Simple Selection Method

[hermes]

A SIMPLE SELECTION METHOD

See my placegetter stats posts on this forum for some background info what I am doing. I'm trying different factors, running them through sample races, looking for the factors that score hits.

Here's a simple method that seems to work surprisingly well. A lazy approach to punting but as I say my stats say that it works.

It combines two simple factors:

*A high proportion of placegetters (and weinners) were last start winners.
*There is a bias towards the lower saddlecloth numbers (weights in handicaps).

So, the selection rule is:

Select the last start winner with the lowest saddlecloth number. If you have two last start winners, numbers 2 and 5, number 2 is the selection.

Don't look at anything else, just those two factors. I didn't expect it would come to much, but I was surprised.

Tried it over 180 races, all races, any races. Results were about 54% place strike and about 30% win strike rate. (53 winners in 180)

The highest win pay was $17.80, the highest place pay was $3.80. The lowest place pay was $1 (three times).

After 180 races betting $1 to win, $4 to place each race. That's 180 x $5 = $900 outlay. Total returns were a healthy $1095. A result not to be sneezed at for such a simple selection method. The average return per race is $6.08 (less your $5 bet) which is a comfortable return rate (for such a simple selection method).

Would it be sustained over a larger sample? Don't know. But on first inspection it provides a good volume of hits, with not too long runs of outs, and prices sufficient to give a return.

I tried other modes of staking. $1 to win, $4 to place works best (for reasons given by other contributors to this forum). I was really after placegetters but this method throws up a good proportion of winners too. You need the winners to make it profitable. No good for placebetting alone- margins are too fine.

But the thing is, it must be done *raw*. I've tried to be smart and find other factors to add to try to improve the results. I tried being more selective about races, for instance. But no good. I tried making the selections more intelligent. Read the form of the qualifier before betting. But I think you get better results if you leave it as it is. The method works precisely because it catches those suprise hits you wouldn't normally have bet on, and it catches a good number of hits in races with big fields which pay better. And it works in steeples, hurdles, 2yo, country, metro, whatever - these races have the same statistical bias this system exploits. If you add more filters and only bet on the safe ones you tend to slash your average return because you miss the surprise winners.

The sample races I tried this on are the same races I've tested other selection methods on. This method often turns up totally different horses, ones missed by other more rational systems. I've trialled a couple of methods based on last start winners, but this one works best. This is a bit annoying actually. I put in hours and hours of stats, I apply logical rules in logical ways, I read up on form, I try to make intelligent, deliberate selections -- and this irrational selection method gives better results. Doh!

Actually, it is not quite irrational. Picking numbers out of a hat is irrational. There is at least some rational basis for this method: it exzploits a bias towards last start winners and lower saddlecloth numbers. Happily, this seems to follow a line of least resistance through the stats, including the market stats, so it just happens you come out ahead.

And the results are very even. It gives a consistent volume of hits, even if only low paying placegetters, that keeps you in the black rather than having long runs of outs that dip you in and out of the red. Some systems will pay in the long run provided you can sustain periods in the red. In this system you stay in the black and your profits gather slowly. Or that's what happened in the sample.

In any case, it works better than other lazy methods like backing favourites. Favourites at $1 to win, $4 to place over the same sample gives you a loss. The Zip Star horse (from the Sportsman) at $1 to win, $4 to place gives you a loss (and a couple of very nasty runs of outs).

Try it out. On paper.

I am aware of this though. Most, if not all, simple systems will *eventually* send you bust. It is just that some will do it quicker than others. If this system "works" that might just mean that it postpones bankrupotcy longer than similiar methods. The difference between a perceived "good" system and a bad system is the time it takes to kill you.

Time wounds all heels - Groucho Marx.

Hermes.

[hermes]

So today (Wednesday 3rd July) the selections are:


Wednesday 3rd July 2002

Selection method: Top weighted last start winner.

Betting: $1 to win, $4 to place.

Canterbury

Race 1

# 1 Baroness Britney.

Race 2

#3 Boon Moon

Race 3

4 Houlihan

Race 6

#4 Grand Raj

Race 7

#5 Isim


Moonee Valley

Race 3.

5 Quasi Stellar


Race 4

#1 Silver Birch

Race 5

#1 Gold Boom

Race 7

#3 Datari

Race 8

#2 Mystic Melody


Doomben

Race 1

#1 Akers

Race 5

#2 My True North

Race 6

#4 Grace's Roses


Murray Bridge

Race 4

#3 Telbon Lotto

Race 5

#1 Prince of Revelry


We'll see.

