PDA

View Full Version : Mining data for patterns paid off


patternfinder
29th May 2007, 07:33.56 PM
Hello,
Well, yesterday was my first day betting on horse racing. I built some data mining programs which look for patterns and make recommendations as to which horses to bet on. Yesterday, I made two types of bets, either $100 to win on a single horse, or a $100 dutch bet on 3-4 horses (I found about a dozen of these, most of them won). At the end of the day I was up $9,703.00. I bet $100 on every race at 12 tracks. I don't know what the total number of races was off the top of my head, 130+ I guess.

I would like for someone who knows how to handicap to look at the picks that my data mining programs made and tell me what your handicapping program would have said to do on these picks. I am looking for someone to work with where I recommend horses to bet on based on my data mining tools and the other person evaluates the picks using their handicapping tools. And when both programs agree, we bet on those picks.

I have attached an excel spreadsheet which lists all of my picks, the winners have the odds next to them. Note, that the odds I list are not the track odds, I bet on bet matching sites, which is why I limited the bet size to $100. It is tough to get filled on orders above that sometimes. I have an idea why this data mining program worked, I'll share this with people who give me good feedback. To read the spreadsheet, the track is the name of the worksheet, the row number is the race number, the entry in the first column is the horse I bet on, if the horse won, the odds are in the second column. If the first column has more than one entry in it, that was a Dutch bet. The last row is the total amount bet for the first column, the total payoffs in the second column, and the total amount won (total payoffs - total amount bet) in the third column. It was much easier to total things this way than to try to figure out "let's see, 5/1 is actually 4 to 1 because you have to deduct the amount bet..." geesh I have a headache. So, I just multiplied the odds x amount bet and deducted the total amount bet at the end. Comes out the same way and much easier to calculate.

Patternfinder

njcurveball
29th May 2007, 07:48.08 PM
I have attached an excel spreadsheet which lists all of my picks, the winners have the odds next to them.

Congrats on your great day! I opened your spreadsheet and found tracks and numbers next to them. I am guessing this is the amount won or lost by track.

I wouldn't know how to recommend a change of strategy to someone who won over $500 at 9 of the 12 tracks played.

$5,000 a day on $100 bets is the Dalai Lama in my eyes, so you do not need any help from where I sit.

Even if you win half of that you can be sure there are maybe a dozen people in the entire country making more and that is not on win bets.

Congrats and keep up the good work!
Jim

DanG
29th May 2007, 07:54.35 PM
I wouldn't know how to recommend a change of strategy to someone who won over $500 at 9 of the 12 tracks played.

$5,000 a day on $100 bets is the Dalai Lama in my eyes, so you do not need any help from where I sit.
Jim
Agreed Jim;
========================

Steve, when you wrote me the other day I didnít dream you were going to lay $13k on the table your first time betting racing!!!

That is supreme confidence my friend.

Best of luck and continue success!

patternfinder
29th May 2007, 07:55.45 PM
Hi Everyone,
I forgot to mention that the first worksheet is the tally by track. I have a hunch that some tracks are better for mining data than others. Smarter players, or dumber players, depending on how you look at it. I'm going to keep this tally over time and see if I am right. I saw a chart on here somewhere where someone listed the tracks by how likely the track was to produce a longshot, same basic idea.
Patternfinder

MikeDee
30th May 2007, 06:16.12 AM
P_F If you make 10K a day, I don't think there is a handicapper in the universe that can help you improve on that. You're in a league of your own.

njcurveball
30th May 2007, 08:47.12 AM
I put my "excel" hat on and went through the different spreadsheets.

In one race at Hollywood you listed

1,2,4,5,6,7,8,9,10

So you had $900 invested in this race? The winners odds were 1.65, so your return was $265? Ouch! At least I can offer the advice of staying away from flat betting multiple horses if the results are negative.

Same as the first race at Hollywood

1,2,4,5,6
I guess the winner here got you $262 for your $500. Another ouch!


Looks like the other 8 horses got you 1 winner at 4.6, so you bet $800 to win $560.

Now what I do not understand is your tally of $1,000 bet, $787 returned?

I counted $2,200 bet and $1,087 returned.

I am sure you know the figures much better than me. I just went by your statement of $100 to win and the numbers in Column A.

Jim

njcurveball
30th May 2007, 09:00.56 AM
I checked some more tracks and I have to admit I am totally confused by the spreadsheet.

If you bet $100 and you are getting 4-1 odds, your return will be $500.

I check Churchill with no multiple horses.

Looks like 4 wins at 4, 4.65, 10.8, and 4.1. These add up to 23.55 which is your number for money won.

As for the multiple horse bets, did you bet $10 each when you list 10 horses?

Donnie
30th May 2007, 12:14.29 PM
Quick question.....

"I bet on bet matching sites".....what exactly does this line mean....

BetFair? which means you are overseas?

Also, I would like to see how you dutched the 5 out of 6 horses in race 1 at HOL ..... with 2 or 3 horses @ odds of 3-1, .....??

Donnie
30th May 2007, 12:21.35 PM
I am lost as well Jim....

CD Race two winner is 2.90/1....pays 7.80
Race eight winner is 2.00/1.... pays 6.00


His second column is 4 for the 2nd race and 4.1 for the eighth race. Betting $100 per horse shouldn't there be a "slight" difference if you have a full point in odds spread?

BTW...last time I looked, my program told me to bet the winner!:D

njcurveball
30th May 2007, 12:37.03 PM
I am looking for someone to work with where I recommend horses to bet on based on my data mining tools and the other person evaluates the picks using their handicapping tools. And when both programs agree, we bet on those picks.



This is a sheet of ice trying to make one program agree with another. Most of the profit comes from higher odds horses. When more and more things agree, prices drop.

I set up something similar and I can generate 70-80% winners with those horses. Unfortunately, the number of plays with 20 tracks is less than one per track per month.

The great thing about HTR is that we have the $ and $$ symbols to alert us with horses who do not show a lot of positives. But they do show enough that they win at nice prices,

The main thing you have to remember with patterns and trends is that sometimes they are just that. 3 red lights in a row does not mean you have a better or worse chance of seeing red on the next light (even though we all might agree we do).

Flipping a coin to see 3 heads in a row also does not increase the chances for the next toss. Unless it is a two headed coin. ;-)

Patterns are specific to the type of horses, the surface, and the track. For those that keep models, they understand that at 5 1/2 furlongs some tracks are dominated by the 3rd fraction. Others by the first fraction.

You are definitely on the right track, Mr. PF so keep up the good work and keep in mind this game may not be as easy as it looks on the first day.


