Anyone know how the 'Percentage Projections' work?

[Robert Fischer]

Laurel and possible others (is it a 1/st expressbet thing?) have a 'Win%' thing and it changes throughout the racing.

Sometimes isn't accurate.

I'm most interested in watching it change during a race while various things happen, and I'd love to know how it works... what drives it... what it's based on...

Nothing that significant, but it's just another thing that's throwing data of various accuracy at you.

[Onesome]

Does Betfair UK do inplay on US horse racing these days? My guess for the obvious choice where they are using the implied % from the betting there.

[Dave Schwartz]

Originally Posted by Robert Fischer

Laurel and possible others (is it a 1/st expressbet thing?) have a 'Win%' thing and it changes throughout the racing.

Sometimes isn't accurate.

I'm most interested in watching it change during a race while various things happen, and I'd love to know how it works... what drives it... what it's based on...

Nothing that significant, but it's just another thing that's throwing data of various accuracy at you.

How can you tell if a percentage is accurate or not? (I'd be more interested in that.)

If a horse shows 44% and wins was the percentage right?

If the horse shows 82% and loses was it wrong?

[coachv30]

Do you only see the top four? I would love to know what the % chance would be on the deep closer in the back as they hit the turn after blazing early fractions.

[classhandicapper]

There are probably a few metrics that make up the input but I’d say if you started with the horse’s running position/lengths ahead or behind and post time odds you’d be able to build a very good line just from that.

[xtb]

I'd like to know what purpose it serves if it is changing during the race, unless someone can cancel their bet.

[Robert Fischer]

Originally Posted by Dave Schwartz

How can you tell if a percentage is accurate or not? (I'd be more interested in that.)

If a horse shows 44% and wins was the percentage right?

If the horse shows 82% and loses was it wrong?

Good question dave.

I'll list an example related to a horse in the above picture.

Decemeber 17th LRL R6.

I think it incorporates win odds(and/or some power ranking), because it seems to start races with good horses/lower-priced w/ higher percentages.
It also seems to use whatever horse digital tracking technology there is to see changes in velocity.

Started some interest watching a horse called Continentalcongres who opened up big, and then 'hung' and got caught in deep stretch ...

First of all = it showed his % greatly increasing pre-turn and turn
Now that either is one model that tells me something about that horse's specific running style, or it is noise.

Secondly he opened up into the stretch and was in the 60%

Finally it abruptly dropped, although he had a big lead, apparently due to his deceleration - so that was a nice job by the Percentages...

however it quickly annointed a horse called I Can Run as a 50% chance winner!!
In reality InMyOpinion easily swept by leading the late flow to win. His chances were probably misread due to the fact that he was 25-1 and the IcanRun was 4-1. The algorithm couldn't handle the odds discrepancy and over-weighted that input.

[coachv30]

[Dave Schwartz]

Originally Posted by Robert Fischer

Good question dave.

I'll list an example related to a horse in the above picture.

Decemeber 17th LRL R6.

I think it incorporates win odds(and/or some power ranking), because it seems to start races with good horses/lower-priced w/ higher percentages.
It also seems to use whatever horse digital tracking technology there is to see changes in velocity.

Started some interest watching a horse called Continentalcongres who opened up big, and then 'hung' and got caught in deep stretch ...

First of all = it showed his % greatly increasing pre-turn and turn
Now that either is one model that tells me something about that horse's specific running style, or it is noise.

Secondly he opened up into the stretch and was in the 60%

Finally it abruptly dropped, although he had a big lead, apparently due to his deceleration - so that was a nice job by the Percentages...

however it quickly annointed a horse called I Can Run as a 50% chance winner!!
In reality InMyOpinion easily swept by leading the late flow to win. His chances were probably misread due to the fact that he was 25-1 and the IcanRun was 4-1. The algorithm couldn't handle the odds discrepancy and over-weighted that input.

So, if a horse wins and his pct was 60% was it right?

If it was 6% does that make it right?

"Did he win?" is a 0.00 or 1.00 answer.

If the horse wins then what would be the correct probability?

[mountainman]

Originally Posted by Robert Fischer

Good question dave.

I'll list an example related to a horse in the above picture.

Decemeber 17th LRL R6.

I think it incorporates win odds(and/or some power ranking), because it seems to start races with good horses/lower-priced w/ higher percentages.
It also seems to use whatever horse digital tracking technology there is to see changes in velocity.

Started some interest watching a horse called Continentalcongres who opened up big, and then 'hung' and got caught in deep stretch ...

First of all = it showed his % greatly increasing pre-turn and turn
Now that either is one model that tells me something about that horse's specific running style, or it is noise.

Secondly he opened up into the stretch and was in the 60%

Finally it abruptly dropped, although he had a big lead, apparently due to his deceleration - so that was a nice job by the Percentages...

however it quickly annointed a horse called I Can Run as a 50% chance winner!!
In reality InMyOpinion easily swept by leading the late flow to win. His chances were probably misread due to the fact that he was 25-1 and the IcanRun was 4-1. The algorithm couldn't handle the odds discrepancy and over-weighted that input.

Having watched (most times intently) a nearly countless number of races, I'm surprised at how infrequently the winner seems evident-or even predictable- before the field straightens away. Since I believe in downloading as much experience as possible, when betting horses I will at least quickly scan, if at hand, the pp's even of races at tracks that don't interest me and then make a point of watching the action play out.

