Playing Spot Plays Forward

[MikeDee]

Now that all the spot plays are in and the winner is soon to be announced. We can ask the big question. How well do the spot plays play forward?

Here are the results for the month of feb 07.

Item Wins Plays Win% aMut $ROI #Lg HI
cent1 9 36 25% $11 1.41 0 $15
dang1 5 21 24% $7 0.84 0 $9
dang2 10 48 21% $20 2.08 3 $41
dang3 9 33 27% $20 2.77 3 $41
Dang4 0 10 0% $0 0.00 0 $0
john1 2 23 9% $10 0.41 0 $13
john2 10 39 26% $25 3.25 4 $109
mark2 6 34 18% $10 0.84 0 $18
mark3 6 37 16% $24 1.96 4 $41
mark4 2 34 6% $29 0.85 1 $41
mike1 7 40 18% $12 1.06 1 $21
mike2 4 47 9% $15 0.62 1 $24
nj1 4 19 21% $38 4.02 2 $107
nj2 11 46 24% $22 2.65 2 $107
nj2r 5 27 19% $29 2.67 1 $107
nj3 1 29 3% $92 1.59 1 $92
nj4 12 92 13% $33 2.15 5 $122
pizz2 4 31 13% $9 0.59 0 $12
qhrick1 5 35 14% $10 0.73 0 $13
qhrick2 2 32 6% $16 0.49 1 $18
114 713

allnoD 88 596 15% $19 1.41 20 $122

The spot plays listed here are the ones that showed a ROI of 1.20 (or there abouts) and up. These are only PL5 spot plays. multiple PL 3 and 4 are to much work and complexity for me to deal with in a single Db.

I did not concern my self with win percent or the number of plays. So even though some of these did not qualify for the contest they all qualify if you cash a ticket.

Most of you who tweaked your play the tweak is included. Except for NJ2 and NJ2r. The revision seemed to be different enough to stand on it's own. If you are following along on your own and you have something different different let me know and I'll recheck my work.

The last line is the composite results with the duplicate plays removed.

The "allnoD" item is unique plays, 713 plays resulted in 596 unique plays.

Not to shabby Team HTR a 1.41 ROI.:cool:

[km]

Hey, that's great stuff Mike - thanks a million. Saved me a lot of follow up testing!

The final accumulative ROI for the month hammers home what others have been saying, the diversity of plays will work in your favor, even if the individual ones sometimes fall flat for one month.

[MikeDee]

March spot play results are in and they are not pretty. A losing .89 ROI for the composite result with a meager 11% win rate.

John2 leads the way showing 2 months of positive ROI's NJ4 and Mike1 also have 2 months with positive ROIs.

Maybe with the on set of spring and good horses moving back north will bring positive ROIs back to the spot plays.

Details in the attached spreadsheet

[AwolAtHTR]

thanks Mike,

I had felt remiss about not following through to check out the contest spot plays.

This does confirm that the 2007 Spot Play contest is NOT LIKELY a '..go forward..' bet.

I do not think the spot play contest rules should be changed.

Personnally, I learned a lot about spot play construction trying to beat the latest high.

I feel the player benefit as a learning experience far outweighs trying to change the contest rules for a fantasy goal that the winner's spot play would work after the end of the contest.

[Zaf]

Thanks Mike !

[DanG]

Thanks for keeping up with his Mike…

March: 35%+ winning favorites and most of us missed that $310 bomber…

Let’s secure a loan from the Beulah Twins and II-down in April. Bomb plays are like marriage….They need constant nurturing and a leap of faith when all appears lost. ;)

[njcurveball]

Unfortunately I have had this same attitude in the past.

Jumping off after a losing month, only to find that the next horse was a $100 winner.

I think the bottom line is the figures for the year these days. I would prefer having one incredible month and a winning year where I make significant money to dribs and drabs of months with small profit.

That is all about someone's Investment attitude and no two people are alike.

I am not doubling down, but I would not lose faith right away.

Jim

[JohnB]

I have several spot plays that have shown consistent profits on a year to year basis, but have in reality been incredibly inconsistent on how those profits have been obtained, when you look at their weekly, monthly, or even quarterly performance.

I believe someone once posted a spreadsheet on this board (I apologize for not remembering who posted this) that calculated the potential losing streak that could mathematically happen based on a spot play's winning percentage. It also calculated what percent of bankroll should be played based on the winning percentage of that spot play. When I got into spot plays, this was a real eye opener for me. If memory serves me correctly, a spot play that returns 19% winners could mathematically possibly go through a 52 race losing streak (I remember those numbers because at the time my spot plays won around 19%). I think the spreadsheet spit back a recommendation of 1.90% of bankroll per play for a 19% win rate spot play.

For me, I have been blessed with a personality that can handle the swings that happen with spot plays. If I have a play that has shown consistent profits for multiple years, I have no problem sticking with a play that is tanking in the short term. I also bet way less of a percentage of my bankroll per race (I have used 0.50% of bankroll per play for the last couple of years), which reduces the risk of eroding the bankroll. This approach has helped me deal with the fluctuating nature of spot plays and has helped long term profits. Case in point - March was a losing month for me from March 1 through March 29. March 30 and 31 were both very good days, and put my profit squarely in the black for the month, and made it a fairly "normal" month for me in terms of my historical monthly profit.

