Sniper Plays / Contest

[km]

Received an email from NJCurveball concerned about the direction of this contest. Much appreciated Jim. The gist of which is that the contest might turn into a race of who can contrive the most perfect backfit (including himself). He coined the phrase "sniper" plays for things such as =

HTR <> 3 (but all others are ok)
S/P = 3, 7, 8 (others are eliminated)
JKY = 169-203 OR JKY 338-396

Look at that last item. Any sample of longshot winners will have some commonality and that's how "sniper" items can be developed - pure backfitting - just tailor it to the highest priced winners to get the ROI up.

My fault for using 600 plays as the minimum, should have made it 1000 or more to avoid this sort of thing. Posh expressed his dismay with this, i agreed with him, but we voted on it with a Poll here, and the majority wanted to keep the contest rules as is and award the $500 to anything that can be verified, no matter how quirky or contrived.

A spot play that includes "sniper" items is no good to any of us. It might win the contest, but it is not likely to go forward and is not useable to the other HTR subscibers.

Luckily many of the plays submitted earlier in the contest are terrific though and are easy to setup and continue to win going forward. So even if we get a loco one to win it, the contest was probably valuable overall for the ideas presented.

So we will carry on until next Friday; anything goes.

I guess i could formulate some ridiculous nonsense to get a 2.50 ROI and donate the $500 to Rick. Rules are rules - there is nothing that says "KM is ineligible" :eek:

Your comments appreciated

[njcurveball]

Hey! Why is there a HUGE target on my back now? :eek:

My concern was that you can do wonderous back fitting with a database. If factors such as RPED = 1 are used, the play has a solid chance of being profitable in the future. If I backfit a play to say RPED > 2 but not = 5 and < 8. The chances of it being profitable in the future are not very good.

Here is my simple suggestion, which of course does NOT apply to this contest, but perhaps for future contests. I am NOT asking for any changes. I am going to play by the "anything goes" rules.

My simple suggestion:

Add another "rule" saying a play has to be also show +10% ROI in 9 of 12 months.

Alright now that I am about as popular as Bernie Williams to the Yankees I am going to go find some salon and shave my head! :cool:


Jim

[MVM]

I would resubmit the idea we discussed earlier and determine the winner by plays X return.

There would also be a number of other measures we could use to determine which plays are realistically the most viable and likely to go forward with a positive return.

We could have an "HTR Oscars" type of award array....anyone who cares to submits their plays, and we have awards in various categories...best in '03, '04, '05...best over last 4 years total....total potential profit based on the pool size being played....it is also fairly simple to determine the possible actual range of returns over an infinite number of plays by using Proportion and Confidence functions....i.e. we can say with 95% probability that SpotPlay763523829046 will return between ? and ?.

If nothing else I think the various measures would provide people with a solid idea of exactly what constitutes a solid spot play. I think many people would be surprised at how much you can learn about and how little you actually know about a play when looking solely at ROI, # of plays, Win rate, average mutuel and IV.

BTW KM, 4 of the 5 highest winning mutuels in the country last year can be "gotten" using robot criteria that they have in common (the most obvious being HTR=9), there's a starting point!

[DanG]

Completely understand the concern.

As I wrote you at the start Ken the way the contest was structured it allowed tweaking. However, I for instance have a very successful play that I use with real money that I guarantee many members would term a “sniper” play. It jumps around with a particular rating and creates completely artificial values that are hardly easily reproduced.

A couple quick things:

I have never understood the rational that saying something is ranked 1st is considered “credible” and eliminating something ranked 2nd is “fitting” data. Many times it is, but it’s not always the case.

If I say nFC >84, but not “#X”, is that “sniping”? :confused: If it is then I snipe every day, because the way I use the rating there is a dead spot above its nitro zone and I won’t bet it.

To me, I love this contest as is and as I wrote the only future modifications I would suggest is split divisions (Robot, Access) and possibly limiting criteria within Access.

[njcurveball]

Dan,

I value your opinion greatly. There is value in #2 rankings and #3 rankings, etc. No doubt it. A sniper play is one that simply takes out a rank for no logical reason.

