Steve Wolson V6.0

[John]

I just received my information pack from Power-online.detailing "The software" V6.0. 6 pages included the new screen and new features.

one screen shows the new longshot Consensus overlay
feature in V6.0.This is a new overlay indicator built into the program that is triggered by the program noting the discrepancy between ML odds and predicted odds. Example:at Belmont 99/22 "Champagne Ending" predicted odds of 2-1 and morning line odds of 8-1.
Horse went off at $46.50-1 of course winning easy and paying $95.00. End of example.

Also, there is a new Results feature. A result window that pops up on any race that has been modeled so you can see the actual results of that race.

also. new modeling options,And there is a Contenders and non-contenders that is important for exotics and Maiden races where first time starters may be a factor.

Several more new features in,"Detail Query Tools" and Black box,Data Bulk Modeling,Plus the ability to manually update Turf surface changes even after modeling.

Breeders Cup Special that will surface on Breeders cup day each year.highlighting the selection in each of the breeders cup races.
[ these selections are generated by models that have been back fitted to the point were where they have been pretty good in pointing out the winner in the Breeders Cup......... .[ discontinued "restocking fee" Full refund in 30 days.]

Will all of the above show a profit. I don't know.

Soon I will be ordering the upgrade and intend to post some selections for all to see.

Maybe we can get Steve Wolson to post some of his selections while explaining if the new features actually picked the winner.

Any Questions to the above, Per his Information Pack .Please contact Steve Wolson wolsons@aol.com

JOHN

[Maxspa]

All,
One of the few software programs (5.0) I've tested that showed a PROFIT. One of the draw backs, I found was, it depended on some huge prices to achieve the positive results and I wasn't comfortable with that aspect of the program. IMHO those priced horses might be aberrant situations and never happen again.
I'll be interested to see John's results with Poweronline's 6.0 program. I checked the web site and found several upgrades!
None of us on this board can forget Steve Wolson's winning postings with a 500 race test.
So John, I'm looking forward to your unique incite and honest evaluation of this program.
Maxspa

[BIG RED]

I would also be interested John. I may start this winter ( for the first time ) to use a program, and I have had this one in my head for sometime to maybe try, for a newbie. I do recall how well Wolson did.

[Doug3312]

Does Wolson have a web site?

[cj]

www.poweronlinesoftware.com (www.poweronlinesoftware.com)

[speedking]

Any updates on the new version yet? Thanks.

speedking

[garyoz]

Being a software junky, I purchased it. I am not ready to provide a full review because I'm a new user. First let me say excellent and friendly support from the programmer Mel. I have no knowledge of the previous versions. Why I decided to keep the software after the trial period (I decided to keep after 2 days of use, not 30 days) is that it has a unique approach to computer handicapping. It is not paceline based, but rather "system" or decision-rule based. There are 23 (by my count) different methods that could potentially be used (some applicable to only Turf, maidens, etc). To me this was a breath of fresh-air. Pretty much, if you select the same paceline, no matter the program, you'll come up with similar picks. Contenders are selected based upon a consensus type point model. (5-top selection, 3-second, 2-third, etc--can be user defined or defaulted) based upon methods appropriate for the race.

The database capabilities for looking at each method is pretty extensive. The black box and eliminator functions are useful. Two caveats, first, you need a pretty large database and two, something I have posted alot about, you are looking at past correlations--and correlations reqress to the mean unless you have a large sample. Takes TSN and Bris files.

Personally, I don't think I would use the software on a stand-alone basis, but it definitely provides an efficient assist to my handicapping. I use in conjunction with figures (either CJ's, or Xtra's or occasionally Tgraph) and traditional DRF handicapping (and of course sometimes other software--I am a software junky afterall). You can download the manual at the web site.

I can provide more information as I get more experience with the software. The programming and usability of the software is very good (IMHO).

[Maxspa]

All,
Very surprised that Gary O, the traditional handicapper, would be involved with PowerLine. That being said, All of us software fanciers will be fortunate indeed to have Gary's opinion regarding this product. He is experienced, honest and has a storehouse of handicapping knowledge. His analyzation skills are unique and well crafted.
I hope we are lucky enough to get his evaluation when he has sufficient time to test the software .
Maxspa

[speedking]

Thanks for filling us in. Much appreciated.

speedking

[garyoz]

http://www.paceadvantage.com/forum/showthread.php?t=19736&page=4

There is a recent post in another thread relevant to Version 6.0. The poster seems to have a lot of experience with the software. Post number 53.

[BIG RED]

http://www.paceadvantage.com/forum/showthread.php?t=19736&page=4

There is a recent post in another thread relevant to Version 6.0. The poster seems to have a lot of experience with the software. Post number 53.

