The Company Line

[Robert Fischer]

enjoy reading the comments and approaches here.


scan for some bolded 'rematch' rivals... I scan for known names, and occasionally for some mutual rivals within a class or circuit, which are mutual to different entries in today's race.

Margin of victory - look at bunched, 'blanket' finishes, but other than that, I get most from the running lines, and then video replay if there's enough to warrant.

[JohnGalt1]

Originally Posted by thaskalos

IMO...the company line is the most useless piece of "handicapping" information in the past performances. I wish it were eliminated, to make room for more extensive trip notes on the particular horse.

I agree.

I use the Bris pp generator and do not print the company lines leaving a blank space on my printed form to write the pace figures. I find the class and purse of the past races to be more valuable.

The key to my success, or lack of success, is knowing what is valuable and what is detrimental to my bottom line.

[thaskalos]

Originally Posted by classhandicapper

I haven't studied the betting patterns of the public or the results of races closely, but this is a very interesting topic.

If A and B are coming out of the same race you can compare their performances without worrying about whether their figures are accurate, how the pace and race development impacted the final time, whether they ran on tracks that were playing differently etc..

If you think A ran a little better than B in that race you are done.

If A and B are coming out of different races and have similar small differences in their figures, there's a margin of error associated with those figures that doesn't exist when they come out of the same race.

I have never agreed with this sort of handicapping reasoning. IMO...the horses in a particular race are all in different stages of their racing development, and it doesn't matter how they have finished in their individual bouts against one-another. That's why they regularly take turns beating each other. As an example...if a particular horse won its race in a prolonged and exhaustive drive down to the wire...then the odds are that this horse will be given a "breather" in its subsequent race...and may finish behind a horse that it beat soundly in the prior start. The company line makes no such distinctions...and the simplistic info that it provides can do more harm than good.

Yes...I suppose the company line could be of some use if "all other things are equal". But the other things are NEVER "equal" in this game.

[classhandicapper]

Originally Posted by thaskalos

Yes...I suppose the company line could be of some use if "all other things are equal". But the other things are NEVER "equal" in this game.

All things are never equal whether you use the company line, your favorite pace and speed figures, or both. There's also always more to the analysis than any single snapshot in time.

However, you have to start somewhere.

I don't care if you get to that starting point with numbers or company lines.

Using figure language, I am simply saying that all else being equal if A ran a 95 and B ran a 90 in the same race, over the long haul that's more accurate than if A ran a 95 and B ran a 90 in two different races.

The complexities of evaluating horses that raced in different races with different track speeds, different winds, different paces, and different quality horses etc... makes evaluating horses from different races more difficult and less accurate than if they faced each other.

So among very similar looking horses, how they did against each other given their respective trips enhances your understanding beyond just using figures as the starting point.

[thaskalos]

Originally Posted by classhandicapper

All things are never equal whether you use the company line, your favorite pace and speed figures, or both. There's also always more to the analysis than any single snapshot in time.

However, you have to start somewhere.

I don't care if you get to that starting point with numbers or company lines.

Using figure language, I am simply saying that all else being equal if A ran a 95 and B ran a 90 in the same race, over the long haul that's more accurate than if A ran a 95 and B ran a 90 in two different races.

The complexities of evaluating horses that raced in different races with different track speeds, different winds, different paces, and different quality horses etc... makes evaluating horses from different races more difficult and less accurate than if they faced each other.

So among very similar looking horses, how they did against each other given their respective trips enhances your understanding beyond just using figures as the starting point.

I understand what you are trying to say...but I am not of the same opinion. This doesn't mean that I consider you "wrong" and me 'right', it just means that we have different ways of seeing things...and that's how things should be in this game...IMO. If we all saw the game the same way...then this game would be doomed right from the start.

I, personally, have never looked at the company line in over a quarter-century...that's why I called it "useless" to me. But if this bit of information is vital to someone else, then I raise my hat to him...and I pray that this info remains in the past performances forever. I am all for "diversity"...in sport, and in life as a whole.

[classhandicapper]

Originally Posted by thaskalos

I understand what you are trying to say...but I am not of the same opinion. This doesn't mean that I consider you "wrong" and me 'right', it just means that we have different ways of seeing things...and that's how things should be in this game...IMO. If we all saw the game the same way...then this game would be doomed right from the start.

I, personally, have never looked at the company line in over a quarter-century...that's why I called it "useless" to me. But if this bit of information is vital to someone else, then I raise my hat to him...and I pray that this info remains in the past performances forever. I am all for "diversity"...in sport, and in life as a whole.

I hear you.

I thought I wasn't explaining myself clearly.

[thaskalos]

Originally Posted by classhandicapper

I hear you.

I thought I wasn't explaining myself clearly.

Quite the contrary. If you have a fault...it's that you explain yourself TOO clearly.

[MJC922]

One of the things my class ratings have proven to me is that when handled properly 'company' i.e. who beat who and by how much is approximately 80% of a horses probability to win a race. You might say the output from a strict 'company' analysis predicts the winner of approx 30% of all races while the crowd favorite wins what is it these days 38%?

That 38% for favorites varies quite a bit by track though, I'm just talking rough numbers here. Anyway it's hard to believe but I can assure you that number approx 80% of a horse's probability is true and it always has been. Someone may even be able to get the number to 31 or 32% with some company analysis that is superior to mine but I believe mine which is 100% computer generated is quite solid and there is an upper limit, probably of 31% on this one factor if I had to guess.