On a related note, I have always thought that I could improve my bottom line if I could properly "time the market" with my spot plays. Unfortunately, I have never been able to come up with a system that I have been comfortable with during testing. Has anyone come up with a "market timing" approach that has worked for them? For example, if we know that rHTR=1 historically returns 0.92 on a WP of 28%, is there a way to jump in on betting the rHTR=1 horses after they have gone through a prolonged dry spell? I have done tests that have focused on betting rHTR=1 horses after they go through a streak of 25% less profit/wins than "expected" (i.e 21% WP and 0.69 ROI), but I have had trouble determining the proper length of "losing time" needed before the play "turns around". I also never was comfortable with the theory of betting a play that had a 3 day "losing streak", but might in fact be in the middle of a 7 day "winning streak". I am sure the folks that trade stocks based on short term technical factors would have a better handle on how to approach plays from this point of view than me.

John

[njcurveball]

John,

You make some very good points. You seem to have a lot of experience with this and you know the ropes.

I probably agree with everything you said. Stocks are often compared with racing, but really they are two distinctly different animals.

In my mind, comparing them with racing, is like comparing poker and blackjack.

Before I lose everyone, the point is that everyone at a Blackjack table can win, but only one can win at Poker (other than the rare push).

Stocks are traded today on perceived value, rather than actual. A recent example of this was the story released of how TJ Maxx compromised the debit card information of 47 million customers. Hearing that news, the stock went up 50 cents! :eek:

In an almost strange way, racing is actually more reliable than the stock market, since the value of the Investment is tied to actual performance.

One of the things that most spot plays depend on is the track. As the rules stated, these spot plays could not prohibit a track or group of tracks.

Being in NJ, you know there are things that are gold at ACRC that will cause you to lose your shirt at Monmouth.

One thing spot plays seem to do is bring out some of the best thoughts and experiences of this group.

Very worthwhile for all of us! Thanks for contributing.

Jim

[DanG]

Seriously great post John;

Sidebar…How Jim found the time to respond while juggling 37 fantasy baseball teams is just remarkable!!! :D

[Troy]

John, I use to play the stock market, I have thought of how I could apply what I learned there to horse racing. I don't know if you or anyone here has herd of it but you could possibly use a moving aveage of the performance of different factors and use that factor in your handdicapping when the fast moving average passes "going up" the slower moving average.

Troy

[JohnB]

Troy -

Thanks for responding. Could you possibly elaborate a little more on what you mentioned regarding moving averages? From what I have read regarding moving averages in the stock market, a rising moving average is considered a "bullish" signal. Do you suggest that it would be a good idea to create moving averages for handicapping factors and play the factors that have a rising moving average? I have always had a fear of taking that approach - it seems that it would be a more prudent approach to try to find a "bottom" as an entry point for a handicapping factor rather than using a rising moving average as a place to jump in. I could be very wrong in thinking this way - it certainly wouldn't be the first time.....

Jim -

I like your poker/blackjack analogy. One interesting thing regarding my daily spot plays - I do not seperate them out by track. I think I may be in the minority by doing it this way, and I am not sure my approach is anywhere near the best approach. One thing I require to put a spot play in action is a lot of bets per year. 1000 plays per year is my bare minimum to go "live" with a spot play, and my best spot plays generally produce many more than 1000 plays per year. For some reason, my brain is much more comfortable with a spot play with a high degree of plays, even if it has a lower ROI. I filter plays by surface, distance, class, age, etc. but I have not filtered down to the track level (yet). I plan on doing research at the track level in the future, but I have not used that as a filter to date. I suspect that it will be a very worthwhile venture that will open up a lot of new spot play avenues.

Dan -

Thanks for the compliment. I have learned so much from you and the others that post on this board. When I signed up with HTR I learned a TON just from reading the old posts on this board, and that continues to this day. I am always amazed by the class of the people on this board, and by the quality of information that is shared here.

John

[DanG]

Do you suggest that it would be a good idea to create moving averages for handicapping factors and play the factors that have a rising moving average? I have always had a fear of taking that approach - it seems that it would be a more prudent approach to try to find a "bottom" as an entry point for a handicapping factor rather than using a rising moving average as a place to jump in.
John
That’s what I’m currently doing John…

It can sure backfire, but I’ve found it to be MUCH more profitable than jumping in after a couple good hits.

Now that I have a two year record of HTR oriented spots, I have more than 90 individual funds and they go in and out of rotation according to performance. To your point they go in off “bottoms” and are removed off “tops”. (Most are track specific, so I do make allowances in time frame for meets sometimes.)

I didn’t truthfully know going into this if it would work, but I’m such a believer in contrarian investing that I had to give it a go. I told myself I would not waver from this strategy once for 12 months and I’m now + 23% in the black.

Sure I feel stupid when I remove a play after a + $1.30 “peak” and it goes + $1.40 next month…But, as they said in the Godfather…”This is the life we have chosen”. Hard choices have to made and my experience is the ones that don’t necessarily “feel” safe in our guts are the ones we actually profit from.

BTW: Less than 1/2 of the individual funds are actually winning, but some are doing spectacular and they are carrying the portfolio as a whole. One in particular I’m very proud of is $1.08 W / $1.04 P (pre rebate) with a 61% W, 84% WP rate and 1,200 plays a year. That’s my “piece of mind ATM” and with no exaggeration its one of my first real success with chalk in my life! :o

[njcurveball]

Great post Dan. The problem with people who lament about leaving "money on the table" is that they spend way too much on getting it.

Looks like you have a nice set-up with diversification that gives you much needed security.

Only one Fantasy Baseball team for me and I started in last, spotting the field 50 points! I am usually in first by Memorial Day, so no worries. :)

And to pull those thoughts together, you may be surprised how managing a Fantasy Baseball team is actually very similar to this discussion.