What I see is a plethora of front-runners surviving duels, closers blowing past slow fractions, non-chalks overcoming rough trips, and seemingly beaten runners mounting one last surge to score. And that's quite aside from all the rocky movers that dig in, pin their ears (sometimes literally), and defeat more athletic rivals.

The implications of this greatly intrigue me , and while race-flow and visuals remain core tenants of my (admittedly old-school) game, I've come to understand that tons of results outright defy (after the fact) traditional concepts of pace, trips, position, and reserve energy.

I do have some theories as to why-not complex ones, since no portion of this post represents any sort of earth-shattering epiphany-but will save them for later as I savor the cowboys writhing in their death throes.

Hope all is well, Jason.

[Mikym]

Hey Dave, I can’t tell you if the percentages are accurate or not, since I don’t know how they are calculated. But if you think about it like poker percentages, it might make a bit more sense.

In Texas Holdem heads up, preflop, an unsuited pair of Aces has an 87.2% to win vs unsuited 7 2, which has a 12.40% to win. (There is a small percentage chance for a tie.). So obviously, you want the Aces, right? Post flop, the percentages could get better or worse. So I THINK if you use that logic in your example above, if the horse shows a 60% chance to win, it doesn’t mean the percentage was wrong if the horse loses.

[Robert Fischer]

Originally Posted by mountainman

Having watched (most times intently) a nearly countless number of races, I'm surprised at how infrequently the winner seems evident-or even predictable- before the field straightens away. Since I believe in downloading as much experience as possible, when betting horses I will at least quickly scan, if at hand, the pp's even of races at tracks that don't interest me and then make a point of watching the action play out.

What I see is a plethora of front-runners surviving duels, closers blowing past slow fractions, non-chalks overcoming rough trips, and seemingly beaten runners mounting one last surge to score. And that's quite aside from all the rocky movers that dig in, pin their ears (sometimes literally), and defeat more athletic rivals.

The implications of this greatly intrigue me , and while race-flow and visuals remain core tenants of my (admittedly old-school) game, I've come to understand that tons of results outright defy (after the fact) traditional concepts of pace, trips, position, and reserve energy.

I do have some theories as to why-not complex ones, since no portion of this post represents any sort of earth-shattering epiphany-but will save them for later as I savor the cowboys writhing in their death throes.

Hope all is well, Jason.

48-32!

Prescott looked awful.

Interested in some of your theories.
That would add a lot to this convo + the board in general lately and that's a shame that all of these knowledgeable players aren't discussing this type of stuff more often.

Races with depth of contenders, and sometimes simply full fields and/or Turf or Synth remind me that I'm often better at a Postmortem Analysis, than I am at Projection Forecast/Projection.

That's obvious, but I think Horseracing naturally leads us towards the Forecast/Projection and Predicting races.
Easily becomes the focus for most players(not most dollars, but most individuals).
A fair amount of players ONLY do Predictions other than the aftermath of a race they wagered on.

Inversion is part of the tool box. Moderately useful for a bunch of stuff. Significantly useful once in a while.

Currently I'm doing a bunch of postmortems and notes and watch-lists, and save for a few 'epiphanies' that 'jump off the screen' in the minutes leading to a race, I'm mostly taking opinions and betting when the value seems to be there. Then kind of letting things happen.

Less 'exciting', but lately been profitable and building a long term thing.

[Dave Schwartz]

Originally Posted by Mikym

Hey Dave, I can’t tell you if the percentages are accurate or not, since I don’t know how they are calculated. But if you think about it like poker percentages, it might make a bit more sense.

In Texas Holdem heads up, preflop, an unsuited pair of Aces has an 87.2% to win vs unsuited 7 2, which has a 12.40% to win. (There is a small percentage chance for a tie.). So obviously, you want the Aces, right? Post flop, the percentages could get better or worse. So I THINK if you use that logic in your example above, if the horse shows a 60% chance to win, it doesn’t mean the percentage was wrong if the horse loses.

Good answer but not really the point.

Decades ago I had a conversation with well-know horse guy who said that if you total up all the pcts of the horses you bet and compare to those that won, they should be close.

He further stated that this proved the pcts were right.

So, I gave him a problem.

"Imagine I bet 4 horses. They had projected win pcts of 40%, 30%, 20% and 10%, respectively. One of the horses won."

I asked, "Ignoring sample size, does this approach prove that the pcts were correct?"

He answered that it did.

I said, these horses were all in the same 4-horse field.

"How could I have been wrong?"

[horses4courses]

Originally Posted by Mikym

Hey Dave, I can’t tell you if the percentages are accurate or not, since I don’t know how they are calculated. But if you think about it like poker percentages, it might make a bit more sense.

In Texas Holdem heads up, preflop, an unsuited pair of Aces has an 87.2% to win vs unsuited 7 2, which has a 12.40% to win. (There is a small percentage chance for a tie.). So obviously, you want the Aces, right? Post flop, the percentages could get better or worse. So I THINK if you use that logic in your example above, if the horse shows a 60% chance to win, it doesn’t mean the percentage was wrong if the horse loses.

I think this is exactly the point.

These percentages have nothing to do with betting.
They are merely an indicator of percentage likelihood of success during a race.
Strength of poker hole cards is an excellent parallel.

[classhandicapper]

Dave,

We can ask the same question about the Win odds, the ML or any handicapper's odds lines.

In fact, I'd argue that most people's odds lines are wrong more often than the real odds and sometimes ridiculously wrong. That was one of the points of my other thread about how people determine when they actually have a good value bet.