Hermes

[hermes]

A small thing:

Its not quite all races, every race. Some races don't have any last start winners. In that case, no bet. That's why there's some races missing in the selections above.

If you want to reduce the number of races to bet on I suggest selecting races at random. If you set parameters like "Fewer than 14 runners" you change the mix of selections. And the method depends absolutely upon getting a felicitous mix of the good, the bad and the ugly (with the good predominating) - see selections above. If you take any steps to change that mix, won't work.

In the long run though I suspect the success or failure of something like this will depend upon finding a mechanism to select fewer races that have the same mix of selections but targetting the better prices.

Cheers

Hermes

[michaelg]

Hi, Hermes.

A few years ago Equestrian Publishing sold a system similar to yours. In the package there were about half-a-dozen sub systems all based on the last start winner with the lowest TAB number, each sub system had its own rules. All the systems in their testing period (almost a year) showed a profit in both Win and Place wagering.

[Equine Investor]

hermes you got my 5 STAR special in that lot....Silver Birch.

Good Luck.

By the way, don't be put off by just random numbers with that system. The laws of random numbers say that you will find winners that others won't merely because you are oblivious to certain form factors.

[hermes]

Missed one, plus one small correction.

Murray Bridge.

Race 5 should be #2 Nafir and of course Prince of Revelry is running in race 6, not 5.

Thought I'd point out the correction before the race rather than after.

Good punting

Hermes

[hermes]

Thanks for that michaelg. You mean someone has already thought of packaging this up as a "system" to sell to lazy punters? Damn! I was thinking of tweaking it a bit to make it more complicated and mysterious, naming it something like "Hermes Miracle Punting Plan" and selling "Hermes Turbo Tips" to subscribers for $10 a week each. After which I could forget about punting and retire in Fiji. But you say its already been done. Oh well... :smile:

[Merriguy]

System doing great so far today. Hope this doesn't cruel things!

5 races so far --- two winners ($6.50, $8.70 NSW Tab), and one third ($1.90). Couple of scrachings. Do you then look for the next qualifier, or just forget that race. I guess you will say it doesn't really matter. Serendipity!!

Thanks for the system. While I appreciate it might fall down tomorrow, it does seem to have quite a few pluses going for it.
Personally I like to have a few bets --- for interest sake if nothing else. Couldn't be one of those who only have one bet a day; yet don't believe in backing the card everywhere there are races. Thanks again.

[Placegetter]

On 2002-07-03 14:28, Merriguy wrote:
Personally I like to have a few bets --- for interest sake if nothing else. Couldn't be one of those who only have one bet a day; yet don't believe in backing the card everywhere there are races.


Merriguy, you need two banks. One that will make you rich and one that will probably send you broke. If you ever get serious you will easily see why, otherwise, I hope you are at least having fun.

Hermes, my wife found the elusive filter you need in your system.

Lowest saddlecloth number from last start winners

PLUS

Cutest name.

Works every time she reckons.

Placegetter

[hermes]

Results:

Baroness Brittany - scratched.
Akers - unplaced
Boon Moon - $5.50 win, $2.60 place.
Houlihan - unplaced
Quasi Stellar - $9.20 win, $2.50 place. (Yeah!)
Silver Birch - $1.70 place (Equine Investor's five star beauty.)
Telbon Lotto - unplaced
My True North - third, NTD. (Blast!)
Gold Boom - scratched.
Nafir - $5.90 win, $1.90 place. (Lucky I double checked the selections or I'd missed this one.)
Grand Raj - unplaced.
Graces Roses - unplaced.
Prince of Revelry - unplaced.
Isim - unplaced.
Detari - $2.10 place.
Mystic Melody - $5.80 win, $2.00 place.

By my reckoning that's $70 outlay ($1 win, $4 place on 14 races) with a return of $77.60.

The system is $7.60 up after fourteen races.

If we'd added Placegetters wife's filter (see post above) we would have eliminated Akers and saved an extra five bucks.

Today's results - fortunately for me - illustrate what I see as all the best features of this hot little system. A good number of reasonably priced placegetters gets you by while you wait for a healthy share of winners and the occasional good one like Quasi Stellar, race 3 Moonee Valley, $9.20 the win, which I would never have selected in a million years. Who selected Quasi Stellar? And it isn't a one-off. This method gets winners like that often enough. Just check back through batches of old races and see. It surprises me.

I reckon you could expect one return like Quasi Stellar every one or two race days with this system. On a day to day basis you *should* be ahead two days out of three. And there is high fun value in this too.