Some guys have good stats against a certain team, are on a hot streak, cold streak, etc. It is far from an exact science. Just like handicapping.

No one wanted to touch Gil Meche with a 10 foot pole in the Draft, but after he beat up Boston, he will be picked up. Do you pitch him off a "top"? Hoping he pairs good efforts? Will he bounce? All good stuff!


Jim

[Troy]

What I thought you could do is make a 7-8 day moving average of the ROI or the WIN % of a paticular factor and overlap that moving average with a 3-4 day moving average of the ROI or Win %, and when the 3-4 day the "faster" moving average passes, crosses over, the 7-8 day "slower" moving average, that would signal it would be time to use that factor. the faster moving average reacts much faster to ROI or WIN % changes so it will cross back and forth over the slower one telling you when to use the factor and when not to.

Troy

[JohnB]

Troy -

Ok, now I get it. Sometimes I have to hear things twice before they sink in :)

I'll look into that approach - thanks for the suggestion!

John

[Troy]

John, here is a few definitions with pictures of different moving average pictures used in the stock market.

http://stockcharts.com/school/doku.php?id=chart_school:technical_indicators:movi ng_average_conve

Troy

[JohnB]

Troy -

Great link. Thanks!

John

[overdog]

[QUOTE=DanG;61505]Thanks for keeping up with his Mike…

March: 35%+ winning favorites and most of us missed that $310 bomber…


Dan: On a lark, I hit that $310, which was logical(?) or at least a decent case could be made for it using Pitz's Key race file.

I have found its best to print out the keys, then cut them out and physically paste them onto the top of my BRISNET form. I mark the interesting keys, THEN, put a large $ sign beside the horse's name on the "form." Only then do I handicap the race or look at any horse in more detail.

I find if I 'cap first THEN look at the keys, my subconscious logic(?) overwhelms my willingness to throw a whacky(?) horse into an exotic betc.

overdog

a.k.a. Fraser Rawlinson

[overdog]

JohnB wrote..."Has anyone come up with a "market timing" approach that has worked for them?"

Hi John. At the risk of being thought crazy(stand in line and take a number) check out the link given below. It mentions Kagi charts.

Its about technical analysis of stocks. T.A. is all about timing, trends and their turn-arounds and any subsequent, resulting strategies. Traders and trading houses that I do some neural network stuff for are always trying to tweak existing theories, charts etc to cover new territories, industries, IPO's etc.

Kagi charts are my latest fun experiment. Think in terms of doing a Kagi chart on a single HTR factor. My current intertests focus on Cramer figs, F2 figs, PER, Pac, SP and AP.

You MAY find it interesting! (Hint: I have tried Kagi-izing these respective factors with BOTH a % change(pos or neg) AND also using an absolute amount of change.)

I feel there are very many similarities between the day to day sentiments of stock traders about a specific stock, industry or overhanging piece of news(i.e. prime rate%) and the race to race sentiments of handicappers opinions of a given horse, track, distance or ??

For the record, I think most will agree that TG or Rag sheets are really technical analysis in sheep's clothing. They are sort of visual trend analysis/projection or extrapolation.

For those still on the fence over whether or not I'm nuts:

I am also experimenting with Japanese Candlestick charts for individual horses, using subsets of age/sex/DISTANCE/surface etc. These I configure to change color(think underlying trend directional CHANGE) relative to the horse's lifetime or last 10 or best of last 5 at the distance.

http://www.investopedia.com/articles/trading/07/kagi_chart.asp

I would LOVE to hear from any of you with your thots on my latest insanity!

overdog

a.k.a. Fraser Rawlinson

[DanG]

Dan: On a lark, I hit that $310, which was logical(?) or at least a decent case could be made for it using Pitz's Key race file.

Fraser,

Nice $310.00 score! :)

After reading your posts I can safely wager you can problem solve in minutes what would take me weeks. :eek:

Its interesting you mentioned the “order” that you handicap in influences your potential selections. I think this is a much underrated area that many fail to address.

In my case I watch every replay of all turf horses I’m not familiar with, or when my notes are lacking. Up to about 6 years ago I always did my replay work the night before and my technical work the next morning. I then tried it in reverse order and my results almost immediately improved.

The “subconscious” you mentioned is a very powerful thing. There are hundreds a subliminal triggers that cause us to wager / or not. If I have a very strong mental image of an animal’s last race it absolutely affects my judgment when I look at HTR. If a horse is just numerically completely overmatched I’m more prone to poor gambles if I have done my visual work first.

It seems you’re not alone in the high praise of Gary’s key-race work. Very generous of him to post this service daily.

BTW: Your 100% correct in your description of the sheets. I was a Ragozin customer for a few years. (Probably explains my borderline animosity to TG ;) ) When the “originality” of graphing performance was mentioned a market trader told me that this has been done for many decades long before Len’s father breached the idea for horses. There is much to learn from market traders that can be directly applied to pari-mutuel wagering.

Now that I finished typing I realize I’m dragging this thread off the topic of Mike’s monthly scorecard. Ken, Rick…feel free to delete this as you see fit.

[njcurveball]

Fraser,

WOW! A $300 horse!!! BIG BOOYAH CONGRATS!!!

Imagine having this horse in a live money tourney! You would be off to Vegas for sure!

Nice job!

Jim

[overdog]

Thanks, altho I cheerfully give all credit to "saviour" Gary.

I love the key race file and love getting under the mathematical hood and tinkering with the file. I import it into Excel and am trying various weighted scoring methods to get a point value for each horse, and separate total for all horses in each race.

