

It is of no use to anyone imo. Very contrived but for what purpose?

[LookingForValue]

Perhaps to make it look like they are helping punters with stats. Before I changed how I analyse data, I would have tried to use this. With help from those on this forum, I changed that. I suspect many will think they are of value. 

[curious]

3 hours ago, mardigras said:

I do find it a pretty bizarre coincidence that every single number for 3yos over 1400m has the exact same benchmark time as R65 over 1400m for each track condition. What are the chances of that?

Took one look at it and binned it. Didn't compute.  Just had another look when I saw this post The above is some kind of miracle though. I wonder what successful punter came up with this? There's no source provided. Hope a few start using the data though.

Edited June 4, 2021 by curious

[curious]

I can't figure out where they get those numbers?

[mardigras]

1 minute ago, curious said:

I can't figure out where they get those numbers?

Neither can I

[curious]

12 minutes ago, mardigras said:

Neither can I

Perhaps someone would come on here and explain how they were derived?