Jim

MVM
30th May 2007, 02:27.16 PM
OK, I'll admit that I am somewhat lost by this thread, but I'll give this a shot.

Based on PFs original posting:

I make part of my living mining data for patterns. My clients are big companies who are trying to solve tough problems.

I have to warn anyone reading this, I know nothing about horse racing. What I do know about is finding patterns in data.

Something that I am curious to test out is how good are fans at a given racetrack at picking winners?

Since the bettors set the odds for the race, and they bet on the horses they think are most likely to win, I'm wondering how good they are at it?

My premise is that for a given racetrack, if you looked at the results of all the races for, say one year, and compared the actual results to the odds, the results might show that some racing fans are better than others at picking winners. Perhaps some racetracks would show actual results consistent with the odds, and other racetracks would show a large inconsistencey between the actual results and the odds.

This would be predicated, of course, on the assumption that the same fans go to the same tracks often.

Has anyone done this analysis before? I downloaded HTR, but it is going to take some phone calls to customer support for me to get it to actually work.

I am guessing that by using a betting exchange he limited his play to the horse(s) that were offered at prices 20% or more higher than the actual tote odds.

Judging by the posting above, PF has not yet accumulated the historical data necessary to "mine" the PPs for patterns (otherwise he would have done the bettor's odds/outcomes analysis himself).

However, his assumption that you can make a profit by consistently getting a solid premium over the actual tote odds is fundamentally sound. My recommendation would be to leave this strategy as is...actually handicapping the races based on PP data would likely actually hurt your bottom line.

tomcat
30th May 2007, 02:31.57 PM
The challenge is: how to get the biggest bang for the bucks. As Mike pointed out, if you can make that much money day-after-day you don't need anyones help. Instead of a puny $100 per play, maybe you bet a car or house per play...just kidding.
It seemed that others pointed that the results seemed to differ with the tracks. I think the research is just beginning. For example does your play seem to favor certain tracks or certain types of races (Mdn, Claming, etc.) or surface= dirt, turf, poly?
I would also suggest an in-the-money count of the above.
The results of all this study could lead to even more winners or fewer losses. If the "in the money" result was fairly high you might start to look at Exactas, Tris and other exotics.

Keep up the good work. keep us posted.

AwolAtHTR
30th May 2007, 02:33.33 PM
Hello,
Well, yesterday was my first day betting on horse racing.
---A==the winners have the odds next to them. Note, that the odds I list are not the track odds, I bet on bet matching sites,

--B== I have an idea why this data mining program worked, I'll share this
--


Wed 30 May 2007

yes, PatternFinder those are amazing results

Refer A==: However, I have found more than one method that the FIRST time was amazing but then with more history, the results could not be duplicated!! A gain of about 9K for about 13K bet is clearly a method with a lot of potential.

Refer B==: My data base tool is Access 97 with subgroup definitions using Visual Basic. I have no experience with commercial data mining tool but believe that I can define, and duplicate, any set of rules using VB to define recognition and betting rules. However, I expect your offer to share means that you want validation of the bets and not to have the rules duplicated by a tool different than the software which you are using to find the bets.

Email and Private messages are disabled on this HTR forum.

email: awol awol -at- hotmail -dot- com
or
phone: 5 0 9 7 3 8 6 4 1 8

looking forward to helping you confirm the results for 20070529 can be duplicated!!

Rick
30th May 2007, 02:36.36 PM
Emails are only disabled by the individuals.

PM is turned off.

ERNIE LOGSDON
30th May 2007, 07:50.00 PM
It's the little things that bug me the most,when I read that some one has found such a promising method to bet horses (and on their very first day )......
never mind we are given a very well thought out spread sheet and a better way of looking at the payouts-to- one dollar vs the expense of making such a bet any where in the world,but at CRC you listed 13 races that you bet $100 each( only deducted for 10 )--at LS,MTH and CD you bet 11 races for $1000------all of that is exceptable in my mind,@#$%^&$% I can't even spell---and I will be interested in buying anything this good "if it's for sale" BUT you MUST tell me how you first found and then was able to get a bet down on the #1 horse IN RACE TEN AT LAD on a day when they only ran 9 races. Here is hoping that you check-n-double check the data for your large clients to insure..........
Good Luck here's hoping that you contenue to enjoy the wonderful world of horse racing, research,beting and shareing with us your sucsess.

rman1049
30th May 2007, 08:00.37 PM
What has me scratching my head is what parameters did you use come out with these plays??
I would also be very interested in purchasing your data, if some data can be supplied....

Thanks Ralph

rmanfr1049@aol.com

patternfinder
30th May 2007, 08:06.35 PM
Hello,
Thanks to everyone for the feedback. For those who were confused by the Dutch bets, I will try to explain better. I built a Dutch betting spreadsheet and hooked up some data feeds to the spreadsheet. In every race I immediately cross off the favorite. If all the rest of the horses combined have a Dutch bet odds of 1.65 or higher, I just bet them all. That is why you saw half a dozen races where I bet the entire field except the favorite. My Dutch betting calculator just takes $100 and divides it up among all the horses bet so that they all pay the same amount if they win, so the bet sizes will vary. The win will always be the aggregate odds on these bets. For the dutch bets, if I won, the odds listed are the aggregate odds x $100. If anyone wants a detailed example of one of these bets, I can send it to you, but there are Dutch bet calculators online. For the races where the Dutch bet was only a few horses that means that the pattern was too close to call on those horses, there was no way to pick just one of them, and the odds on them were great, so I just bet them all. But again, it was $100 spread among all of them. Not $100 each.

Before I actually bet any money, I did data mining on historic results going back a few months. I actually found the pattern that I ended up using by accident. I was back testing a pattern that I thought looked promising and that pattern was TERRIBLE, so I just did the opposite of it.

It is difficult for me to explain how the data mining works, because what I see are patterns in the numbers, without understand what the numbers actually mean. I know this sounds crazy but I do this kind of pattern finding all the time without knowing what the categories mean, sometimes I dont' even label the axes. Its kind of like if you printed out the value of PI to a few thousand places and had nothing else to do on a rainy Sunday, after looking at the numbers for a while you wll start to see repeating patterns. Its kind of fun actually.

I did test this pattern live for a few days without actually betting, and I modified it slightly because I saw a lot of unusual betting right at the start of the race. What do you guys call that? The post? The gate? This money comes in literally a second before the start. It is obviously computer trading. I have no idea why these bets are being made. This pattern was so obvious it literally jumped off the screen. I have no idea what it means in racing terms.