Very much a work in progress. Trickiest part is the class of the last race of those who've returned, compared to the class of today. Also, factoring in recency is important. When a horse is the last in a 12 horse field to return, undoubtedly the race is old(er.)

Should I divide each horse's points by his days off, or by the days he's been off longer than the average of the field's returnees to date. I COULD go on...but mercifully, won't :)

As everyone here knows, when you love the daily (hell..."minute-ly!) intellectual challenges of it.....its a true labor of love.

My two other main projects are:

1. Devising my own class rating.

What if you divided a race's purse into the pace segments and considered each race to be 3 separate races: to the first call; the internal turntime; and final fraction? So instead of the winner getting 60% of the entire purse, each horse who was fastest in each of the segments would get 20%(60% win share/3)? Second in each segment would get 20%( or whatever place % is,) divided by 3. Same for 3rd and 4th.

2. Track DAILY bias calculator/corrector. (Mountaineer, lately, anyone?)

Compare a given day's race results (winner's normal run style, and/or lengths behind at calls 1 & 2, and or win, place and show energy %) to the age/sex/surface/class/distance pars, and/or last 20-30 races and calculate how much to penalize a day's "abnormal" winners, and credit a day's bad bias losers.

As always, I remain available to discuss this further, on this board or via email, if the threads are too boring or ? for the majority.

Fraser

[DanG]

Great stuff as always Frazier,

I do something very similar to what you outlined in #2 regarding the profiling of specific days and assigning a “bias score”. This is very useful when they run back and adds a third dimension to animals that run from the extreme running styles. (I.e.…RS F, S and R).

The last few years I am a big believer in completely reconstructing horses prior few races from the perspective of “before” they took place, rather than strictly in a revisionist sense as most approach them. Perfectly valid observations and conclusions are applied before they leave the gate. Too many players disregard their information going into the race choosing to completely alter their perception based on a single result.

BTW: (A personal HTR / PPX is what I'm attempting to describe.)

#1 regarding attaching race segments to purse values is an original concept as far as I know, but I would be very careful attaching purse money in modern racing. Many horseplayers (myself included) were using EPS 25 years ago and it was a viable factor. It is currently little more than noise if applied in its raw form.

If one takes the time to work through all the permutations of state bred funding, ‘rascino enhancement, fluctuating exchange rates and the very real breeder driven graded status vs. 60% of the 600k Winstar Derby. I feel in the modern game time would be better spent going in another direction other than pure purse money. Just my opinion obviously and knowing you a new approach would be attempted, something I never would have dreamed of.

BTW: When your graphing factors what software do you use if I may ask?

[overdog]

Hi Dan;

Mark Cramer talks about dialectics in his books. The use of polar opposites. I say 'looking for the dark lining in silver clouds." We are all spending countless hours searching for ever better algorithms to uncover winning horses. The biggest single leap my handicapping ever took was when I decided to uncover/discover "what do LOSING favorites have in common?"

My class experiment is actually showing great promise. Perhaps because, as you mention, and like we all know, EPS, in its "pure" form, is of extremely limited value. Note the emphasis on pure. With very slight tweaks, it still can have value. (Surprisingly small tweaks, in fact!)

I take due notice of the class "muddying" factors you mention. In most cases, only VERY recent shippers, upset the apple cart. State vs homebred is also easily... adjustable.

I have attached a quick and dirty Excel sheet, showing one possible approach. Hastings main sprint distance is 6.5 furlongs. The DRF timed segments are 2 furls; 4f then of course 6.5.

For simplicity sake, ASSUME A WIRE TO WIRE WINNER. Further assume purse money is distributed for the three "races" within the whole race: Start to the quarter; start to the half, and the normal, start to the finish.

2 furls is 30.77% of 6.5 so give the first call leader 30.77% of the winners share(of 60%).
(OR, for the entire $10,000 purse...as of course the proportions will stay the same.)

Do the same for the half mile(4 furls), which is now 61.54% of the total distance of 6.5 furls. And of course 6.5 is 100% of 6.5 so add that on too. We now have a total of 192.31%. The cumulative purse awarded this way is $11,538, and obviously 192.3 % of the winners true share of $6000. This way the winner will SOMETIMES get the cla$$ award of the actual winners share in your state/province. But IF and ONLY IF s/he went wire-to-wire and won all three "races."

Off pace horses are heavily penalized using this method.

This is easily corrected when we apply accurate weights to each of the fractional parts of the race. So for turf races where FAR off the pace horses win, we might know for example that only 7% of first call leaders win; and 18% of second call leaders prevail to the wire, etc. So we would knock the first call leader's cla$$ rating down to 7% of whatever he "earned" under my 3 call class system for leading to the first call. And so on.

It's a pain in the butt, altho not terribly hard to implement and the results WILL surprise you!


As for charting....if you mean stock market charts, I use some of my client's stock market charting software.

Otherwise, and even IN many of my personal applications, I invent my own in Excel. I can send you an example if you want.

Fraser

[DanG]

We are all spending countless hours searching for ever better algorithms to uncover winning horses. The biggest single leap my handicapping ever took was when I decided to uncover/discover "what do LOSING favorites have in common?"

Very interesting Fraser;

Thanks very much for sharing!

Much to digest there, but your quote above is as important as any one approach in our game. That’s such a high percentage of the key to the money room. When we correctly anticipate a portion of the 68% failed favorites, our advantage soars!

Not to say solid favorites can never be exploited, but the “bad fav” that yourself and Ken focus on is worth all time we invested in it.

