Playalong/Workshop on "Inside Information"

[Robert Fischer]

Originally Posted by Robert Fischer

track morning line =

my morning line odds =

my selections = (^ IF 2 is the favorite, and 7 gets some good support(as 2nd choice), I'll take 7 to win here.)

So my won easily, paying $5.20.
There wasn't really a lot of interesting tote action, in a race that was a lot less chaotic than the previous example. The main point here was that I thought the was better than the for a 'reality in which the would be favored and in which the was well supported as the 2nd choice. When this indeed played out to be true, I had a greater confidence that the reality in which I had handicapped for was in line with the betting market.

[lansdale]

Originally Posted by Robert Fischer

What is 'Inside Information' at the track?

& Is it possible for one of us, the horseplayers, to identify when inside-info may exist, and adjust accordingly??


- In this thread, I will attempt to address those questions from my own POV, as well as provide a sample race for everyone to play-along and test the concept.

I invite you to play along, and I welcome any input on the subject of 'Inside Information'.

Hi RF,

Very interesting idea for a thread. And it think it wouldn't be a bad idea to see if there is still anything of value to be gleaned from tracking odds directionality. Would also be of interest to me to see if a 'lemming' factor could be isolated.

But first, one question: wondering how you define 'hotness' or max. downward odds movement, in this race. If you are using the ML, it seems to me that the #10 dropped the most, then, the #1 and the #4. The 11, 12, and 6 all rose. In this context, how can the 1 be the hottest horse?

Re the 'lemming' factor, per Flatstats recent posting of the odds 'shrinkage' of UK winning favorites (if I'm reading this correctly), what about the possibility that a horse dropping too far from its ML is a sign of 'lemming' money, and thus negative. In this race e.g., the 10 was .58 of it's ML odds, the 1 - .7, and the 4 - .82. Of course, as Thask has pointed out, the ML is arbitrary, but for this reason, possibly exploitable at many tracks.

Cheers,

lansdale

[DeltaLover]

Originally Posted by Robert Fischer

So my won easily, paying $5.20.
There wasn't really a lot of interesting tote action, in a race that was a lot less chaotic than the previous example. The main point here was that I thought the was better than the for a 'reality in which the would be favored and in which the was well supported as the 2nd choice. When this indeed played out to be true, I had a greater confidence that the reality in which I had handicapped for was in line with the betting market.

I think that the was the worst betting option in the race, given its profile and final odds. I would have never bet it on top....

[Nitro]

Originally Posted by thaskalos

What you suggest here is "naïve" as well. There are no significant "betting patterns" in today's game. There once was a time when the timing of these bets was deemed significant, because it revealed the actual time of the bet being made at the betting windows. Today...more than half of the money is wagered into the pools as the horses are walking into the gate...and the rest of the money is funneled into the pools whenever the out-of-town tracks feel like sending the bets to the host track.

Tote board analysis, as we once knew it, is dead.

Your statement obviously shows that you're not up to date or misinformed about how today's betting pattern analyses actually work. But that's okay, I'm certainly not going to try and convince you or anyone else around here about how lucrative a real tote analysis can be.

I agree that is was probably easier 10 years ago. But when you think about it, does anyone actually believe that late money is insider action? I doubt that any intelligent bettor with solid info is going to wait for the last minute to place a serious wager. Besides that, anyone who understands these betting patterns knows the limited significance of monitoring individual betting pools.

Originally Posted by Robert Fischer

The methodology I have covered in this thread is very fundamentally sound.

You may know of, or wish to offer some more advanced tote analysis techniques. You are welcome to introduce them in here, or in your own thread.

Well, I must admit it is fundamental, but sound is another story. As one poster put it this thread is about having some fun. So who am I to spoil it. My personal enjoyment is occasionally making a nice score with what I believe to be worthwhile information. Unfortunately I don't have access as to how the betting patterns I use are derived, so any explanation on my part would probably be an injustice to the developer. As far as my own threads go I’ll stick with starting selection threads. They’re a bit more revealing, and demonstrate just how alive and worthwhile a real tote analysis can be.
.

[Robert Fischer]

Originally Posted by lansdale

Would also be of interest to me to see if a 'lemming' factor could be isolated.

...

Re the 'lemming' factor, per Flatstats recent posting of the odds 'shrinkage' of UK winning favorites (if I'm reading this correctly), what about the possibility that a horse dropping too far from its ML is a sign of 'lemming' money, and thus negative. In this race e.g., the 10 was .58 of it's ML odds, the 1 - .7, and the 4 - .82. Of course, as Thask has pointed out, the ML is arbitrary, but for this reason, possibly exploitable at many tracks.