I'm not using a handicapping program, I'm not looking for the kinds of leading indicators handicappers look for, I'm just looking at raw numbers in the betting patterns. But, I ignore the favorite. Why? Because if all you did was bet the favorite race after race, how long before you go broke? I don't think the smart money bets the favorite.

Keep in mind that 130 trials means absolutely nothing. So, I am not saying that these results mean anything. That is why I want a handicapper to evaluate my picks and see if there is a reason that this horse might win.

I don't think that picking 52 winners in 130 races is a fluke. That can't be dumb luck. And if you look at the horses I did pick they all share a common characteristic. Leaving out the dutch bets, the odds on the winners were between 4/1 and 7/1 mostly with a few above 7/1. And a lot of money went down right at the gate on all of these. That isn't a fluke. I am guessing that when these picks are put into handicapping software they are going to show some similar characterstics that mean something in terms of the ability to win a race. LIke the jockey is stoned, or the trainer has nice breasts.... just kidding. But, you know what I mean.

Patternfinder

patternfinder
30th May 2007, 08:17.28 PM
someone pointed out some small errors in math. I put this spreadsheet together at 2:00 am. And I had to do a lot of cutting and pasting from another program. The mistake for Louisana Downs, where I had 10 results listed and they only ran 9 races, I had the winner and the runner up horse for the 9th race when I copied the data over, so I had an extra row, but if you look at the entry for row 10, that is horse #1, who was runner up in race 9. The other mistake where I always had 1000 in bet total even at tracks that had more than 10 races was due to the fact that the formula I wanted to use in excel wouldn't work and I just copied and pasted from the first sheet, being it was 2:00 AM.

After correcting the errors, the total profit for the day was only $9,203. sorry. I could have just looked at my balance in the site where I bet. <laughing> Which I just did.

MVM
30th May 2007, 08:47.46 PM
I'm even more lost now....are we talking about betting into the mutuel pool or a betting exchange?

Are the number patterns in race PP data or the flow of money on the toteboard?

From what I read, it would absolutely have to be the latter of the 2 in both cases.

If you have somehow found a betting exchange that allows bets after the off bell, allowing you to leverage this late tote information, that is definitely something we would all be interested in hearing about.

One thing I would be careful about if you are betting into the mutuel pools, the "bet all horses except the favorite if the dutch bet odds are 1.65 or higher". You are living on borrowed time with this one, as you can only get the required dutch odds when the favorite is 3-5 or lower.....those 3-5 shots (over the long run) are not going to lose at the 60% required to break even on this bet.

It's great weather for a change up here today, so i'm gonna take off these boots and try and soak up some rays, I've actualy heard that the rays emitted by our closest star are a superior antibacterial.

Donnie
30th May 2007, 08:53.08 PM
Steve (Patternfinder)-
How did you find several months of historic data? Most of the vendors online only hold a month or two's worth of results/racefiles at any one time. Not knowing anything about racing, you seem to have accomplished more than the average punter, in terms of setting up your approach and attack.
Also I wonder how you managed the different tracks with the same post times , and betting multiple horses per race while waiting till the last second to get your bets down? I think short of using a betting interface this task is almost impossible. But I would be interested in hearing your method of "getting the bets down". Not to mention how do the "betting match sites" play into this?


Best of Luck!

Donnie
30th May 2007, 08:55.13 PM
Mike-
We have your rain...and anytime you want it back, just holler! Everyone coming to DSM this Friday---bring your boat!!

patternfinder
30th May 2007, 10:03.38 PM
I'm even more lost now....are we talking about betting into the mutuel pool or a betting exchange?

betting exchange



Are the number patterns in race PP data or the flow of money on the toteboard?

flow, but not on the board at the track. I watch several screens of betting exchange data.


From what I read, it would absolutely have to be the latter of the 2 in both cases.
betting exchange and flow of bets on several exchanges. I have multiple monitors showing different things.


If you have somehow found a betting exchange that allows bets after the off bell, allowing you to leverage this late tote information, that is definitely something we would all be interested in hearing about.

No, I didn't say that. It is sometimes tricky to know when the race is starting, the patterns dont work if I go too soon. I can tell when the race is starting by the bet flow, there is a sharp drop off just as the race starts.


One thing I would be careful about if you are betting into the mutuel pools, the "bet all horses except the favorite if the dutch bet odds are 1.65 or higher". You are living on borrowed time with this one, as you can only get the required dutch odds when the favorite is 3-5 or lower.....those 3-5 shots (over the long run) are not going to lose at the 60% required to break even on this bet.


I disagree, I am going by what the pattern says, it is almost always right. Especially on the Dutch bets. Look at my results and you will see what I mean.

patternfinder
30th May 2007, 10:09.38 PM
Steve (Patternfinder)-
How did you find several months of historic data? Most of the vendors online only hold a month or two's worth of results/racefiles at any one time. Not knowing anything about racing, you seem to have accomplished more than the average punter, in terms of setting up your approach and attack.
Also I wonder how you managed the different tracks with the same post times , and betting multiple horses per race while waiting till the last second to get your bets down? I think short of using a betting interface this task is almost impossible. But I would be interested in hearing your method of "getting the bets down". Not to mention how do the "betting match sites" play into this?


Best of Luck!

I searched long and hard to find this data. I don't know anything about racing, but I know how to do research. And I know how to build trading models, and I know how to backtest.

I managed the different tracks with several different screens up at the same time, I have several large monitors. My office looks a bit like NASA. It is very easy to bet multiple horses on multiple races when you have multiple betting programs running, I can bet with one click. The dutch bets are calculated for me, then I just click. I wrote these programs. I don't bet at the track, I only bet on match sites, but someone told me today that the horses I am betting had better odds at the tracks. But, the data I use to determine the bets is coming from the match sites. I think it will be too complicated to be watching the patterns in one screen, then switching to another screen to place the bets. But, I might be able to automate that.

I'm not sure if "betting match site" is the right term. It is an electronic bet matching site. One person creates an order for a bet, another person accepts the bet, the bets are matched electronically. The site charges 3% commission. I have accounts at several.

patternfinder
30th May 2007, 10:15.30 PM
What has me scratching my head is what parameters did you use come out with these plays??
I would also be very interested in purchasing your data, if some data can be supplied....

Thanks Ralph

rmanfr1049@aol.com

I didn't use "parameters". I just looked at amounts bet. At first I did back testing on historic data, then I did live testing while paper trading. After doing the live testing for a while I modified the trading rules to also watch bet volume right at the gate, from about 10 seconds before starting up until you can't place any more bets. This was actually the best indicator. You'll have to help me to get more information than that. I did actual betting on Monday using the modified trading rules. It worked very well.