PS: I was wondering reading your post if you “weighted” according to profiles for surf / dist / track etc…I smiled when you alluded to the “pain in the butt” factor. I have much experience in that area and it is a challenge to say the least! :eek:

Thanks again; All the best this weekend. Word is spring is arriving soon to all our HTR family north of our border. :)

[Huguenot]

Fraser,
Have the Kagi Charts approach borne any fruit for you in tracking trends? I read the material in the link and it was very interesting.



JohnB wrote..."Has anyone come up with a "market timing" approach that has worked for them?"

Hi John. At the risk of being thought crazy(stand in line and take a number) check out the link given below. It mentions Kagi charts.

[overdog]

[QUOTE=Huguenot;61748]Fraser,
Have the Kagi Charts approach borne any fruit for you in tracking trends?

Yes they have! I am tracking 3 or 4 different amounts of % change in horse's last race figs, as my approach.

So far they are interesting, most especially with young developing horses, exactly the kind we track for the Kentucky Derby.

A day before the Derby I will post some derby picks, based on my own "Kagi" together with my odds lines for each. This far ahead and in a (presumed) huge field, I have no idea how many final possible "live" horses it will discover.

The approach is less distinct with other subsets. But I lost my 2006 data to a hardware failure and have too small a database to project from. Also as the season wears on, and more 3 year olds (and possibly lightly raced 4's) start, and I have more data, I am hopeful it will offer some opportunities. Saratoga and Oak Tree at Santa Anita may also be good.

Also, with solid claimers who seem to specialize at ONE distance, it has had success. Currently with 6 furls, 1 mile and 1 1/16th.

Go to the same web site and enter a search on "point and figure" charts too. That can pick nice "turn around" or sudden, big improvement horses too. Will take some playing around to find the best size of changed figs to track, with the Cramer figs or Per, Pac, or SP or AP.

I suggest about 5%. So a horse whose PER is usually 90-95....try using change of 4 or 5 pts.

Good luck!

Fraser Rawlinson

[overdog]

Hi everyone;

Thot I'd start a new(?) thread, an offshoot of a question in JohnB's post about market timing, spreadsheets and losing streaks etc.

He mentioned the sheet listed a 19% win rate and that someone had said he could have as many as 52 successive losses, at that rate.
Also, that he was betting 1/2 of 1 % of bankroll on an angle.

There are some problems with those numbers. To know exactly how much of bankroll to wager on a spot play, we also need the average price of those 19% winners.

I have attached an Excel workbook with 3 separate sheets.

They are quite similar but each demo a slightly different look at the same sets of numbers.
In all three sheets, down column "A" are numbers of consecutive losses, starting with 10.
Across the top row, are different win %'s starting with 15 and going up 1 % each, thru 19%.

Sheet one, so labelled at the bottom, is percent chance. So each cell contains the % chance of a given # of consecutive losses associated with the win% at the top of each respective column. In row one.....for 10 consec losses at 19%, that is a 12.2% chance.

FOR THE MATH CURIOUS. A win % of .19 is the same as a losing chance of .81. And 10 consecutive such losses is .81 X .81 X .81, until you have 10 separate .81's. In Excel the formula is = .81^10.

The second sheet lists the odds against that many consec losses at a specific win %. (For 10 losses in a row at 19% win rate, that 12.2% equals odds against of 7.2 to 1.

The third sheet shows how many SETS of 10 consecutive races, (or spot plays) you would have to go through before the odds would be 50-50 that it would/would NOT happen.

In this case, 5.0 sets of 10 races. (Don't forget this can include rolling pick 10's!) So the first 10 race set includes races 1 thru 10, the 2nd set includes races 2 thru 11....etc.)

For the record John, the % chance of 52 consecutive losses with a 19% winning spot play, are 0.00174%.

The second sheet says that translates to odds of 57,381.4 to 1 against.

And..the number of SETS of 52 consecutive races that you'd need to see to have a 50-50 shot at actually losing that many...is...drum roll...39,771.1

I feel totally safe in saying that if you played EVERY single race all day in a Vegas race book, and played the outside post every time, you could go for a month and not have that losing streak.

Hope this helps.

I will be sending two separate sheets along. One uses random number generation and simulation that members can play with, and adapt to their personal parameters. The other addresses SOME issues about the likelihood of some of Ken's contest spot plays, will carry forward. It will show assorted confidence levels that the win% of the spot plays will not slip more than x%.

Fraser Rawlinson

[overdog]

Here is a spreadsheet which you can adapt.

I've had to cut this down dramatically to fit into this forum's file size upload limit.

Anyone with minimal Excel skills can adapt this extensively.

Currently it uses JohnB's mentioned .005 % of bankroll. (1/2 of 1 %)

To change that go into D2 and change the .005 to whatever % of bankroll you want to bet.

With this sheet everytime you press F9 (the function 9 key) it will recalculate the entire sheet.

I also suggest extending the columns I and J to the right out by 50's up to 750 or even 1000.

Row two shows the betsize you'd be making at bet #50, # 100, etc etc. If your spot play doesn't have long term profit built in, this will shrink, quickly or slowly depending on the parameters of your play.

What you are looking for is when the betsize in row 2 drops below $2.00. I have conditionally formatted the row to display the figures in bold, italic RED FONT, when the bet size drops that low. That will show how many bets you'd get to make before your suggested % bet based on your declining bankroll, would necessitate a bet below the universal $2.00 win bet minimum.

Row 3 shows the total money you'd have bet up to the point that your bankroll declined to below the bettable minimum.

If your spot play shows a long-term profit, these numbers will grow.

Questions, comments etc are welcome,

overdog

a.k.a. Fraser Rawlinson

[DanG]

Is it the weather that stimulates the minds in Canada?