Lemmings, sheep, the herd-instinct... oh my
It's a jungle out there.

I'm not sure how to necessarily calculate it, but I can look at a race like that one, and at least raise the question.

Did , , and really comprise 68% of the win-probability for that specific 12 horse turf race??

I don't think that it would be unreasonable to question that.

Were the or the Hot horses, or Lemming horses?

I think that is harder to say with any certainty before the race is run.

Was the daily double pool from the previous race favorite (favoring the ) 'smarter' money than the win pool (leaning toward ), or just more accurate in that specific sample? Or was leveraging involved?

It's all certainly interesting.

[thaskalos]

Originally Posted by Nitro

Your statement obviously shows that you're not up to date or misinformed about how today's betting pattern analyses actually work. But that's okay, I'm certainly not going to try and convince you or anyone else around here about how lucrative a real tote analysis can be.

I agree that is was probably easier 10 years ago. But when you think about it, does anyone actually believe that late money is insider action? I doubt that any intelligent bettor with solid info is going to wait for the last minute to place a serious wager. Besides that, anyone who understands these betting patterns knows the limited significance of monitoring individual betting pools.

Well...I happen to know for a fact that there are some very intelligent bettors, with solid info, who wait for the last SECONDS to place their serious wagers...but as you say...what reason could I possibly have to try to change your mind?

[Stillriledup]

Originally Posted by thaskalos

Well...I happen to know for a fact that there are some very intelligent bettors, with solid info, who wait for the last SECONDS to place their serious wagers...but as you say...what reason could I possibly have to try to change your mind?

This is correct.

The longer you wait, the more information you can gather on warmups, money adding to the pools, late scratches, horses flipping in the gate, late rider changes, etc.

You need every edge you can get, there's no edge to betting early other than maybe once in a while not getting shut out of some or all of your plays.

[ReplayRandall]

Originally Posted by Robert Fischer

Lemmings, sheep, the herd-instinct... oh my
It's a jungle out there.

I'm not sure how to necessarily calculate it, but I can look at a race like that one, and at least raise the question.

Did , , and really comprise 68% of the win-probability for that specific 12 horse turf race??

I don't think that it would be unreasonable to question that.

Were the or the Hot horses, or Lemming horses?

I think that is harder to say with any certainty before the race is run.

Was the daily double pool from the previous race favorite (favoring the ) 'smarter' money than the win pool (leaning toward ), or just more accurate in that specific sample? Or was leveraging involved?

It's all certainly interesting.

RF, thank you for this entertaining and interesting thread. I'm following the journey of thoughts, comments and analyzation which seem to be leading to a possible "breakthrough" conclusion. You've got my attention, so by all means, please continue.........

[lansdale]

Originally Posted by Robert Fischer

Lemmings, sheep, the herd-instinct... oh my
It's a jungle out there.

I'm not sure how to necessarily calculate it, but I can look at a race like that one, and at least raise the question.

Did , , and really comprise 68% of the win-probability for that specific 12 horse turf race??

I don't think that it would be unreasonable to question that.

Were the or the Hot horses, or Lemming horses?

I think that is harder to say with any certainty before the race is run.

Was the daily double pool from the previous race favorite (favoring the ) 'smarter' money than the win pool (leaning toward ), or just more accurate in that specific sample? Or was leveraging involved?

It's all certainly interesting.

Hi Robert,

I didn't mean 'lemming' in the subjective sense - I was asking what the statistical value of directionality is, and whether it can be exploited. Sam Wang at Princeton Electronic Consortium was asking this same kind of question about late movement in voter polls before the recent election. Intuitively, it seems that we sense that winners tend to be horses whose odds are dropping - but, if so, by how much? If you take a look at Flatstats' charts of odds shrinkage in winning ML favorites, the mean is ca. 90%. One wonders what the comparable odds drop (if any) there would be for higher ranked horses. The question I'm asking is, could a horse be dropping 'too much' in a way that reveals that the 'uninformed' money was driving it rather than the smart money. There may be a statistical sweet spot for dropping odds that such a horse might fall outside of.
And I think, to really examine this, there needs to be some objective measure of what a 'hot' horse is, which is why I mentioned those percentages.