I'm still looking for someone to put my picks into a handicap type program and see if there are any patterns that deal with factors that handicappers care about. I would be willing to bet that there are some patterns there. I don't know enough about this to evaluate that. So far, no one has done that. That is what I am interested in talking about.

patternfinder
30th May 2007, 10:26.03 PM
The challenge is: how to get the biggest bang for the bucks. As Mike pointed out, if you can make that much money day-after-day you don't need anyones help. Instead of a puny $100 per play, maybe you bet a car or house per play...just kidding.
You can't bet more than $100 on the sites I use, no one will match the bet. Well, they might eventually, but time is of the essence and you can't take the chance you wont get a matched bet. I sometimes spread the bet among multiple sites, depending on how soon I am making the bet and what kind of volume is on that race at a given site. I also have to sometimes offer better odds to get a matched bet.

It seemed that others pointed that the results seemed to differ with the tracks. I think the research is just beginning. For example does your play seem to favor certain tracks or certain types of races (Mdn, Claming, etc.) or surface= dirt, turf, poly?[/QUOTE}
I have no idea what these terms mean. I just look for patterns in money flow. Later I will do this kind of analysis. Right now the sample size is too small. I will say that I am using 12 tracks. Someone asked me why I picked those 12, the answer is that they are the ones that the betting sites I use will let me bet on. I also don't bet, what is it called? The races where the horses are pulling the wagon. I forget what it is called. Wagon racing? I dont bet on those. Dont ask me why. It isn't rational.

[QUOTE]
I would also suggest an in-the-money count of the above.
The results of all this study could lead to even more winners or fewer losses. If the "in the money" result was fairly high you might start to look at Exactas, Tris and other exotics.

Keep up the good work. keep us posted.

Yes, this seems possible. I started looking at this. One issue is that I have to watch the money flows of different races on different sites at the same time. Then bet before the race starts, right before. I have to do all the calculations in my head. Because I used to day trade stocks, this is okay. But, doing trifecta calculations in your head? On multiple races on multiple screens on multiple sites? I'm not the rainman.

You have my spreadsheet. You can look at the tracks and see if the ones that were big winners have something in common. I suspect that is where the hottest babes go to watch races.

patternfinder
30th May 2007, 10:32.35 PM
I checked some more tracks and I have to admit I am totally confused by the spreadsheet.

If you bet $100 and you are getting 4-1 odds, your return will be $500.

I check Churchill with no multiple horses.

Looks like 4 wins at 4, 4.65, 10.8, and 4.1. These add up to 23.55 which is your number for money won.

As for the multiple horse bets, did you bet $10 each when you list 10 horses?

No, not $10 each. I bet so that any horse that won would payoff the same amount. That means different amounts for different horses because they are at different odds. I don't think I bet on 10 horses in any race. I'll have to check that. I did have one race where I bet 9 horses. The 1.62 odds are the combined odds. So, for the $100 bet, the payoff was $162. It didn't matter which horse won, the payoff was $162. Some of the long shots you would only bet $2-3 on, others you would bet more. I have a program that figures all this out for me. I just have to be able to do a quick calculation in my head to see if a Dutch bet is feasible in a given race, then I put the numbers into the spreadsheet, it tells me what to bet, then I hit a button on the spreadsheet and it feeds the Dutch bet to another program.

patternfinder
30th May 2007, 10:45.51 PM
Quick question.....

"I bet on bet matching sites".....what exactly does this line mean....

BetFair? which means you are overseas?

Also, I would like to see how you dutched the 5 out of 6 horses in race 1 at HOL ..... with 2 or 3 horses @ odds of 3-1, .....??

I use sites that act as electronic bet matchers. One person offers a bet, another person offers a matching bet, the site matches the bets up electronically. Its kind of like a stock trading site. One person offers a lay bet at 10-1 odds, another person offers a for bet at 10-1 odds, the site matches the bets electronically.

I dont use BetFair, but they do something similar to the sites I do use. My profile says where I am.

The total odds on the 5 horse dutch were 1.62 for 1. Not great, but doable. There weren't 3 horses at 3-1. One I think. I get better odds than the tote board at the track on the longshots, much better. Slightly worse odds on the lower odd horses. Makes the dutch plays doable when they don't look that way using the tote board odds. Assuming you have picked the right dutch play. <grin>

Catch me online some time when races are going on and I will tell you the results of the races before they run. That will make it more real to you.

patternfinder
30th May 2007, 10:57.46 PM
I am lost as well Jim....

CD Race two winner is 2.90/1....pays 7.80
Race eight winner is 2.00/1.... pays 6.00


His second column is 4 for the 2nd race and 4.1 for the eighth race. Betting $100 per horse shouldn't there be a "slight" difference if you have a full point in odds spread?

BTW...last time I looked, my program told me to bet the winner!:D

You are using X TO y odds, I am using x FOR y odds. The difference between TO and FOR is, if you bet $100 and I pay you 2 TO 1, I pay you $200 plus I give you your original bet back, for a total of $300. If you bet $100 and I pay you 2 FOR 1, I give you $200 including your original bet, so to calculate how much you actually won, you have to subtract your original bet, which means you won $100. I'm sorry I didn't make that clear.

The actual payoff was pretty close to the track odds.
Using your numbers with $100 bet: My numbers with $100 bet
Race 2 2.9 TO 1 = 390.00 Race 2 4 FOR 1 =400.00-100 (original bet) = 300.00
Race 8 2 TO 1 = 300.00 Race 8 4.1 FOR 1=410-100 (original bet) = 310.00

Assuming you bet at the track and I bet at the match site, you did a little bit better than me on this one. But the track doesn't give me the data I need.

patternfinder
30th May 2007, 11:05.59 PM
OK, I'll admit that I am somewhat lost by this thread, but I'll give this a shot.

I am guessing that by using a betting exchange he limited his play to the horse(s) that were offered at prices 20% or more higher than the actual tote odds.
No, look at the spreadsheets and you will see the odds that I was paid, most are slightly below the track odds. Only the longshots were higher than the track odds. The odds have nothing to do with it.


Judging by the posting above, PF has not yet accumulated the historical data necessary to "mine" the PPs for patterns (otherwise he would have done the bettor's odds/outcomes analysis himself).

I am using one year of historic data which I have backtested, plus several weekends of live testing while paper betting, and one day of actual betting. I am already mining for patterns, that is how I know how to do this. What I don't know is why do these horses win from a "horse racing" perspective. I can already tell you which tracks have the smartest (or dumbest) bettors. I am looking for confirmation on my theories from people who know what they are talking about.