For example, I find my Maiden Claiming play does very will for NVI < 23, however there is a "dead spot" between 23 and 25. Then it does well again. Is there any logic that 23.01 should be bad and 22.99 and 25.01 be good?

Perhaps I have a bad case of "linear thinking"? It would be like you telling me the horse is a go in NFC > 84, but do not play if it is between 85.1 and 85.5.

Perhaps a more germaine example would be if a few FTS bombs won and their FTI was > 50. However when I did a query I find NFTI = 59 was a loser, so I say NFI > 50, but not equal to 59.

This all makes me think back to the movie where they say there is a fine line between winning and losing.

I really enjoy your posts and learn a lot from them.

thanks for sharing,
Jim

[AwolAtHTR]

Wed 21 Feb 7

I believe a spot play needs to be valid for RECENT bets and the history for a year will give a person confidence to bet the spot play. Yes, the currently submited plays have APPARENTLY been back fit for best results. However, here is a spot which I found using K1 K2 K3 as the core condition, IE, chalk NOT long shots"

BET 00054 00017 31% 54% 59% 2.13 1.45 1.06 $13.5 0001 $36 2.52
ALL 00068 00022 32% 54% 62% 2.13 1.44 1.06 $13.2 0002 $36 2.51

The BET line used only the thirteen tracks which I can bet using AmeriTab. Whereas the ALL used all tracks from HDW. The data used was for 1 Jan thru 18 Feb 2007, yes, ONLY six weeks, and my money will be on this one until it stops working. What I found utterly amazing was that the ROIs with fourteen (14) more samples by the ALL set were so close to the BET sample set!!

So, IF the contest said '...best performance..' from '..cut off day..' until the '..end day..'
In other words, the cut off day would be the last day to submit and the END day would be say a week or month into the future. Sure, Ken (or others) would have to run the Spots for each day or week whenever a Contest update would be reported.

I did not expect this Spot play to hold up for a year because it has a bias to Three year olds in Jan and Feb. I expect other spot plays would be valid for ONLY a couple of months EACH year!

Another Spot play contest rule could be that ONLY the administrator will see the rules. The players will only know the performance (ROI for Win, Place AND Show?) that is the target for the next spot play to be submitted. Then, a player could win the contest and not have the play possibly diluted WHEN, not IF, other HTR players decide to use the spot play.

Another Spot play contest rule could be that ALL Win, Place, AND Show must show a profit to qualify! In other words, one or two long shots for the Win will NOT be a winner!!

This has been a good contest with the rules as defined and GREAT seed of ideas for future contests.

so, basically my opinion is: Contest with Spot play CUT OFF day, Winner by END day.

duane

[DanG]

Jim,

Trust me, the respect is mutual.

We do disagree on if “Illogical” sequencing of ratings can go forward. My second best play would be considered highly illogical and it’s actually contains a ("weird") foundation rating that I lean on.

If I use nWO, nFC, nPED, etc greater than “X” I have to use every rating above it? I have proven to myself this is not the case, but I admit I am adding my own twist to the recipe.

Good healthy debate.

Spot plays are very alluring to horseplayers. They take emotion out of the equation and approach that elusive “black box” that eludes so many. So much will be learned from this exercise that we all will benefit from.

I guess I chose to write because I submitted a play that had various criteria (one being HTR <>5). Next I read the word “sniping and HTR <> 3 is given as an example. I use a play that works well that excludes several middle range ranks while allowing the top and bottom end. To some this is “Illogical” and can’t possibly go forward, to me it works within my framework.

BTW: Since I wrote that play on 2/11/2007 it’s had 21 plays at 19% and an ROI of $2.34.

Jim, I really wish I could show you one method in particular I use that has really been profitable, but as you know you can’t show your entire hand in a pari-mutuel game. I think we would both get a kick out of it and I’ll bet you might think to yourself…”This shouldn’t work” But, thank goodness, it does,

PS: I know spring is coming because you changed your avatar! Baseball and the Triple Crown are close and A-Rod and Jeter are back page stories again…All is right in the world! :D

[njcurveball]

Dan,

You and I might sit down and enjoy a good dinner where I tell you Ted Williams was better than Babe Ruth and you tell me I am stupid to say that.

