HANDICAPPING IS RANDOM

[biggestal99]

Originally Posted by AskinHaskin

Well that enlightens the masses as to the general cluelessness in these hallowed halls.


Talk about inaccurate thinking.



Sure, if the results were (truly) random, that would be the case, but as horse race results are anything but random, the above quoted statement is baseless.

You need look no more than beyond the percentage of winning favorites to immediately see that race results are not 'random'.

"Random" is lottery number drawings.

The fastest horses win the most races. The fastest ping-pong balls know no greater success than the others.

If you pick random horses. You get random results.

If you picked random horses over the course of the year to win. You lose to the track take out. That’s it.

Suprisenly enough you would pick the fastest horse every so often.

Allan

[cj]

Just to set the record straight, if you pick a horse at random, you'll lose more than the takeout. This is due to breakage and also that longer odds horses tend to lose more than the takeout, much more. And if truly random, you'll be betting a lot of these horses.

[Robert Fischer]

Originally Posted by cj

Just to set the record straight, if you pick a horse at random, you'll lose more than the takeout. This is due to breakage and also that longer odds horses tend to lose more than the takeout, much more. And if truly random, you'll be betting a lot of these horses.

This is important.

And this is 'flat-betting' every horse hypothetically. Gamblers go broke, several times, well before they had a chance to bet random horses over a long period of time.


As to randomness in general - It's a horse race, not a procession!

randomness mixes up the individual outcomes and helps the gambling incentive a great deal...

and insiders try to profit from unaccounted extrinsic randomness

[AndyC]

Originally Posted by cj

Just to set the record straight, if you pick a horse at random, you'll lose more than the takeout. This is due to breakage and also that longer odds horses tend to lose more than the takeout, much more. And if truly random, you'll be betting a lot of these horses.

The breakage point I get but what percent of the random horses would be underlays to the breakage and what percent would be overlays? And wouldn't your random pick of the overlays offset the underlays?

[cj]

Originally Posted by AndyC

The breakage point I get but what percent of the random horses would be underlays to the breakage and what percent would be overlays? And wouldn't your random pick of the overlays offset the underlays?

How do you get an overlay to breakage? You either get what you should have without breakage or less.

My point is, as a group, horses at 2-1 lose less compared to takeout than horses at 10-1, and horses at 50-1 lose a lot more than horses at 10-1 compared to takeout. If truly picking randomly, you are going to get a lot of horses at big odds that are not going to lose only the takeout. I won't even out.

I've run simulations picking a random horse in each race and you always lose more than the takeout, and not just about the 1% or so breakage adds.

[biggestal99]

Originally Posted by cj

Just to set the record straight, if you pick a horse at random, you'll lose more than the takeout. This is due to breakage and also that longer odds horses tend to lose more than the takeout, much more. And if truly random, you'll be betting a lot of these horses.

I disagree. if you pick random you will end up close to the median of odds in a race so what about 7-1 or so. which will what about 10% of the time


Allan

[biggestal99]

Originally Posted by cj

I've run simulations picking a random horse in each race and you always lose more than the takeout, and not just about the 1% or so breakage adds.

since you've done the hard work.


how much of a sample did you use?


what was the winning percentage of random horse picking?


what was the average payout for hitting a random horse winner?


thanks for the data that you produce.


you are never too old to earn something.


Like the blinkers data that Tom produced the other day,


That was eye opening.



Allan

[castaway01]

Originally Posted by biggestal99

I disagree. if you pick random you will end up close to the median of odds in a race so what about 7-1 or so. which will what about 10% of the time


Allan

The odds aren't random though, just your picks are. There are a lot of times you'll end up with a random 50-1 shot in a race where the horse's true odds should be 200-1. Because the odds (no pun intended) are against you randomly ending up with that horse the 1 in 200 times it wins, you'll lose much higher than the takeout over time.

[Jeff P]

I penned an article for the HANA Blog back in December, 2013 that illustrates the effect CJ is talking about.

Turning Fans into Horseplayers:
http://blog.horseplayersassociation....seplayers.html

Players' Expectation and Random Selections in WPS pools
One of the hurdles the industry faces in converting the first time track visitor and new would be racing fan into a long term horseplayer is the severity of the negative gambling experience imprinted during those first few track visits.