However, his assumption that you can make a profit by consistently getting a solid premium over the actual tote odds is fundamentally sound.
That isn't what I am doing, and I actually get slightly worse odds on most bets. I don't even look at the tote odds. I couldn't tell you on any one bet if I am getting a premium or paying a premium on the odds. And I don't care.


My recommendation would be to leave this strategy as is...actually handicapping the races based on PP data would likely actually hurt your bottom line.

I disagree. I think there is a "horse racing" reason why my picks win. And the reason they show up in the patterns is because smart people are betting these horses. Otherwise there would be nothing to see in the patterns. I also think that whatever these "horse racing" factors are they are not obvious to amateurs.

MVM
30th May 2007, 11:48.47 PM
I have a program that figures all this out for me. I just have to be able to do a quick calculation in my head to see if a Dutch bet is feasible in a given race, then I put the numbers into the spreadsheet, it tells me what to bet, then I hit a button on the spreadsheet and it feeds the Dutch bet to another program.

I built some data mining programs which look for patterns and make recommendations as to which horses to bet on.


One issue is that I have to watch the money flows of different races on different sites at the same time. Then bet before the race starts, right before. I have to do all the calculations in my head. Because I used to day trade stocks, this is okay. But, doing trifecta calculations in your head? On multiple races on multiple screens on multiple sites? I'm not the rainman.


I'm dizzy, and I have no idea where the data-mining program(s?) fit into this.

Quote #1 definitely details a rather circuitous route though (you don't happen to pass through Minneapolis on your way from Hartford to Suffolk do you?) I think with a little interface work, you could get rid of the spreadsheet (??) and multiple screens and let the computer do more of the heavy lifting.

Having some passing familiarity with the data mining field myself, I was curious about some of your approaches (I'm always looking to pick the brains of fellow QAs).

What kind of data warehousing schema are you using (or planning on using)? Myself, i'm far from a purist, and store far too many derivable fields! Intake once, output infinite is my guiding mantra.

Are you using diamonds, trees, cubes or some other type of mining model? With race data, i've definitely found one of the aforementioned 3 to be the most useful, but probably because i'm lazy and find it easier to work in a GUI (that gave away my preference).

And with the money flow data you are mining, i'm curious how frequent your refresh rate is, and whether most of the (positive or negative) patterns that have been discovered are a result of moving though the plane in any particular direction?


I'm still looking for someone to put my picks into a handicap type program and see if there are any patterns that deal with factors that handicappers care about. I would be willing to bet that there are some patterns there. I don't know enough about this to evaluate that.

Ah, the crux of the matter. If someone were to identify all the selections (both winners and losers), and work backwards from this end result 9the selection), would you want to know what differentiated the winners from the losers, or what common traits (patterns) the selections as a whole shared?

No idea why, but the last sentence in the quote above "I don't know enough about this to evaluate that." would seem to be something that you would want to keep intact, given your past successes with pattern identification sans knowledge of the related subject matter. Besides, upon reading that statement, I was somehow reminded of a quote from the 1:22 minute mark of the SNL skit below...

http://www.youtube.com/watch?v=1n7mP3UixfU&NR=1

patternfinder
31st May 2007, 06:30.09 AM
I'm dizzy, and I have no idea where the data-mining program(s?) fit into this.

Quote #1 definitely details a rather circuitous route though (you don't happen to pass through Minneapolis on your way from Hartford to Suffolk do you?) I think with a little interface work, you could get rid of the spreadsheet (??) and multiple screens and let the computer do more of the heavy lifting.

Eventually I will. I have two other full time interests, so my time to devote to this is a bit limited.


Having some passing familiarity with the data mining field myself, I was curious about some of your approaches (I'm always looking to pick the brains of fellow QAs).

What kind of data warehousing schema are you using (or planning on using)? Myself, i'm far from a purist, and store far too many derivable fields! Intake once, output infinite is my guiding mantra. [/QUOTE

I keep things as simple as possible. I didn't have any assumptions about the data, I sorted the horses by total $ of bets, and looked to see if any patterns showed up in which of these horses won, then I backtested, then I live tested. There is more to it than that, but I'm not going to share that with anyone until I find a partner and they will have to sign a non-disclosure agreement.

[QUOTE}
Are you using diamonds, trees, cubes or some other type of mining model? With race data, i've definitely found one of the aforementioned 3 to be the most useful, but probably because i'm lazy and find it easier to work in a GUI (that gave away my preference).
[/QUOTE

No, I don't use these. I just put the columns of data on the screen and scan for patterns (with my eyeballs). While live testing I found the patterns I actuallly use, these are all based on bet volume during the last few seconds before the race starts.

[QUOTE}
And with the money flow data you are mining, i'm curious how frequent your refresh rate is, and whether most of the (positive or negative) patterns that have been discovered are a result of moving though the plane in any particular direction?

There is no refresh rate, the data is streamed in real time, kind of like playing a song which is actually sitting on the server. Or, another way to say it, the refresh rate is instantaneous. I don't know what your second question means.



Ah, the crux of the matter. If someone were to identify all the selections (both winners and losers), and work backwards from this end result 9the selection), would you want to know what differentiated the winners from the losers,

I don't care about this.


or what common traits (patterns) the selections as a whole shared?

I would want to know what were the common traits among my picks. Break my picks into two groups, winners and losers. What is common among the winners? What is common among the losers? Someone was helping me live test yesterday and he kept asking me "what next two horses do you like?". Then he would see what the 3 picks I made did in terms of win-place-show, perfecta, and trifecta. At first I was irritated at this, but we hit so many perfectas I soon got over it. Note, we were not betting real money during this testing, just paper trading.


No idea why, but the last sentence in the quote above "I don't know enough about this to evaluate that." would seem to be something that you would want to keep intact, given your past successes with pattern identification sans knowledge of the related subject matter.

Theoretcially, true. I don't have the free time to become expert enough on horse racing to figure out what the handicapping factors mean in a reasonable time frame. Which is why I am looking for a partner. Note, when I say partner I don't mean "I give you picks for free and you bet with my money". Whoever I do this with will have to put up a stake equal to mine. I'd rather not say what that is in an open forum.

I'm looking for someone who can take any of my picks and say "well of course, #2 won in the third race, that was HolyCow ridden by JimBob, the #1 ranked jockey, owned by Bubbles, the #2 ranked trainer, and HolyCow likes to run on turf and JimBob was high and Bubbles had on a really short skirt".

Donnie
31st May 2007, 07:14.39 AM
ummm....so with last second decisions being made here, how does your "partner" gain the same advantage, when they (you) wouldn't know who or how much you are going to bet until the very last seconds before the race. Not to mention they would probably not have the same setup you have? Or does this...