Have you an opinion yet, gary?

[garyoz]

Thanks for asking. I downloaded 86 cards and results for Tampa Bay from the Bris Archives last weekend for last year. I thought that would be a starting point for a database for a meet that I like to play. I'm trying to determine a fair methodology for measuring performance and displaying results. I will try to do it for the Tampa Meet prospectively and also provide summary data for last years meet retrospectively.

In just eyeballing the Tampa results there are plenty of positive models and apparently profitable relationships. The question is whether they will hold up over time? Are they real causal relationships or merely correlational artifacts that will sink to "average" results over time when the sample size is increased (aka "regress to the mean")?

One thing that I like about the program is that is is very unique. It is not like the standard quasi-Sartin paceline selection-based software. With most software programs, if you select that same paceline you will pretty much come up with the same picks, more or less.

A key question is whether you are interested in somewhat of a black box approach, for identifying contenders, establishing oddslines, finding horses to eliminate, etc. As I posted before, personally, I'm too much of a traditional handicapper to go over to a black box approach, but the selections, oddsline, etc. can be used to inform my traditional approach and help point me to overlays and toss some underlays.

I don't mean to present myself as a big time authority on the program. I'm just learning it. It has many really nice functions. I hope to have the time to provide some summary statistics in the next few weeks. I haven't had much time during the last several weeks, but hopefully that will change soon.

[BIG RED]

Thanks, we are all time limited with the holidays fast approaching. GL

[garyoz]

I checked this past weekend’s performance at Tampa, based upon 86 racing cards that I downloaded to start building models. I was traveling, so didn’t even get to play, so strictly looking at it retrospectively. Most important, it is for only two racing cards.

There are so many ways to use Power-Online, that it is difficult to decide what to look at (this is both good and bad). I decided to focus on the Consensus Picks, the Black Box picks and Longshot Consensus Plays.

The core of the program consists 24 different methods or “systems” that can be applied to rank horses in a given race. Some of these models are race type specific for maidens, or turf races, etc. Typically they consist of race eliminations, horse eliminations then some decision rules. The consensus picks default to a score that is similar to the DRF box, where first pick is worth 5 points, second pick is worth 3, etc. (This can be modified to a user's preferences). The consensus scores are then summed for each horse across all the methods. Each method is given an equal rating. The Top consensus pick had 3 winners each on Saturday and Sunday ( I eliminated a deadheat race from consideration and one maiden race from the 11 race cards) out of 10 races. The prices were short, but Tampa didn’t have a lot of big prices this weekend. A base $2 bet totaling $40 would have generated $38.80. Nothing to get excited about, but you would expect this type of approach to generate chalk. 3 of the winners had a Power Online oddsline lower than the off-odds. So just playing overlays would have shown a positive ROI.

The Blackbox picks sort through the database and identifies the system or method that provided the top ROI and top winning percentage in past races matching the conditions of today's races and lists the horse from that category. As many as 4 horses can be listed, but in most races 2 or 3 are listed. So, the BB provides multiple selections. Only 7 of 20 races had winners listed in one of the Blackbox models. Disappointing, but a small sample. I’ll try to look at a larger sample size.

Finally the best for last, the consensus long shot play picked the $46 winner in the 9th on Saturday at Tampa. There were a total of 7 longshot plays on Saturday. The winner was the only consensus long shot to hit the board. On Sunday, there were 9 consensus longshots, with 2 thirds and one second. The $46 winner on Saturday was the key to a $10k Pick-4. The consensus long shots on Sunday that finished in the money rounded out some nice tri’s and exactas. Seems consistent with what my friend Maxspa said, that if you are patient, you can get some big prices.

Obviously need to do more work. This was only a two day sample. You can use the software to model thousands of different combinations (some esoteric such as contenders in the 3rd spot in trifectas with non-contenders on top, etc.). Or you can see what models show a positive ROI in particular types of races. As stated before, the problem is will the relationship that was profitable in the past hold up over time? As Harvey Pack used to say, “If you missed the wedding you don’t want to go the funeral”

I'll update as I analyze more races. Once again, I don't profess to be an expert with this software and anyone with more knowledge, feel free to jump in. It is an interesting program--I still think best used as a handicapping aide.

[Maxspa]

Gary,
Look forward to your updates! They are informative and as always contain quality handicapping views! Your reports always look at both sides of a situation and you tell it like it is. The readers are fortunate to have an experienced person evaluate this program!
Maxspa

[John]

Gary and Max,

I recently upgraded.I also have found many ways to use POLSW.I started with Aqueduct inner track because I had a large data base to model.I did have my share of winners 45 plays 27% winners $90.00 bet $89.00 returned.just about broke even.