You prove your logic at the Windows, so there is no need to listen to anyone else. What did RIchard Dreyfus say about talking about the races. This isn't the Olympics! LOL!

Heck, I have had conversations with people who say Energy distribution is hogwash and they are professionals. My comments were directed to Ken as suggestions for future consideration.

They were not directed at anyone or any play. I have a tool called Oracle Data Miner. It can come up with some weird things. I don't think anyone would like it if I found a play that was profitable between 1455 and 1506 Eastern time. Although that play may actually be my bread and butter.

My comments were more for the support of the non database players than anything else. Please do not take any of them personally as they were not meant to be.

OK, I have to run back to my bed in rehab now! Good thing I shaved my head so they can't test me. :D

Jim

[DanG]

Dan,

You and I might sit down and enjoy a good dinner where I tell you Ted Williams was better than Babe Ruth and you tell me I am stupid to say that.

Jim, but could the “splendid splinter” pitch? :D

Nothing was taken personally at all, I just wanted to point out that there are several ways to skin a cat (horse) and some of them might not even make “sense”. ;)

[MikeDee]

Maybe you want to change things and then maybe you don't.

I have been following the spot plays that have a 1.19 or better ROI (only PL5) very closely and as a group they have done very well going forward.

(I'm not proud I'll take a 1.20 annual roi any day)

This includes plays that fall short of the min.

DanG mentioned it earlier, if you think of them as a group of stock picks, a diversified portfolio, on some days one is up and the other down and over time you net out with a positive ROI.

If you play them as a group, so what if some have a very few plays. On a group basis you have lots of plays each day and you can afford to wait for the few plays that the short spot plays find.

We can quote conventional wisdom that a single spot play built on a few super bombers will not play forward, but on the other hand a dozen or so spot plays all coming from different angles, that had super bombers in their past, might just catch a few super bombers going forward and all you need to do is to be on board when the ship comes in.

I don't like the idea of keeping the plays secret. It really cuts down the interest and gives Ken zero marketing value for running the contest in the first place. If your play is so good and so special that you can't share it with the HTR group, that's fine, keep it to yourself, stay out of the contest.

So I say "lets' not over analyze this". Let the contest finish and then continue the monitor the best of the spot plays and lets see were they are in 3, 6, or 9 months.

Lets see what would happen to hypothetical bank roll based on daily play of a group of spot plays over a period of time into the future. Lets test that conventional wisdom.

Maybe this can become an annual event held in Feb. Sorta of the "American Idol" of horse racing:D

[dehere]

My money is on Lakisha Jones.

And I'm with MikeDee on this - no need to over-analyze this thing. While the "winning" play may be heavily oriented towards back-fitting, there are obviously some fine plays coming out of this contest. It'll be interesting to see which of the various plays perform the best over the next three, six or nine months. Maybe the next contest should include some sort of turf-vivor format. The first round, based on history, provides the contest with the initial qualifiers and the second round, based on what happens going forward, actually determines the winner?

[ERNIE LOGSDON]

Mike: I've been trying to promote sharing for a long time now,after the contest I plan to post some plays that fell short for a contest play but still do well enough for me to watch.
and to make my point here is my "Old standby"
as stated by others "don't dismiss or( knowingly smile to big) as to simple until you watch it for a few days"
monday 2-19-07 op r9 found $48 winner..........

spot play is: QP= 6+ and $, $$ and PED =450+ ..........<Ernie>

[DanG]

I don't like the idea of keeping the plays secret. It really cuts down the interest and gives Ken zero marketing value for running the contest in the first place. If your play is so good and so special that you can't share it with the HTR group, that's fine, keep it to yourself, stay out of the contest.
D
I follow this up to a point.