Thanks very much Fraser for your generosity and knowledge. These are very helpful / instructive tools.

[overdog]

Nah Dan....just us math perverts with too much time on our hands and a laptop and Excel handy...:)


Fraser

[overdog]

Htr Spot Plays....Safety Factors

The attached spreadsheet should be of major interest to all horseplayers, or at least those smart enough to be HTR subscribers.

If you save it as is and then copy the bottom row and all its formulas down, you can track your own spot plays, (or anyone elses for that matter!)

It shows some sample rows for both "normal" favourite win %'s (33) and also three rows at the 19% win rate JohnB mentioned.

For both the 33% win rates of faves and John's 19% spot play, I have used those same winning %'s but with three radically different sample sizes.

Namely faves at 33 of 100, 333 of 1000 and 3330 of 10,000. John's spot I used 19 of 100; 190 of 1000 and 1900 of 10,000.

I also listed several of the spot plays in KM's very entertaining and highly educational datamining contest.

The data shown is whatever KM included when he scored them. Some of the submitted entries are red flagged (red cell backgrounds) because they were DQ'd due to insufficient plays or too low R.O.I..

The STOP! column is just to tell me to stop entering data. (Too many spreadsheets constructed at 4AM necessitate this!!!)

The sheet lists safety factors which I will explain using JohnB's sample spot play, the 19% winner.

The column headed Win% is obvious. The yellow columns headed, 90.%; 95% and 99% mean that the win% for that spot play on that row is THAT certain, to be no lower than.

For John's spot play with only 100 races, it means...

DESPITE the 19% win rate attained, the sample is small and therefore "YOUR EXPERIENCE MAY DIFFER!!!"

John, you can be approx 90% certain your angle's TRUE WIN % is 14%.
With the same 19% win rate over 1,000 races you can be 90% certain its not lower than 17.4%
Finally, with a large sample of 10,000 races its 90% likely that the true win rate is not lower than 18.5%

But notice how rapidly your win rate can shrink. With only 100 races, you can be 99% that your actual win long term win rate is not lower than 9.9%

The 3 columns headed r90; r95 and r99 show the % "decay" of the statistically likely win rate from your original angle.

It shows that with only 19 wins from 100 plays, and the drop from the 19% win rate to the possibly as low as 9.9%, your actual win rate could be 48% lower than expected.

The mathematically curious can dissect the formulas and find the statistical formula that computes the 90; 95 & 99% safety factors. (For other math perverts, they are taken from standard Z-scores, based on assorted standard deviations inherent in small, medium and large samples.)

Hope this is of help.

Fraser Rawlinson

[Mark]

Fraser, your chart is a wonderful reference tool.

Thank you.

[AwolAtHTR]

Nah Dan....just us math perverts with too much time on our hands and a laptop and Excel handy...:)


Fraser

aah Fraser, you have given me the opportunity to discuss a back burner project.

The per cent of bankroll for spot play management is reasonable. However, a person really just '..pitching pennies...' out of their bankroll to have fun betting. So, a person can just relax and have fun betting knowing that they will never run out of money. Steve Fierro in his Four Quarters book supports this plan along with recommending that a person have at least one and even TWO reserve bankroll. In other words, to bet ten dollars per spot play (or event with a Pick) the bankroll of ten thousand ($10,000) is required. Oh, so to play this penny pitching game, I should have thirty thousand dollars to have fun betting ten dollars a race!!! aaah, let me admit to being a gambler playing a game of skill who has bet over a hundred dollars in a race and had less than one thousand dollars in my betting budger.

Now, let us say we do not want to pitch pennies by playing spot plays but go to the track and have some fun betting. So, the daily budget is the reference measure for each event. (aka DBRM). Using the ticket history from the daily betting, the daily handle can be computed (aka DHDay.). The bettors measure of productiviity will be how many times did the bettor bet the daily bankroll, (ie, DHday/ DBRM). Some refer this to churning. The daily gain (aka GainDaily) is the cash at the end of the day. Oh, an alternative name for this field would be (aka2nd, CashOut). The daily gain ratio will be how many times was the daily bankroll duplicated but as a simple multiple (ie, INT(), Interger value). The computation is simply GRDaily = INT( GainDaily / DBRM )

So, each person will have a different set of GRDaily and the number of zeroes will tell you how many times did the player leave broke.

In other words, sometimes a player will leave the track with five hundred dollars after betting two thousand dollars and started with a only a hundred dollars.

----------------------
ok enuf about the general nature of redefining bet management plan.

-------------
let me try to describe the back burner project:

a. bettor expects to lose daily bankroll.
b. bettor expects to Win most months..
c. bettor expects to Win annually.

The problem is creating the data base to track and report the betting.

The extension of the basic problem is the Betting Improvement report which will recommend how a person change bets (using SAME picks) to reduce the number of losing bankroll days.

----------------
so, Fraser, the issue is: How do I describe the project for a Math pervurt?

aaah, in other words, a data junkie has an idea and wants someone to join the project that has fun playing with the numbers.

AWOL
aka data junkie

[overdog]

Before we can "partner(?)" on anything I need to know:

(1) What exactly you are "AWOL" from ? :)

Seriously, what I think you're asking is really quite simple. So...some background.

Unless I miss my guess, you want to "optimize" your return from racing.

Much of my consulting practice revolves around my math specialty which is optimization.
I haven't "Googled" optimization theory lately, but it may help you get a (hopefully) brief explanation.

A gross over simplification from another form of gambling, BlackJack (21) may help explain optimization as applied to finance(investing) or gambling on races.