Which is why I am looking for a partner. Note, when I say partner I don't mean "I give you picks for free and you bet with my money". Whoever I do this with will have to put up a stake equal to mine. I'd rather not say what that is in an open forum.

...mean they just give you their money to bet? Why not post the amount in an open forum? You posted your win $$. Don't be shy. You need to cull the deadwood out anyway....might as well be just as up front with the requirements where all can see. For a second I thought that you were just gonna give them your picks and they would bet on their own. Suddenly, to me, your partner ship is gonna cost someone a few hundred thou to get involved. Correct me if I am wrong, but I tend to read between lines....more is said there than anywhere else. (you seem to think the jockeys are drugged and the trainers are all women as well. But that is just one of your inferences that really throw me off. Not exactly professional remarks by someone who is as knowledgable as you at first come off as.)

One other question.....why go back to paper after such a huge score on Monday? Also why change to perfecta betting? Suddenly the entire dynamics has changed, in less than 3 days no less!!

Sounds like you have the golden goose by the throat. Like KM says, "The only true test is with your real money". If you have a winning angle or method or system or program, why bother digging deeper into anything....?

MVM
31st May 2007, 09:45.24 AM
There is no refresh rate, the data is streamed in real time, kind of like playing a song which is actually sitting on the server. Or, another way to say it, the refresh rate is instantaneous. I don't know what your second question means.


I was referring to the stored data that you are mining. Don't worry about the other 4 questions, you told me all I need to know.


Note, when I say partner I don't mean "I give you picks for free and you bet with my money". Whoever I do this with will have to put up a stake equal to mine. I'd rather not say what that is in an open forum.

njcurveball
31st May 2007, 10:05.42 AM
I'm looking for someone who can take any of my picks and say "well of course, #2 won in the third race, that was HolyCow ridden by JimBob, the #1 ranked jockey, owned by Bubbles, the #2 ranked trainer, and HolyCow likes to run on turf and JimBob was high and Bubbles had on a really short skirt".

This is the easiest part of all that you wrote. Anyone with HTR can do this and you could as well if you had the program. Are you already a subscriber?

Of course you don't have to subscribe to tell which trainers have the short skirts. I can honestly say I have never seen a trainer in the winners circle with a mini skirt on.

BUT a very good idea to attract the customers! :D

patternfinder
31st May 2007, 05:49.15 PM
ummm....so with last second decisions being made here, how does your "partner" gain the same advantage, when they (you) wouldn't know who or how much you are going to bet until the very last seconds before the race. Not to mention they would probably not have the same setup you have? Or does this...
I always bet $100. True, they won't know who to bet on until just before the gate.


Which is why I am looking for a partner. Note, when I say partner I don't mean "I give you picks for free and you bet with my money". Whoever I do this with will have to put up a stake equal to mine. I'd rather not say what that is in an open forum.

...mean they just give you their money to bet?
I didn't say that, you did. We would pool our money somehow and share the results. I have done this with other things like blackjack and daytrading.


Why not post the amount in an open forum? You posted your win $$. Don't be shy. You need to cull the deadwood out anyway....might as well be just as up front with the requirements where all can see. For a second I thought that you were just gonna give them your picks and they would bet on their own. Suddenly, to me, your partner ship is gonna cost someone a few hundred thou to get involved.

Why would I just give someone my data? I'm not the salvation army. I don't expect them to just give me their data either.


Correct me if I am wrong, but I tend to read between lines....more is said there than anywhere else. (you seem to think the jockeys are drugged and the trainers are all women as well. But that is just one of your inferences that really throw me off. Not exactly professional remarks by someone who is as knowledgable as you at first come off as.)

I was kidding around, having a little fun. Lighten up.


One other question.....why go back to paper after such a huge score on Monday? Also why change to perfecta betting? Suddenly the entire dynamics has changed, in less than 3 days no less!!
I didnt' change to perfecta, one of my friends on here wanted to get a feel for my picks with live action, he suggested the perfectas. Yesterday the volume was too low to be a good indicator. Also, I had worked 12 hours at my day job and I was too tired to use real money. No, the entire dynamics has not changed. My friend said that perfectas would increase the outcome. He was the one who saw that the patterns gave not just the winner but also the place and show.


Sounds like you have the golden goose by the throat. Like KM says, "The only true test is with your real money". If you have a winning angle or method or system or program, why bother digging deeper into anything....?
130 races is too small a sample to say anything with certainty. I would like to know why the patterns work from a handicappers perspective. Also, the day I bet with real money was a holiday. I have to validate that the holiday did not skew the results.

And, I disagree with the statement, "the only true test is with your real money". Patterns can only be recognized from practice, trading systems have to be constantly tested.

patternfinder
31st May 2007, 06:07.05 PM
This is the easiest part of all that you wrote. Anyone with HTR can do this and you could as well if you had the program. Are you already a subscriber?

Of course you don't have to subscribe to tell which trainers have the short skirts. I can honestly say I have never seen a trainer in the winners circle with a mini skirt on.

BUT a very good idea to attract the customers! :D

Yes, I could do if I had the time. I downloaded the program but have not subscribed yet. I guess I will just do the handicapping myself.

My girlfriend made the comment about the trainers and the short skirts, my girlfriend is a bit wild.

HaroldFA
31st May 2007, 07:27.05 PM
Hi Steve:

Not my intent to rough you up anymore regarding your Spreadsheet. However, in Post 25
you stated that: "I will say that I am using 12 tracks. Someone asked me why I picked those 12, the answer is that they are the ones that the betting sites I use will let me bet on. I also don't bet, what is it called? The races where the horses are pulling the wagon. I forget what it is called. Wagon racing? I don't bet on those. Don't ask me why. It isn't rational".

Well one of your 12 Tracks was the Meadowlands - MED. That's my home track and it is almost exclusively a Harness Track. There is a short rather meaningless Thoroughbred Meet in the fall. Right now it's running Harness Races as is almost always the case. So according to your Spreadsheet you actually did bet $1000 on 10 Harness Races - and apparently made a nice profit of $391. So - I guess that I'm doing you a small favor by pointing this out.

By the way, I also derisively refer to Harness Races as Cart Races - have absolutely no interest in them. As far as I'm concerned MED is just a simulcast facility, and the vast majority of their simulcast revenue is derived from Thoroughbred Racing.