And all with a selection system that is so simple George Bush junior could do it without having to hardly look up from his pretzels.

A couple of refinements:

*It wouldn't hurt to eliminate races less than 8 runners to avoid the dreaded NTD as in Doomben race 5, #2 My True North, today. It shouldn't upset the stats. On the other hand, you should be able to pick two horses in a field of seven. I don't think it matters, but NTDs really irk me.

*If you want to include a safety device, I suggest stop betting after four outs and only resume after a strike. The chances are that the strike you miss by doing so will be one of the small priced placegetters this system collects and missing it won't cost you as much as you saved on outs, and the chances are also that you'll collect the next winner after the one you missed because this system throws up its share of doubles (two strikes in a row). Maybe. I haven't tested this. Today you might have missed Quasi Stellar and picked up Silver Birch, or you might have missed Detari and picked up Mystic Melody, if these strikes had of been the resumes from the break.

In any case, in my sampling this system seems remarkably resilient against long runs of outs. It breaks up the runs of outs with lots of short priced placegetters. In my sample of 180 races there was one run of seven outs, two of six, none of five, three of four outs, three runs of three outs and twelve sets of two outs. That is pretty stable. A safety device wouldn't have saved you much. No doubt you'll hit runs of eight outs, nine, ten, etc. but again the sheer volume of placegetters this system gathers guards against it happening too often.

So I don't think spectacular and catastrophic runs of outs are the weakness in this system. More likely it will strangle you slowly with a slipping average return.

The good number of double strikes and triple strikes the system yields is promising for the apllication of various staking schemes that depend on doubles and triples. Then the system will either strangle you very rapidly, or it will show itself to be a true weakness in the fabric of horseracing that will get you rich with little effort. Which do you think it is?

And remember: Like all systems it is self-defeating. The more it succeeds the less it works. There is no solid ground. You push one side and the other side moves to compensate.

A good day. But I remember last Wednesday.

Hermes

[becareful]

Question - why not back for win only??? Sure you would have a longer time between "drinks" but based on your initial post ("no good for placebetting alone - margins too fine") and todays results you would be better off. For today $26.40 in divs from 14 races - so at $5 a race (ie. put your place pool in win) you would be up $62 for the day.

[hermes]

I was just being careful becareful.

It won't work for placegetters only but it will for win only - if it "works" at all - , except you then expose yourself to long dries between drinks, as you say. And if you take out the placegetters and play win only I suspect runs of outs could be severe. There is nothing to buffer them.

One of the features I find attractive here is that it nets lots of placegetters which breaks up a run of outs and saves you from cycles of boom/bust.

Betting win only would be more fun. More ups and downs. I'm a conservative soul. I'm looking for stability.

You seem to be having fun with the mini system.
Good luck.

[hermes]

Many punters in this forum, including myself, had a bad one last Sat. Oakfield Duke, Ex-files, Zed-Files, Timidity. And so on. This simple selection system scored though.

Thirty one eligible races. Outlay at $1 win, $4 place = $155. Return = $166.50.

At win only, return = $250.50. (Underlines becarefuls point!)

Check them out. The winners were Gullcatcher, Wyngrove, No Stops (at $15.60 to win), Mr Attorney (at $11.50), Zabenz, Judanazo, Zedimbi, Glenwest ($7.10), Zip Infatuation.

Going OK.

Hermes

[hermes]

I've extended my sample to get an idea of the long term prospects of this mini system. I don't have hard figures. I mainly did it by eye, but it seems to me that, in the long term:

*If you bet win only, you'll go bust pretty quickly. You hit long runs of outs punctuated by small paying winners. After a while there are just not enough good paying winners to off-set the outs.

*If you bet each way you'll go bust slowly. Again, after a while the volume of outs exceeds the returns provided by the low paying placegetters. It will grind you down.

So, in conclusion:

If you are a fun better looking for a good day out, this system is recommended. High fun value and on lots of days you'll do well pulling in some wild and wooly winners. It is much better than following favourites.

But do NOT adopt this system as an investment strategy or your long-term betting strategy. You'll lose. I haven't worked out the gorey details. Don't need to. You can see the graph trends. You'll lose.

Hermes

[Bhagwan]

One way to increase your profit but keeping the idea of win & place,
is to bet 1Win+3Place instead of the 1W+4P
This should increase you profit on turnover by 20 points, because the betting ratio has been changed from 20%W*80%P to 25%W*75%P

[hermes]

Thanks for the tip Bhagwan.