There are two possible aims of optimization: Maxi-Min and Mini-Max. The former is to MAXImize the MINimum profit you can show from some activity. I.E. if the activity is maximizing the minimum profit at BlackJack: doubling down on a two card count of 11 is the biggest single tactic in the game.

Mini-MAXing is undertaking a tactic to MINI-mize the MAXimum loss you can absorb from a risk or activity. In BlackJack, think hitting a hard 16 when the dealer has a 7 or higher showing. That tactic, long term LOSES money....BUT hitting 16 loses LESS money (long-term) than doing anything else with a 16 and facing a dealer's 7 or better.

I don't know if you've remembered reading that I was a member of the team that developed the DR Z system years ago that was only the 2nd ever mathematically PROVEN gambling "system." Thorp's "Beat The Dealer" was the first and spawned an "industry" of casino playing counters. So I DO have some experience designing math methods that work.

So you have to deal with what the investing(betting?) public THINKS works in a given market and find out what REALLY works.

I titled this email Portfolio Theory because its something that myself and many others, globally, are doing regarding betting on multiple games of skill; sports betting, poker, and of course racing.

You divide your total bankroll into 3-4 or ?? separate revenue "streams." You then keep all up to date, and shift money in & out of each of the separate types of investment(bets) according to recent trends etc. Segmentation DRAMATICALLY decreases your overall risk.

As demonstrated in the EXCEL sheet I sent, assuming you wanted to play say 5 of the HTR Spot Plays, you'd develop some models, run some simulations, start playing, and then decide how often and how much circumstances of each of the spot plays dictated that we tweak each individual one, up down or stay the same.

Models can be RELATIVELY simple at best, and excruciatingly complicated and problematic at worst.

I see you live in WA which is just south of me in BC. Any time you want to pop up I-5, give me some advance notice and we can talk. Or.......??

Fraser Rawlinson

[AwolAtHTR]

-----

(1) What exactly you are "AWOL" from ? :)

Seriously, what I think you're asking is really quite simple.
-----

I titled this email Portfolio Theory because its something that myself and many others, globally, are doing regarding betting on multiple games of skill; sports betting, poker, and of course racing.

---
Segmentation DRAMATICALLY decreases your overall risk.
---
Models can be RELATIVELY simple at best
---
and excruciatingly complicated and problematic at worst.
--------
me in BC.
---------


thank you for your interest in the problem.

I would like to read your complete email concerning Portfolio Theory.

oh my AWOL is A--Winner--Or--Loser
and the list of states for where I am from is the protracted way I answer the question: Where are you FROM?

whereas the brief answer is: A problem solvers' colony.

When I have the project defined in terms of the analysis parameters,
I look forward to meeting you to discuss the validation of the procedure.

[njcurveball]

I don't know if you've remembered reading that I was a member of the team that developed the DR Z system years ago that was only the 2nd ever mathematically PROVEN gambling "system." Thorp's "Beat The Dealer" was the first and spawned an "industry" of casino playing counters. So I DO have some experience designing math methods that work.



I have to offer my admiration and congrats then! You were part of a "strategy" change for an entire generation of horse players.

Thanks for your contributions both then and now!!!

Jim

[overdog]

Thanks Jim.

Recent research suggests people who spend time daily in stimulating mental pursuits can significantly reduce the likelihood they'll get Alzheimer's, and/or dramatically slow its advance.

I suggested to a track mgr that I go to Geriatric wards and teach handicapping. He looked at me like I should be sent to the looney bin, myself. (He's probably right, altho for other reasons. My wife extends the "women are from Venus and men are from Mars" by adding and "Fraser is from west of Pluto.")

I LOVE this game! Despite playing it for close to half a century, EVERY day I get some crazy thought...."What if I tried.....etc etc."

I have a more than passing interest in NJ. I spent many months there, separate times, from June 98 thru Oct 2004, working on assorted neural network and other software projects. Stayed in Oakland, near the New York state line, on a small farm with retired standard breds on the property. Loved it. Got to the Meadowlands fairly regularly, to Monmouth 3 times, (Point Given winning the Haskell) and Atlantic City once. Visted the NY OTB in Suffern NY almost daily, for R & R. Into NYC on the train frequently too. My partner died prematurely at 52 of a heart attack or I'd still be "commuting."

I admire your skills and am working to catch up with my Access knowledge. Want to be a dark horse in next year's Best Spot Play Contest. I think a more interesting category might be "Highest Win % showing a profit, over the most plays over 500."

Best of luck always,

Regards,

Fraser

[Huguenot]

Overdog,
could I see just one of your kagi charts as an example to help get me started?

BTW, I did buy the Dr. Z book about 20 years ago -- and when I try and develop spot plays on low-odds horses, the place odds are much more generous than with higher odds plays.
With rebates -- assuming you're not counting on the $2.20 return that off-shore places won't rebate -- I wonder if you can make a nice steady income with a spot place method that wins at 60-70% of the time and returns a consistent .98 to $1.03 ROI with about two plays a day.

[njcurveball]

I have a more than passing interest in NJ. I spent many months there, separate times, from June 98 thru Oct 2004, working on assorted neural network and other software projects. Stayed in Oakland, near the New York state line, on a small farm with retired standard breds on the property. Loved it. Got to the Meadowlands fairly regularly, to Monmouth 3 times, (Point Given winning the Haskell) and Atlantic City once.




WOW! You get around! Shame you got to ACRC after the "fire sale"! You probably saw a non working tote board shell, no lights and all turf racing (not a bad thing there). You may have even got to buy a hot dog or some Dominos pizza from the high school band.