Interestingly,If I bet the odds/on consensus plays[ the picks in yellow and if you demand a minimum 10% jockey AND a 10% trainer, the results change as follows:

23 plays / 11 wins 47.6% wins

$46 bet / $84.00 returned - a big 80+% profit.

I have started with day one at Tampa.Larger Fields and better prices.I have no data base like you Gary. So I will have to build a model day by day.Once my data base has sufficient races and results. I will be able to run Consensus from the options and use money index to generate whats winning at Tampa.

For now I am useing POLSW grid screen. Gary I want you to go back and look at the past races. Look at the color codes.Yellow is top pick, Blue is 2ND, Green is 3rd, and Pink is 4Th. The Longshot plays are indicated when the entire horse's name is in color, rather than just the Consensus rating.

I am amazed at how many winners are among the first three colors. In almost every race the winner is yellow,blue or green.
Example: Gary.look at the 6Th race Sunday your green won $22.00 and your blue,is 2ND. exacta $99.60. A shame that there is no late pick 4 or pick 3 all Tampa does offer is Late DD.With the late scratch in the 10Th race #10 and off the turf.POLSW made #1 blue,2ND choice and in the last race POLSW made #4 yellow, 1st choice. DD $42.40. Not to bad.

Yes, Max is right. If you can wait until the odds are in your favor. you can make money.If the First 3 POLSW choices Yellow.Blue and Green are the same tote board favorite. I pass the race. The way I bet, $12.00 exacta is meaningless.

Gary I look forward to hear more from you on POLSW and I hope the above helps.

[midnight]

All this is well and good, and eventually you can find something that wins over a PREVIOUS sample. The trick is to find something that works forwardly.

If I remember correctly, the HTR rating (in the HTR program) showed an ROI of 1.03 in 2001 if used with post positions 1 to 5. In every other year, the ROI was between 0.88 and 0.92 for the same situation.

The point is that if something can be found to show a profit over about 50,000 races (of which probably 28,000 were played since the top HTR rating would be outside of the first five posts about 44% of the time), then one can see that caution has to be exercised when using too many contingent factors to come up with a profitable "black box" method.

[garyoz]

For now I am useing POLSW grid screen. Gary I want you to go back and look at the past races. Look at the color codes.Yellow is top pick, Blue is 2ND, Green is 3rd, and Pink is 4Th. The Longshot plays are indicated when the entire horse's name is in color, rather than just the Consensus rating.

I am amazed at how many winners are among the first three colors. In almost every race the winner is yellow,blue or green.


John, thanks for your comments and advice. The reason that I focused on the three picks (Top Consensus, Black Box, Consensus Longshot) is that you need a precise definition in order to test performance.

In terms of the top-4 consensus, I believe that they didn't win several races on both Saturday and Sunday--I tracked that but I don't have the data with me. Picking winners out of the Top-4 shouldn't be that difficult. I believe that it is the Synergism Program that assumes the Top-5 selections will win 92% of the races. This seems pretty reasonable to me. I'll measure based on the Top-3 picks as you recommend. I'd expect the win rate to be around 70%.

As you suggest, the program appears good at selecting contenders. However, what contender do you go with? That is why I think the program is very useful in conjunction with traditional handicapping approaches.

In terms of Blackboxes and using backfitted models to forecast the future, I'm not sure that it will ever work in this game. They are fun to play around with.

[midnight]

I believe that it is the Synergism Program that assumes the Top-5 selections will win 92% of the races. This seems pretty reasonable to me. I'll measure based on the Top-3 picks as you recommend. I'd expect the win rate to be around 70%.


Synergism picks 92% of the winners in its top 5? That seems high to me.

[garyoz]

As I recall, Synergism allocates 92% of the probability of winning to its top-5 picks when creating it oddsline. Personally I think that program has the best oddsline of any I have used. I've also heard great things about the HSH oddsline.

Don't forget that any use of the favorite in a contender list will get you to about 33% off the bat. That would leave 4 horses to make up the other 59%. Doesn't seem unreasonable to me. Have to control for field size, so the statistic may be meaningless on its own.

[John]

I am sorry, Some members PM me and I deleted all my new messages.If you want, PM me again or e-mail me at jrmartino@yahoo.com......Thanks

[John]

HI,
Wanted you to know that so far at Tampa Downs POLSW is doing good.

Big fields and a lot of POLSW longshot picks.POLSW has picked a longshot almost every day so far. Right now Tampa is running only two days a week. The first of the year wiil start a normal schedule.

Aqueduct has return to racing today.Consensus/odds-on and longshot plays that have a bug rider are doing well.

John
HAPPY NEW YEAR.