How bad is it (really?)

If WPS takeout is 16% you might think the novice horseplayer, who doesn't know a thing about betting horses, and whose selections therefore might realistically be no better or no worse than random picks, should have an expected return no worse than 84 cents per each $1.00 wagered (1.00 minus the takeout percentage) right?

You might intuitively think that - but you'd be wrong.

Believe it or not, the true long term mathematical expectation of horse race bettors who make the equivalent of random selections in WPS pools is a net loss approaching 25 cents per each $1.00 wagered.

That's approximately 1.5 times worse than the percentage net loss most of us might intuitively think results from a 16 percent takeout rate!

Below is a cut and paste of some numbers pulled from my calendar year 2013 database. The database includes every thoroughbred race run in North America during calendar year 2013 from Jan 01, 2013 current through a few days ago Dec 08, 2013.

The top part of the data readout shows what would have happened to the player unfortunate enough to have bet $2.00 to WPS on every starter in every race during the time period covered by the database (approx 350k starters in all.)

The ROI numbers towards the bottom of the WPS columns reflect the return for each $1.00 wagered: Net losses bordering on minus 25 percent for WPS bets.

Data Window Settings:
Connected to: C:\JCapper\exe\JCapper2.mdb
999 Divisor  Odds Cap: None

SQL:  SELECT * FROM STARTERHISTORY
      WHERE [DATE] >= #01-01-2013#
      AND [DATE] <= #12-31-2013#
      ORDER BY [DATE], TRACK, RACE


Data Summary          Win         Place          Show
-----------------------------------------------------
Mutuel Totals   529184.70     529707.00     531006.60
Bet            -699134.00    -699134.00    -699134.00
-----------------------------------------------------
P/L            -169949.30    -169427.00    -168127.40

Wins                45440         90386        131813
Plays              349567        349567        349567
PCT                 .1300         .2586         .3771

ROI                0.7569        0.7577        0.7595
Avg Mut             11.65          5.86          4.03

Next up, the second part of the above sample with the data broken out by odds rank with no attempt to break ties:

By: Odds Rank

Rank       P/L        Bet        Roi    Wins   Plays     Pct     Impact
-----------------------------------------------------------------------
 1   -14901.00   93348.00     0.8404   17228   46674   .3691     2.8396  
 2   -15208.70   88526.00     0.8282    9642   44263   .2178     1.6758  
 3   -18372.90   88404.00     0.7922    6540   44202   .1480     1.1382  
 4   -18563.50   88650.00     0.7906    4681   44325   .1056     0.8124  
 5   -22870.40   87296.00     0.7380    3078   43648   .0705     0.5425  
 6   -22460.60   81236.00     0.7235    2034   40618   .0501     0.3852  
 7   -20465.90   65288.00     0.6865    1155   32644   .0354     0.2722  
 8   -14106.40   46530.00     0.6968     616   23265   .0265     0.2037  
 9   -10315.90   30140.00     0.6577     294   15070   .0195     0.1501  
10    -8129.40   17866.00     0.5450     110    8933   .0123     0.0947  
11    -2351.20    7388.00     0.6818      46    3694   .0125     0.0958  
12    -1884.80    3616.00     0.4788      13    1808   .0072     0.0553  
13     -245.60     578.00     0.5751       2     289   .0069     0.0532  
14      -63.00     258.00     0.7558       1     129   .0078     0.0596  
15       -2.00       2.00     0.0000       0       1   .0000     0.0000  
16       -2.00       2.00     0.0000       0       1   .0000     0.0000  
17       -2.00       2.00     0.0000       0       1   .0000     0.0000  
18       -2.00       2.00     0.0000       0       1   .0000     0.0000  
19+      -2.00       2.00     0.0000       0       1   .0000     0.0000

This part of the data sample shows the favorite-longshot bias and hints at why horse racing might be at least attractive to some players from a viable gambling standpoint.

Note that win rate and roi (or player's expectation) varies with the odds.

The player who makes a serious go at studying the past performance records of horses will discover that horses have both positive attributes and negative attributes in their records. Positive attributes incrementally increase the likelihood of strong placings. Conversely, negative attributes result in incrementally worsened placings. Of course, the betting public realizes this and a high percentage of those attributes (as collectively seen by the crowd) are reflected in the odds.