Perhaps you even heard my brother call the races. He was at the right place at the right time to succeed Larry Lederman, and did a great job for a few years only to see the dream crash when they pulled the plug on live racing.

He even lost his "4 day gig" when Philly Park bought the track. They send Keith Jones down for punishment I think. He is a great guy but HATES being sent there for a few hours for what has to be little or no pay.

You also have some great ideas! Racing is a dynamic game more about profits than winners. You seem to be grounded in the right places!


thanks for sharing,

Jim

[overdog]

Huguenot wrote "...

(1) could I see just one of your kagi charts as an example to help get me started....

(2) I wonder if you can make a nice steady income with a spot place method that wins at 60-70% of the time and returns a consistent .98 to $1.03 ROI with about two plays a day.

Re: (1)

I have been doing them by hand on a notepad of quadrille(square) graph paper. I am in the process of transferring the method to Excel. In excel I was trying to do it entirely by "hand." This means using the Format; Cells; Border choices and then entering the lines one by one. Then outlining the relevant sideways lines with a single line, and upwards improvements or downwards regressions in double wide lines, in say the Cramer fig Visually it looks ???? I will see what I can do. It may take til early part of next week as weekends are busy for me.

RE: (2)

As a Canadian I can bet offshore, BetFair, Pinnacle, etc. And yes, you can do "quite nicely" betting place on over win bet, under place/show bet horses. Even more so if you also play harness, where shorter price faves are the norm.

I have gone to tracks of both type and used nothing but the morning line and tracking the flow of money on the tote thruout the betting and then made place/show bets. No handicapping whatever, and only marginal extra expense for a simulcast program. No "Form" etc. I usually pack a Sharp handheld programmable calc/computer which can calc minimum place and show prices before the start and decide if the return is high enough. Doing this I can bet upwards of 10% of bankroll with almost complete impunity. At some small(er) tracks, notably harness, you CAN run into small pool problems, where I bet of over $75 or $100 might seriously dent the place/show price. In those locales I have an extra small routine in my Sharp which tells me how much to bet to avoid denting my own price by more than a fixed %. (I use 20%) SO if the place price is currently paying at LEAST $4.00 (harness) and a $100 bet would lower it to $3.20 or less, then I might bet only $75.

You can routinely grind 30-40+% R.O.I. daily. You DO have to take money OUT of your bankroll regularly, (every 3-4 days at harness) maybe weekly at a smaller to medium T-bred track, to avoid constantly having to downsize your bets. Playing a 50% Kelly Criteria bet or even a 25% (quarter Kelly) makes bankroll loss all but impossible, and is sort of like having 150 ATM accounts all across North America. Its a low stress way to play the races.

Its NOT much fun.:(

Fraser Rawlinson

[overdog]

AwolAtHTR wrote "...I would like to read your complete email concerning Portfolio Theory.."

No you wouldn't ! :)

The originator won a Nobel Prize (Economics). His paper is a very tough read.

But if you insist....a clear(?) explantion is wikipedia's at the following link.

http://en.wikipedia.org/wiki/Modern_portfolio_theory

If you skip over the equations and just glance at the diagram/graphs and read the text, the rationale becomes clear.

Best thing? It also works.

The overriding principle is to reduce the "scatter" or "spread" or standard deviation of win prices of your total portfolio of bets. In the same way a mutual fund invests varying amounts in different industries and countries, you would have a safer(more predictable) return in the WPS pools over exactas, tris, supers etc. Your average losing streak would be much longer in the exotic pools. So your % of total bankroll might be only 5% allocated to supers, while 60% was set aside for win bets....etc. You might only bet 1/10 of 1% on pick 6's but perhaps 4-5 or 6% of your separate win bet, bankroll.

Regards,

Fraser Rawlinson

[Paladin]

...
You can routinely grind 30-40+% R.O.I. daily. You DO have to take money OUT of your bankroll regularly, ...Fraser Rawlinson

I really don't mind. I think they call that "profit".
I enjoy reading your interesting posts Overdog. :)

[DanG]

I really don't mind. I think they call that "profit".
I enjoy reading your interesting posts Overdog. :)
Agreed!

Fraser is always an interesting read.

BTW: Congrats on your Canucks last night!!! :)

[overdog]

To Paladin and Dan;

Thanks for the kind words. HTR forum is a great "board." Notably for the very high levels of expertise of all posters. A lot of entertainment value as well, and finally, the tone we take with each other is always warm, friendly and helpful.

Paladin: love your handle. Loved the TV Show. He was the quintessential professional, in ANY field. Some reviewer once wrote that "Boone's face looks like a relief map of Nevada." I loved that and so did Boone apparently. Miss Twilight Zone too.

Dan: winning the Spot Play is a serious credential on a man's resume, IMHO.

As for grinding 30-40%, a lot of people get bored with $2.60=$2.80 to show. I use those as they appear, to help defray expenses, etc. Have gone 17 days in a row without losing a Dr. Z show bet. I find them a good way to reduce boredom on days when we're all choking in clouds of chalk dust.

Good luck to all HTR'ers...

Fraser

[Paladin]

...Paladin: love your handle. Loved the TV Show. He was the quintessential professional, in ANY field. Some reviewer once wrote that "Boone's face looks like a relief map of Nevada." I loved that and so did Boone apparently.
Overdog,
This link should bring back some memories:

http://www.hgwt.com/ballad.wav

________________________________________

[overdog]

There are some old re-runs on TV now, and I've introduced him to my 9 y.o. granddaughter> She said it best...."Paladin is WAAAAY Cool!!!" :cool:

Thanks for the clip!

Fraser