Looking at racing from a high level then it's a very predictable game with a high takeout -- plus with the odds volatility these days limiting the ability to do an acceptable risk/reward calculation (except on the biggest of racing days like the BC or Derby etc) I do wish everyone who still plays the best of luck.

[classhandicapper]

Originally Posted by MJC922

One of the things my class ratings have proven to me is that when handled properly 'company' i.e. who beat who and by how much is approximately 80% of a horses probability to win a race. You might say the output from a strict 'company' analysis predicts the winner of approx 30% of all races while the crowd favorite wins what is it these days 38%?

That 38% for favorites varies quite a bit by track though, I'm just talking rough numbers here. Anyway it's hard to believe but I can assure you that number approx 80% of a horse's probability is true and it always has been. Someone may even be able to get the number to 31 or 32% with some company analysis that is superior to mine but I believe mine which is 100% computer generated is quite solid and there is an upper limit, probably of 31% on this one factor if I had to guess.

Looking at racing from a high level then it's a very predictable game with a high takeout -- plus with the odds volatility these days limiting the ability to do an acceptable risk/reward calculation (except on the biggest of racing days like the BC or Derby etc) I do wish everyone who still plays the best of luck.

Are your class ratings based on multiple races?

Are they weighted as in more recent races count for more?

I tried to automate a multi-race process for testing purposes, but I abandoned it after I finished it. It performed slightly better than using just "last race", but I found that there were so many subjective aspects to interpreting which races to use and how to weight them I couldn't code it all. Watching the races, looking at the trips, and doing a subjective analysis is better for me.

Not that this kind of comparative analysis is for everyone, but it sometimes surprises me there's so much resistance to it. Timeform Europe and the Racing Post have had enormous success creating ratings based on who beat who by how much at what weights. So it's not like this is some novel idea or that it has been totally discredited. There are just some situations where it works better than others.

[MJC922]

Originally Posted by classhandicapper

Are your class ratings based on multiple races?

Are they weighted as in more recent races count for more?

I tried to automate a multi-race process for testing purposes, but I abandoned it after I finished it. It performed slightly better than using just "last race", but I found that there were so many subjective aspects to interpreting which races to use and how to weight them I couldn't code it all. Watching the races, looking at the trips, and doing a subjective analysis is better for me.

Not that this kind of comparative analysis is for everyone, but it sometimes surprises me there's so much resistance to it. Timeform Europe and the Racing Post have had enormous success creating ratings based on who beat who by how much at what weights. So it's not like this is some novel idea or that it has been totally discredited. There are just some situations where it works better than others.

Every race every horse has run in north america is utilized. I think after six months a race is essentially locked down and becomes static, anything within the past six months is recalibrated each day however a race likely 'moves' very little in most cases the older it becomes. Once a race is 'rated' it's of course easy to rate each horse's effort within that race. From there you have all of the individual's races rated and it's time to rate the individual horse. The intent is to rate the current class of the horse today so certainly the most recent races are what matter and consistency does play a significant role. I brute forced the optimization to be as predictive as it could be. It's the output coming from a process which is 100% mechanical. I'm not sure how Timeform UK does their stuff, I thought it was subjective to some extent.

[classhandicapper]

Originally Posted by MJC922

Every race every horse has run in north america is utilized. I think after six months a race is essentially locked down and becomes static, anything within the past six months is recalibrated each day however a race likely 'moves' very little in most cases the older it becomes. Once a race is 'rated' it's of course easy to rate each horse's effort within that race. From there you have all of the individual's races rated and it's time to rate the individual horse. The intent is to rate the current class of the horse today so certainly the most recent races are what matter and consistency does play a significant role. I brute forced the optimization to be as predictive as it could be. It's the output coming from a process which is 100% mechanical. I'm not sure how Timeform UK does their stuff, I thought it was subjective to some extent.

It must have taken quite awhile to develop all that. Impressive.

Timeform is more or less making "projection class ratings" in the same way speed handicappers make "projection speed figures".

Each horse comes into a race with a series of class oriented ratings from prior races. They look at the result of today's race (who beat who by how much at what weights), and then project the ratings for each horse for today's race based on a subjective analysis of the result.

That's more or less what I do, but I don't put emphasis on weight carried. I put more emphasis on the way the race developed. So for me, horses that didn't win are often rated higher than those that finished ahead of them.

To be honest though, other than for my automated ratings I use for testing ideas, I don't even actually bother creating those subjective ratings and recording them somewhere. I know the horses. So most of my notes are trip, bias, and race flow related. When horses come back, I just refresh my memory by looking at the chart and my notes.

[mountainman]

If you're intimately familiar with the horses involved-to the point of remembering flow, who fired how big, and who went on to do what, company lines can be useful. But while poor company lines are often the kiss of death, good company lines must be kept in perspective and can only illuminate in the context of these and other more intricate factors.

At mnr, company lines tell me most of the story. At Mahoning, they speak to me. But at any other track, I might as well be reading gibberish.

[Tom]

Company lines can be very easy, too - like at Belmont, they are often the entire field that ran in that race!

[classhandicapper]

Originally Posted by mountainman

If you're intimately familiar with the horses involved-to the point of remembering flow, who fired how big, and who went on to do what, company lines can be useful. But while poor company lines are often the kiss of death, good company lines must be kept in perspective and can only illuminate in the context of these and other more intricate factors.

At mnr, company lines tell me most of the story. At Mahoning, they speak to me. But at any other track, I might as well be reading gibberish.

I agree.

Without knowing the trips, they can easily be misleading.

[HalvOnHorseracing]

Perhaps it takes a little more time, but wouldn't it be best to just get access to replays and watch them instead of relying on the comments?