As far as this race went, without running it through any software and trying not to redboard, I think I would have made it 10,1,4,12. The three that you mention were also the top Bris PP - to use a recent TM stat re speed fig weighting, the 10,1,12 had the best figs in their most recent race. I agree with you re the 4 - it's most recent races didn't stack up. If, out of this group you, therefore, eliminated the 4 and just dutched the other droppers, the 10 and the 1, using directionality would have worked. But how would it work in the more average 8 and under fields? My guess - not much value. But I would be interested in seeing some serious research on this. I wouldn't doubt that someone somewhere may already be using it.

Cheers,

lansdale

[lansdale]

Originally Posted by lansdale

Hi Robert,

I didn't mean 'lemming' in the subjective sense - I was asking what the statistical value of directionality is, and whether it can be exploited. Sam Wang at Princeton Electronic Consortium was asking this same kind of question about late movement in voter polls before the recent election. Intuitively, it seems that we sense that winners tend to be horses whose odds are dropping - but, if so, by how much? If you take a look at Flatstats' charts of odds shrinkage in winning ML favorites, the mean is ca. 90%. One wonders what the comparable odds drop (if any) there would be for higher ranked horses. The question I'm asking is, could a horse be dropping 'too much' in a way that reveals that the 'uninformed' money was driving it rather than the smart money. There may be a statistical sweet spot for dropping odds that such a horse might fall outside of.
And I think, to really examine this, there needs to be some objective measure of what a 'hot' horse is, which is why I mentioned those percentages.

As far as this race went, without running it through any software and trying not to redboard, I think I would have made it 10,1,4,12. The three that you mention were also the top Bris PP - to use a recent TM stat re speed fig weighting, the 10,1,12 had the best figs in their most recent race. I agree with you re the 4 - it's most recent races didn't stack up. If, out of this group you, therefore, eliminated the 4 and just dutched the other droppers, the 10 and the 1, using directionality would have worked. But how would it work in the more average 8 and under fields? My guess - not much value. But I would be interested in seeing some serious research on this. I wouldn't doubt that someone somewhere may already be using it.

Cheers,

lansdale

I should have added that the lemming money is the uninformed money - bet on a horse that is dropping (and possibly dropping the most) and running out or at least, or not winning - however you choose to define it.

[DeltaLover]

Originally Posted by lansdale

Hi Robert,

I didn't mean 'lemming' in the subjective sense - I was asking what the statistical value of directionality is, and whether it can be exploited. Sam Wang at Princeton Electronic Consortium was asking this same kind of question about late movement in voter polls before the recent election. Intuitively, it seems that we sense that winners tend to be horses whose odds are dropping - but, if so, by how much? If you take a look at Flatstats' charts of odds shrinkage in winning ML favorites, the mean is ca. 90%. One wonders what the comparable odds drop (if any) there would be for higher ranked horses. The question I'm asking is, could a horse be dropping 'too much' in a way that reveals that the 'uninformed' money was driving it rather than the smart money. There may be a statistical sweet spot for dropping odds that such a horse might fall outside of.
And I think, to really examine this, there needs to be some objective measure of what a 'hot' horse is, which is why I mentioned those percentages.

As far as this race went, without running it through any software and trying not to redboard, I think I would have made it 10,1,4,12. The three that you mention were also the top Bris PP - to use a recent TM stat re speed fig weighting, the 10,1,12 had the best figs in their most recent race. I agree with you re the 4 - it's most recent races didn't stack up. If, out of this group you, therefore, eliminated the 4 and just dutched the other droppers, the 10 and the 1, using directionality would have worked. But how would it work in the more average 8 and under fields? My guess - not much value. But I would be interested in seeing some serious research on this. I wouldn't doubt that someone somewhere may already be using it.

Cheers,

lansdale

I do not think that the topic of the thread is one that can be treated by a statistical approach.

Quite the contrary!

Board movements and hot - cold horse assessment, belongs in the artistic rather than the scientific part of the puzzle and should be approached as such!

Each race has two kind of characteristics: those who can be treated macroscopically (where applying statistics makes sense) and those that specify its individual attributes (where intuition and experience plays the first role to detect their validity) making it an event that never occurred in the past and will next be repeated in the future .. What we are examining here, clearly belongs to that later case and we should continue to examine in a per - race base, possibly adding some more comments about how each horse was bet...

[DeltaLover]

Yesterday (NOV 13) I liked a the Ferrero in the 6th in DEL MAR.