Wish you all the best with your Data Mining Project. I think that to the extent that you've been successful the over-riding correlating factor is VALUE - rather than any particular set of Handicapping Factors. VALUE - that sounds nebulous, but it is something that we can try to quantify by calculating Fair Odds based on a composite Handicapping Score.

patternfinder
31st May 2007, 08:40.48 PM
Hi Steve:

Not my intent to rough you up anymore regarding your Spreadsheet. However, in Post 25
you stated that: "I will say that I am using 12 tracks. Someone asked me why I picked those 12, the answer is that they are the ones that the betting sites I use will let me bet on. I also don't bet, what is it called? The races where the horses are pulling the wagon. I forget what it is called. Wagon racing? I don't bet on those. Don't ask me why. It isn't rational".

Well one of your 12 Tracks was the Meadowlands - MED. That's my home track and it is almost exclusively a Harness Track. There is a short rather meaningless Thoroughbred Meet in the fall. Right now it's running Harness Races as is almost always the case. So according to your Spreadsheet you actually did bet $1000 on 10 Harness Races - and apparently made a nice profit of $391. So - I guess that I'm doing you a small favor by pointing this out.

By the way, I also derisively refer to Harness Races as Cart Races - have absolutely no interest in them. As far as I'm concerned MED is just a simulcast facility, and the vast majority of their simulcast revenue is derived from Thoroughbred Racing.

Wish you all the best with your Data Mining Project. I think that to the extent that you've been successful the over-riding correlating factor is VALUE - rather than any particular set of Handicapping Factors. VALUE - that sounds nebulous, but it is something that we can try to quantify by calculating Fair Odds based on a composite Handicapping Score.


Hello,
I did not know that Meadowlands is a wagon track. Given that I am just looking at raw numbers, I guess it doesn't really matter. Don't worry about beating me up. I can take it.

I did notice that over half of the winners I picked were at odds between 3 and 5. Today was the same. I'm not sure if that means anything or not. Is there something significant about horses at odds between 3 and 5?

MVM
31st May 2007, 09:09.25 PM
This is a really good price!

http://www.amazon.com/Stearns%C2%AE-Lightweight-Utility-Stocking-Waders/dp/B000HSUECE/ref=pd_bbs_1/105-1070233-2654842?ie=UTF8&s=apparel&qid=1180663534&sr=8-1

Donnie
31st May 2007, 09:23.48 PM
:D:DThanks Mike! But I've already got mine on!! :D Put 'em on a while back!

njcurveball
1st June 2007, 08:51.13 AM
My girlfriend made the comment about the trainers and the short skirts, my girlfriend is a bit wild.

NOW that sounds like you know good value to me :D

MikeDee
1st June 2007, 11:51.13 AM
Thanks Mike for the referral. Do they have anything for the smell?

Donnie
1st June 2007, 11:56.27 AM
:D LOL!! You guys crack me up!!

njcurveball
1st June 2007, 12:45.13 PM
I'm not using a handicapping program, I'm not looking for the kinds of leading indicators handicappers look for, I'm just looking at raw numbers in the betting patterns.

I would be curious what software you are using and what database if you care to share?

I have the Oracle Data mining tools and Oracle Database. In my limited trials using it, the profitable patterns were more based on back fitting than logic.

For instance, if the #3 rating did better than #2, the tool naturally said bet #3. But what usually happens in a case like that is the #2 factor will start winning and the #3 factor will start losing. Regress to the Mean is one of the favorite songs on this board.

From what you have written you have a database of betting patterns and you have back fitted the data to find successful patterns you hope will go forward.

Given the fact that 75% or more of the money is bet through Simulcasting, the fact that a natural 3-1 is going off 4-1 with 1 minute to post isn't a big deal these days.

Without a reliable barometer, it is hard to tell if a horse who should be 5-1 is bet down or an overlay.

If you are talking 3-1 to 5-1 range this is usually the 2nd and 3rd choices in a race. I think you said one of your "rules" is to toss the favorite. So it makes sense you land in this area.

If you have patterns that track not only the winner but 2nd and 3rd as well, then you better be making a million or more at your other job. Information like that could be worth $10,000 or more a day at the Windows with 12 tracks in play.

Imagine being involved in 120 races a day at $1,000 per race. That is $120,000. With only a 10% profit, you could make $12,000. You could hire anyone to do your "dirty" work then. Give you more time with your hot gf.

Heck, I would not only give you database expertise for $1,000 a day, I would even bring you coffee! :D

MVM
1st June 2007, 02:06.46 PM
Thanks Mike for the referral. Do they have anything for the smell?

Sorry Mike, I think this one is going to necessitate a call to Servicemaster.

Pattern Detection Data Mining Tool (http://kenstange.com/ken/BullshitDetector.htm)

I scored this one a 13.

Victor
1st June 2007, 06:19.42 PM
store this in the "what's in it for HIM" category.

Nobody needs a partner unless they do, which means what?

What "handicappers" know ain't for sale because nobody will pay what it's worth. It ain't for sale. Our money ain't for sale and our knowledge ain't for sale and we wish you nothing but continued success.

Victor
1st June 2007, 06:59.29 PM
What is daytrading?

The answer, at absolutely no cost to you, is gambling.

Spin the roulette wheel and track it. The wheel is fixed and there is no way to win usually or in the "long run" look it up expectation math and such.

Patterns don't predict the outcome of future events. Daytrading might be a good idea in theory so why give that all up for horseracing! Ask any Dinosaur about the K-T catastrophe and if they saw THAT coming!

I think horseracing could be a good idea for those who approach it correctly, just as the Markets are too, and the rest will have to prove me wrong about any mechanical approach that needs a partner that will succeed and provide the bankroller with a good deal. You're on the right track though, to think maybe "handicapping" could be the answer . . . :rolleyes:

Every race is different so why not approach each one differently? Makes sense to me but what do I know? I'm pretty much dumbfounded by the whole thing and that's saying something and I'm not the first either.

It's hopeless but not entirely hopeless . . . since computers and models and patterns can't do that.

Good luck is not a pattern thing either.

Good luck actually operates before the pattern.

patternfinder
1st June 2007, 07:29.43 PM
I would be curious what software you are using and what database if you care to share?

I have the Oracle Data mining tools and Oracle Database. In my limited trials using it, the profitable patterns were more based on back fitting than logic.

For instance, if the #3 rating did better than #2, the tool naturally said bet #3. But what usually happens in a case like that is the #2 factor will start winning and the #3 factor will start losing. Regress to the Mean is one of the favorite songs on this board.

From what you have written you have a database of betting patterns and you have back fitted the data to find successful patterns you hope will go forward.

Given the fact that 75% or more of the money is bet through Simulcasting, the fact that a natural 3-1 is going off 4-1 with 1 minute to post isn't a big deal these days.