How does one share a play that requires hundreds of complex queries to construct custom value tables? Followed by statistical analysis through Excel.

By not “sharing” I was referring to plays that are not easily duplicated within Access or the Robot. As MVM said, he has plays that are far off the beaten path as I’m sure many of us have.

[MikeDee]

Dan and anyone else,

Please don't construe my comment about sharing to anything you have posted or not posted. Of course, it wouldn't make sense to post something that is not obtainable via the robot or via access or anything that does not come close to the criteria like single track spot plays.

No one should be expected to post a play that they don't want published.

I was referring to the idea of keeping all the spot plays secret and only posting results as a way to encourage more spot play postings. I just don't think future contests should be decided on secret spot plays.

[DanG]

Gotcha…Understood Mike and I agree.

I’ve learned a great deal just following the logic of the way people are constructing these plays. It is certainly revealing where the value is in racing and where there just isn’t many juicy prices left.

40% of all MSW 3yo pay $7.00 and under. :eek:

[njcurveball]

My money is on Lakisha Jones.



Off topic, and probably unrecognizable for those who do not watch American Idol. I do not watch a whole lot of TV, but I am loyal to a few shows, this is one of them.

To keep it brief, this is supposed to be a Grade 1 race and there are very few stakes horses entered. Lakisha Jones is a Grade 3 at best.

Tonight we say goodbye to the claimers, Sundance and Alaina.

We will see how much they "polish" Lakisha in the coming weeks. I would love to bet against her with Blake Lewis (the beat-boxer) and Chris Richardson.

OK, we now return you to your local horse racing forum in progress. ;-)

Jim

[njcurveball]

Jim, but could the “splendid splinter” pitch? :D




Ummm, could Babe Ruth fly a jet? Williams is the only Hall of Famer I know of that served in two wars, WWII and Korea.

Any web search lets the record speak. I would love to see Babe Ruth playing today. How many intentional walks would he have where he swung at two pitches. ;-)

Power at the corners? When Williams hit 43 home runs, his 1b (Billy Goodman) hit 0, his 3b (Johnny Pesky) hit 2. WHY did they even pitch to him?

Ruth had some guy named Gehrig behind him, that helped a little. ;-)

Seriously now, who would you say is the greatest baseball player of all time?


Jim

[DanG]

Jim I’m sure you’ll agree it’s impossible to compare eras.

As far as context, 59 bombs in 1920 when the leading team in MLB had 64 :eek: shows equal offensive domination as Ted did in his era.

However, extra credit goes to Ruth who could eat 7 hot dogs between double headers and still hit two bombs in the night cap. :D

This is comparing Secretariat to Man ‘O War. We would both love to pitch with them on our side.

[njcurveball]

hey Jim---
can you explain your rational behind the MLO <>10...? Why exclude the 10-1 mlo's only??


Donnie,

I pulled your question over here to offer something to both you and the other readers. First let me start by saying this is just pure back fitting on my part. I would not make this a spot play as I have been burned too many times before elminating one value in the middle of others. And speaking of others, this may work for many others, just not successful for me.

A simple case was the $$ horses on the turf that I so smartly put a condition to leave out the ML 30-1 horses. Of course the 2nd day after I set the report up a 30-1 ML bomb won! And of course my biggest smack in the face was Giacomo who had every condition for a spot play, except his ML was something like 50-1.

Ok, now perhaps something helpful. I don't know if this is true today, but the morning line makers of old would make a horse they could not understand 10-1. I made the ML at ACRC for a few years and this was the accepted "rule".

So if a first time starter came to town trained by Donnie N with no workouts. If I could find nothing good or bad, the line went down as 10-1.

One negative then 12-1, a few negatives 15-1, a lot 20-1, and so many I couldn't cross the horse out fast enough 30-1. Although I was warned not to make a horse 30-1, since that would OFFEND the trainer. Truth!

If the horse had a positive, then it went 8-1, and the more positives the lower. Positives could be a hot trainer, hot jock, inside post, outside post, depending on bias, etc.