I don't think there's any question that the newbie would be horseplayer faces a steep uphill climb. But most of us assume the hill to be climbed is 1.00 minus the takeout rate. (As you can see from the above numbers that is clearly not the case.)

In my mind getting people out to the track by itself isn't going to cut it.

The real question as I see it is this:

How do you create a good enough GAMBLING EXPERIENCE for the new would be horseplayer to make them want to come back for more?... A LOT MORE? (Enough to where the "fan" turns into a horseplayer?)

It's hard to believe more than five years have passed since I wrote that article. (Seems like just yesterday to me.)

Even though the above numbers are from 2013, the same effect is evident in the data - no matter which year you are looking at.



-jp

.

[JJMartin]

Originally Posted by Jeff P

I penned an article for the HANA Blog back in December, 2013 that illustrates the effect CJ is talking about.

Turning Fans into Horseplayers:
http://blog.horseplayersassociation....seplayers.html




It's hard to believe more than five years have passed since I wrote that article. (Seems like just yesterday to me.)

Even though the above numbers are from 2013, the same effect is evident in the data - no matter which year you are looking at.



-jp

.

So then the top 2 or 3 favorites should be at least a slightly better value when big long shots are present versus a more evenly distributed field.

[biggestal99]

Originally Posted by Jeff P

I penned an article for the HANA Blog back in December, 2013 that illustrates the effect CJ is talking about.

Turning Fans into Horseplayers:
http://blog.horseplayersassociation....seplayers.html




It's hard to believe more than five years have passed since I wrote that article. (Seems like just yesterday to me.)

Even though the above numbers are from 2013, the same effect is evident in the data - no matter which year you are looking at.



-jp

.

I will change my mind.

Randomly picking from the first three choices will give you losing to track takeout.

You can’t beat facts.

Allan

[AndyC]

Originally Posted by cj

How do you get an overlay to breakage? You either get what you should have without breakage or less.

My point is, as a group, horses at 2-1 lose less compared to takeout than horses at 10-1, and horses at 50-1 lose a lot more than horses at 10-1 compared to takeout. If truly picking randomly, you are going to get a lot of horses at big odds that are not going to lose only the takeout. I won't even out.

I've run simulations picking a random horse in each race and you always lose more than the takeout, and not just about the 1% or so breakage adds.

I meant an overlay to the takeout.

After reading the responses and seeing Jeff's random bet study I have finally gotten my head around this issue. Betting equal amounts on random selections surely produces higher than takeout + breakage losses. And the reason why is because of the equal weight betting. If every horse were bet in the same percentage as actually bet in the pools the result would be a loss of the takeout + breakage.

Given that many players handicapping prowess resembles a random selection the takeaway would be to bet more on lower priced horses and less on longshots.

[JJMartin]

Originally Posted by AndyC

I meant an overlay to the takeout.

After reading the responses and seeing Jeff's random bet study I have finally gotten my head around this issue. Betting equal amounts on random selections surely produces higher than takeout + breakage losses. And the reason why is because of the equal weight betting. If every horse were bet in the same percentage as actually bet in the pools the result would be a loss of the takeout + breakage.

Given that many players handicapping prowess resembles a random selection the takeaway would be to bet more on lower priced horses and less on longshots.

Then you would lose even more money but with a lower negative ROI.

[Robert Fischer]

if you stick to favorites and big fields for your random selections, you will be closer to takeout and a better value relative to long shots, - - however, these will not be overlays in any sense of the word.

If you get offered a 25% rebate; by all means


otherwise this stuff is just masturbation


It's a parimutuel game. You need significant opinions of where the public is wrong. You need to actually to be able to see and understand the game.

[JJMartin]

Originally Posted by Robert Fischer

if you stick to favorites and big fields for your random selections, you will be closer to takeout and a better value relative to long shots, - - however, these will not be overlays in any sense of the word.

If you get offered a 25% rebate; by all means


otherwise this stuff is just masturbation


It's a parimutuel game. You need significant opinions of where the public is wrong. You need to actually to be able to see and understand the game.

They may not all be overlays but if the premise is correct, the value should increase on the lower priced horses making such race scenarios more profitable especially if you are not a random capper.