Without a reliable barometer, it is hard to tell if a horse who should be 5-1 is bet down or an overlay.

If you are talking 3-1 to 5-1 range this is usually the 2nd and 3rd choices in a race. I think you said one of your "rules" is to toss the favorite. So it makes sense you land in this area.

If you have patterns that track not only the winner but 2nd and 3rd as well, then you better be making a million or more at your other job. Information like that could be worth $10,000 or more a day at the Windows with 12 tracks in play.

Imagine being involved in 120 races a day at $1,000 per race. That is $120,000. With only a 10% profit, you could make $12,000. You could hire anyone to do your "dirty" work then. Give you more time with your hot gf.

Heck, I would not only give you database expertise for $1,000 a day, I would even bring you coffee! :D



No, I do not have a database of betting patterns. Backfitting is a waste of time. I look at the data and see trends. I won't know what the patterns are ahead of time, so there is no way to "have a database of betting patterns". I use oracle's database but I don't use their data mining tools, I wrote my own. I haven't found any patterns that are reliable for the #2 and #3 horse. I found in trying to pick 1-2-3 I get too distracted from the main objective which is to pick the winner. I don't look at the odds, I only look at the momentum of the bets, momentum is simply volume over time, whether the odds are "better" than the tote board or "better" than 'natural', I wouldn't know.

The way I look for patterns is probably different than what a lot of people think that means. I don't know what the data means, I just look for irregularities.

I wouldn't bet $1,000 a race, the risk of ruin is too high, in order to bet $1,000 a race, you would need a $1,000,000 stake. I use the Kelly criterion to size bets and the highest risk I am willing to take on any given bet is 1%. IF you are talking about a positive expectation game. Right now I am at $100 per race, I may increase that later. I am comfortable with this risk level, and I fully expect to have lengthy drawdowns. That happens in daytrading and blackjack all the time.

patternfinder
1st June 2007, 07:38.56 PM
What is daytrading?

The answer, at absolutely no cost to you, is gambling.

Spin the roulette wheel and track it. The wheel is fixed and there is no way to win usually or in the "long run" look it up expectation math and such.

Patterns don't predict the outcome of future events. Daytrading might be a good idea in theory so why give that all up for horseracing! Ask any Dinosaur about the K-T catastrophe and if they saw THAT coming!

I think horseracing could be a good idea for those who approach it correctly, just as the Markets are too, and the rest will have to prove me wrong about any mechanical approach that needs a partner that will succeed and provide the bankroller with a good deal. You're on the right track though, to think maybe "handicapping" could be the answer . . . :rolleyes:

Every race is different so why not approach each one differently? Makes sense to me but what do I know? I'm pretty much dumbfounded by the whole thing and that's saying something and I'm not the first either.

It's hopeless but not entirely hopeless . . . since computers and models and patterns can't do that.

Good luck is not a pattern thing either.

Good luck actually operates before the pattern.


Daytrading can be gambling. Depends on the day trader and what they are trading. Also on why they are trading. The way I daytrade is I find volatile stocks (volatile intraday) and then track their trend over ten minute intervals and bet the trend. Sometimes this works, sometimes it doesn't. It mostly works.

Yes, roulette is a negative expectation game, I don't play roulette.

You are right, patterns cannot predict the future, they can only tell you what has been happening lately. Anyone who uses patterns for trading or betting is hoping that what has been happening lately will continue. At some point, of course the pattern will stop working, and then you have to find new patterns.

Never said I gave up daytrading.

Yes, I agree that horse racing is a good idea when approached correctly. There are a lot of people smarter than me who think that horse racing is becoming an efficient market and is showing signs of behaving similar to a stock market or the currency markets, two other efficient markets. Which is why I am interested in this.

All I know is that I found a pattern that so far has been repeatable over 300 races. Will that continue? Who knows. But, that is why I would like to work with someone who knows racing to find out if there are patterns of handicapping variables in my picks. I know the picks work, I would like to know if there is a horse racing reason why they work, as well as the pattern of the bet momentum.

I tend to believe that good luck goes to the prepared.

njcurveball
1st June 2007, 07:42.14 PM
I use oracle's database but I don't use their data mining tools, I wrote my own. I haven't found any patterns that are reliable for the #2 and #3 horse.

#2 ranked or #3 ranked was just an example. I probably should have put some more detail behind it.

Are you writing the front end in java? vb? I said back fitting since you said you have a years worth of data. Have you been testing this for a year?

As for the $100 = 1% bet. That is a comfy $10,000 bankroll. Shame you are in the US or you could probably make a solid rebate check.

Jim

patternfinder
1st June 2007, 07:46.57 PM
store this in the "what's in it for HIM" category.

Nobody needs a partner unless they do, which means what?

What "handicappers" know ain't for sale because nobody will pay what it's worth. It ain't for sale. Our money ain't for sale and our knowledge ain't for sale and we wish you nothing but continued success.

I don't 'need' a partner. I just know how to evaluate risk and I think that the risk of using just bet momentum by itself may not be acceptable over the long run, so I would like to have my picks evaluated using handicapping factors. I think there is a pattern there. I think it will be easier to estimate risk using that approach. I'm not really sure what your point is.

patternfinder
1st June 2007, 08:03.28 PM
If anyone wants to talk to me seriously about combining patterns with handicapping you can email me at ibeecurious at yahoo dot com. I'm willing to give you the results of my picks everyday, and I am willing to give you picks live, a few seconds before the race for awhile so that you can see that what I am talking about is real. We can use yahoo chat or something similar. After awhile you will see that I have something real to offer.

I won't be coming back to this forum. There are too many people here who just want to talk trash. I am looking for people who want to talk seriously.

Patternfinder

Victor
2nd June 2007, 05:26.20 PM
ok Pattern, at least you have apparently given me some due respect as to the nature of my objections, and that's a good thing.

Next time anybody wants to shut down a thread of mine should better think again.

We gots math and we gots probability and we gots whats left of human judgement to kind of determine what could happen so wheres do we go from here?

I say nowhere in particular because the jokers clowns without the ability to make those kinds of determinations are still here and it's possible to take advantange of them.

The right tools for sale could help but never replace the basics. They are for sale but only to the right people I imagine who could make use of them.

Partners are always a question of who gives what to whom and to what end? In this case it's horseracing profits and they only come from superior insight over the long term and nothing less.

I didn't mean to "run you off" pattern (so please come back :)) only meant to set you straight. There are no patterns that saw it coming or will see it coming or will ever see the future coming.

Got that? and remember before thinking I'm some kind of a wack job, my background, is more than considerable in this field and a few others.

Plus, I've read the gambler by Dostoyevsky.