So there actually is logic behind eliminating a horse in a Maiden Claimer with a morning line of 10-1. Especially when they are ranked #1 in EP and the ML maker did not make them 8-1 or even 12-1.

I am probably going to get much wilder with my plays now. It is very hard to come up with a spot play that can double your money. But the contest certainly has made me put together some wild combinations and I think I am actually learning to swing for the fences more because of it.

Jim

[qhrick]

Great topic and discussion

Backfitting obvious in many of the submissions, certainly in mine. Tho I did elect to stop at a point. Personally could care less about winning, just a challenge. And given the rules, backfitting certainly allowable and expected. I can think of many ways to make this a more "legit" contest and it will prob be discussed in detail should there be another challenge

But on some very positive notes
- this has created unprecedented interest
- could be a great marketing tool for the product

For me, it represents a learning experience. I don't really have much time to spend on this challenge so I decided I would pick an area of interest and see what I might come up with. I chose one of the consensus ratings ( HTR ) and tried to exploit it. That made some research manageable. And I simply starting mixing in a few variables. And in the process, I got more familiar with HTR in general. Time investment only a few hours, and well worth the effort.
The 2nd play I submitted actually did OK in 2005 and started 2007 nicely, so perhaps at least some validity in some of the criteria

Thanks to Ken for the contest
-rick

... and now, to the backfit cave, gotta find the magic 3.0 ROI combo!!!!!!!!!!!!

[DanG]

Jim, I understand completely what Donnie was asking and appreciate your response. I happen to be in the minority that has no problem eliminating the “median” value to allow the strong ratings and the randomness to occur.

Great topic and discussion

Backfitting obvious in many of the submissions, certainly in mine. Tho I did elect to stop at a point. Personally could care less about winning, just a challenge. And given the rules, backfitting certainly allowable and expected. I can think of many ways to make this a more "legit" contest and it will prob be discussed in detail should there be another challenge

But on some very positive notes
- this has created unprecedented interest
- could be a great marketing tool for the product

For me, it represents a learning experience. I don't really have much time to spend on this challenge so I decided I would pick an area of interest and see what I might come up with. I chose one of the consensus ratings ( HTR ) and tried to exploit it. That made some research manageable. And I simply starting mixing in a few variables. And in the process, I got more familiar with HTR in general. Time investment only a few hours, and well worth the effort.
The 2nd play I submitted actually did OK in 2005 and started 2007 nicely, so perhaps at least some validity in some of the criteria

Thanks to Ken for the contest
-rick

... and now, to the backfit cave, gotta find the magic 3.0 ROI combo!!!!!!!!!!!!
Rick, this is brilliantly written and really sums up the intent and spirit of the contest.

Thanks again to Ken for this stimulating exercise.

[pizzz]

Ken in response to all the backfitting questions and how to run your contest i believe the answer is quite simple.

Did you intend this contest to see who could attain the highest ROI through backfitting or did you want the superior minds of your board to develope spot plays that have realistic chances to be used year after year.

Simply did you want this to be a programming contest or ideas contest.

I have no input into the actually contest. I dont program and i dont look to the robot for spot plays.

It has been exceptional reading and learning for me through.

I will assume that the sniper[backfitting] plays as they are named will get the highest ROI

I will then guess that if you test them in past or future databases that they wont show the same great results, so what is the value in them?

Backfitting as i see it has zero forward value, but the spot plays that some have developed certainly might.

Ken i don't think you meant this do be a backfitting contest. But instead an attempt to help everybody's ROI in the future.

Just the opinions of a trip and pace handicapper.

Regards,
Mike

[DanG]

One point concerning the morning line.

I’ve yet to see a play (With a significant sample size) that wins universally across every track. Not saying it doesn’t exist, just haven’t seen it.

Obviously the year round circuits make up a disproportionate amount of a complete racing database. At least 124+ tracks yearly conduct racing where charts are produced. Yet, 3 tracks make up nearly 12% of your DB. (CT, MNR and PHA)

Concerning the contest it doesn’t allow being track specific, so it stands to reason that if your play suffers on the year round circuits, it will not yield good results.