His angles that I considered were as follows:

Second Time Out
Won as a FTS
Claimed Last Time
Moves to outer post for first time
Tries Artificial Surface for first time
Took the lead in shorter distance

Considering the other horses in the race I think I might have an overlay, something that it is confirmed by running the race though a computerized model, that happens to agree with me, stating that the EV for each $100 I will bet on it will be around $35 (!!!).

I am now sold!
I place my bet when the horses are loading and the odds on the are9-1

A few blinks before the bell the horse drops from 9-1 to 5-1 at once!

What does this mean?

Do I have a terrible UNDERLAY or in reality the is a hot horse who is about to run a great race? Note also, that here, we are not talking about Turf Paradise or Finger Lakes but for a track with very large pools which in theory are very difficult to move so sharp in a single blink!!!

Anyway, the is proven to be a no factor in the race, where the top two crowd's choices are finishing 1 - 2 and he just manages to complete the bottom spot of the super. I have just managed to burn a few hundred and keep on wondering about the whos and the whys, this dog, was bet so heavily..

[TrifectaMike]

Originally Posted by lansdale

Hi Robert,

I didn't mean 'lemming' in the subjective sense - I was asking what the statistical value of directionality is, and whether it can be exploited. Sam Wang at Princeton Electronic Consortium was asking this same kind of question about late movement in voter polls before the recent election. Intuitively, it seems that we sense that winners tend to be horses whose odds are dropping - but, if so, by how much? If you take a look at Flatstats' charts of odds shrinkage in winning ML favorites, the mean is ca. 90%. One wonders what the comparable odds drop (if any) there would be for higher ranked horses. The question I'm asking is, could a horse be dropping 'too much' in a way that reveals that the 'uninformed' money was driving it rather than the smart money. There may be a statistical sweet spot for dropping odds that such a horse might fall outside of.
And I think, to really examine this, there needs to be some objective measure of what a 'hot' horse is, which is why I mentioned those percentages.

As far as this race went, without running it through any software and trying not to redboard, I think I would have made it 10,1,4,12. The three that you mention were also the top Bris PP - to use a recent TM stat re speed fig weighting, the 10,1,12 had the best figs in their most recent race. I agree with you re the 4 - it's most recent races didn't stack up. If, out of this group you, therefore, eliminated the 4 and just dutched the other droppers, the 10 and the 1, using directionality would have worked. But how would it work in the more average 8 and under fields? My guess - not much value. But I would be interested in seeing some serious research on this. I wouldn't doubt that someone somewhere may already be using it.

Cheers,

lansdale

Hi lansdale,

Following the flow of money bet to determine "informed" betting is chaotic at best. Identifying "bad" money is doable. And leads to more stable solutions and useful information for betting purposes.

Mike

[Robert Goren]

I am in the process of developing a method for detecting "smart money" and predicting a win percentage for it using BRIS Prime Power to set the expected odds. It is still a ways from becoming even a Beta. The more I work on it, the more hopeful I become that it will be something useful. I think I have several unique ideas to incorporate in it.

[Robert Fischer]

Originally Posted by lansdale

Hi Robert,

I didn't mean 'lemming' in the subjective sense - I was asking what the statistical value of directionality is, and whether it can be exploited. Sam Wang at Princeton Electronic Consortium was asking this same kind of question about late movement in voter polls before the recent election. Intuitively, it seems that we sense that winners tend to be horses whose odds are dropping - but, if so, by how much? If you take a look at Flatstats' charts of odds shrinkage in winning ML favorites, the mean is ca. 90%. One wonders what the comparable odds drop (if any) there would be for higher ranked horses. The question I'm asking is, could a horse be dropping 'too much' in a way that reveals that the 'uninformed' money was driving it rather than the smart money. There may be a statistical sweet spot for dropping odds that such a horse might fall outside of.
And I think, to really examine this, there needs to be some objective measure of what a 'hot' horse is, which is why I mentioned those percentages.

Words like subjective and/or objective come out, and people usually think of polarized stances.

When You are tackling this, or when I am tackling this, or when Sam Wang is tackling this (don't make me bring Sam Wang into this thread, /just kidding ), we're doing essentially the same thing.

We are starting with an an insight that clearly sees and and understands the system.

Then, we are looking/waiting for systemic events or scenarios to occur that we understand to be profitable.

Ultimately, all things being equal, it's better to have a machine calculate it, so we can go to the beach, but we are doing the same fundamental things.

-
Do you have a link for the Sam Wang 'directionality' article? I see several interesting Princeton Consortium articles by Wang. , Thanks.