Removing “10-1” for example…How does that impact your DB as a whole compared to the large sample tracks?

CT, MNR & PHA make up 15% of the 10-1 Morning line horses in your DB. That’s a 22% increase over their general population. Eliminating a “certain” morning line number does not impact all tracks in your DB the same. To me it’s a clever way around not being able to eliminate a poor producing track.

The real issue becomes one of functionality. The fact that an Access player can test & remove a particular value and the robot player at times can not.

As I have said…If I had a graduate degree in statistical analysis, I would probably never have discovered that one of Ken’s ratings produces a non-linear value table that is far superior to its rank. I may have never pursued it because it does not make sense in traditional terms.

Two of Ken’s best Algorithms are not completely linear in terms of the whole number rating. (WO & PED) Ken and I believe Mike Dee worked very hard on “pattern reconignition” in the WO rating and found gold in producing a powerful fitness indicator.

Am I saying throw all fundamental mathematical theory out the window and start betting bizarre factors with real cash, of course not. But, a very apparent thing to me in this exercise is one mans “back fit” is not necessarily that of another. Again…The sometimes vague rub always is…within reason.

PS: When I first started with HTR through the archives there was a group of guys who constantly discussed spot plays. The names are lost on me know (Fred4Now? etc) I know was one. If you can find any of their archived chats it was very educational. Some of these guys were WAY outside the box! :)

[tomcat]

For years at our seminars, we have always taken the position that HTR members engaged in a collegial relationship and our seminars have always been one of everyone helping everyone. This has developed in a lot of new friendships and better handicapping for all.
I'm sure Ken has the contests with a side objective of helping everyone with their handicapping.

With that in mind, that $100+ winner at GP yesterday was a result of the spot play....
Chestnut beaten by 8 lengths or more and must not have wraps nor medication. I don't even know if that is accurate but the guy I was sitting with had the $2 WPS ticket and that was his explanation for his selection! So Ken, we need to add Chestnut to the db (just kidding).

Good luck all, keep winning!

[DanG]

BTW-I:

Eliminator Alert!!!

Ron Tiller was the first I saw put this classic stat in print. MNR, 30-1 and up morning line. (Approximately the last two years)

5 / 1567 for a whopping 0.3% and $0.14 cents!!! :eek:

BTW-II… (SA race-7, yesterday)

Nice PED =1 (460) $99 bomb at SA yesterday. The HTR hit parade keeps rolling!!! :)

[dehere]

Dan - Jim

Not that this has anything to do with anything, but I was wandering around an antique store this week end and saw this poster for sale. I thought of you guys and your little discussion about Ted vs Babe when I saw it. Kind of a nifty poster i think.

[njcurveball]

I love it!!!! A man using a 50 ounce bat and holding it like a tooth pick. NO helmet, NO shinguards, NO elbow guard, etc. etc.

Also notice the catchers knees look unprotected! And these guys played for less than the normal working man. My Great Uncle Art, a slick fielding lefty first baseman was asked to play by the late great Connie Mack. He was offered LESS than his day job and promptly told Mr. Mack you cannot make mans money playing a boys game!!!

75 or so years later, the boys can make more money in a year than many lifetimes of working for us men.

Great picture Dehere!!!

Jim

[DanG]

Henry!!!

What a find that is! :)

If we could only take HTR back in a time capsule and watch the Babe, Citation and the great Ted Williams hit against a shift the way a real man would…by pulling the ball down their throats!

Thanks for that!

BTW: Jim, I know you’ve been to the BB Hall of Fame. That 50 ouncer of the Babe is a riot! The handle looks like something from a Bugs Bunny cartoon. What a great era.

[njcurveball]

Dan,

Never been to the Hall of Fame. In 1999 our league had their All Star Game and I chose not to go (something about a 22 year old gf). Of course I got the call the next day saying how I would have started on DoubleDay Field. Ouch!

I will get there someday! I saw the "Babe Ruth Museum" in Baltimore. Everything about him has almost a legendary aura.


Jim