System started off well today but fell in a heap at the end.

Rose Hill R1 #1 Eastwest Success - Win $2.70, P. $1.30.
R2 #5 Prsently - 4.40/1.40
Flem R3 #4 Malu, 3.10 the place.
Sunshine R3 #10 Sequently, 4.40 the place.
Sunshine R4 #3 Jestica 8.40/2.00
Rosehill R5 #6 Bringing Joy, 1.40 the place.
Flem R5 #5 The Big Ask - 5.30/2.20 (A great win! Great race!)
Then a loooong run of outs.
Until Flem R8 #6 Freegold, 3.90 place.
Chelt.R7 #6 Court Hero, 4.20/1.80


Pulled in some goodies, but too many outs on the day. At each way $1, $4. Outlay $125. Return = $112.80. Betting win only = exactly even.

Collating results, and despite what I wrote earlier in this topic, win only looks the better way with this mix, doesn't it? But you get substantial runs of outs before it puts you ahead. Going through a few weeks results now, each way betting is poor value - just behind. Win only has us in front, pulling in some good wins. Long term prospects doubtful I reckon. What do you think Bhjagwan? Can I make anything of this?

THE GUNGADIN FACTOR

A refinement to the system. There must be a way to filter out the likes of Sunshine Coast R1 #11 Gungadin, a selection in this sysatem. I know we are looking for the occasional long shot, but Gungadin is too much. We can squeeze extra value out of this system by eliminating runners like Gungadin, a wasted bet, the longest of long shots.

So eliminate horses with the Gungadin factor:

*The bottom average prizewinners.

If your lowest numbered last start winner has the worst average prizewinnings in the race, forget it.

That will leave the happy mix of all-sorts and take care of Gungadin.

Hermes

[Merriguy]

Hermes

It seems that the answers to some of your queries are to be found, in seed form at least, in some of what you have said yourself.

1) In your first post you said that "You need the winners to make it profitable." Therefore the place component is obviously not pulling its weight (no matter how satisfying it may be to get a return from the TAB or whatever on a regular basis), and so you should stick to the win bets.

2) The 200 or so results that I presume you have (original 180 plus recent races), should give a reasonable idea of what should be the cut-off point about saddlecloth numbers. Are the few winners over, say, about number 8, or whatever, giving a reasonable return? I'd like to hear your answer to that one!!

3) And the next step is surely to find a progression that will take advantage of your very good strike rate of 30% winners. Perhapos the forum can help there.

All in all I think it has great prospects. Congratulations.

[hermes]

Thanks Merriguy. Good advice. Will do some further figures including looking for a saddlecloth cut off. Lots of # 5 and 6 overall, I've noticed, but not sure about win onlys. In fact, it seems I'll need to recalculate all of this on the win only figures. Becareful had it right several posts back. He wrote:

"Question - why not back for win only??? Sure you would have a longer time between "drinks" but based on your initial post ("no good
for placebetting alone - margins too fine") and todays results you would be better off. For today $26.40 in divs from 14 races - so at $5 a race (ie. put your place pool in win) you would be up $62 for the day."

You were right becareful. Win only is the way. It's doing better. A strike rate you could do something with.

I'II do some more figures. Wish I had a push-of-a-button database.

Hermes

[becareful]

Thanks for the positive comment Hermes! One idea you might like to look at to get rid of the "Gungadin" factor would be to eliminate horses over a certain price (maybe $25?). I don't know if this would work but it would be worth looking at when reviewing your data.

[Luckyboy]

Hermes,

'becareful' is on the right track with prices. You may want to look a lttle further into pre-race prices. It's a very good filter.

My own analysis over the last two years highlights that only 3% of winners have had a pre-race price (Friday paper) greater than $20.

Luckyboy

[Bhagwan]

Try this.
WIN BETTING ONLY
Approach No.1
Have a betting bank of $150 for every $1.00 win bet.Level stakes

Only bet on the critters showing $5.00+ in the paper.
Only bet on the Mules with TAB Nos. 1-10
(1-10 have a strike rate of 90%)

PLACE BETTING ONLY
Approach No.2
Try this staking plan from Eqestrean Investor
Bet sequence 1-2-6-18=$28 then stop & start again if you dont have a strike on that sequence.
You could have 6 banks of $28=$168

You said you have a place strike rate of 50%+
Test it on past results.

Leave out races with 7 & less runners.
Selections TAB No.1-10 only
No bet on selections paying less than $3.30 in the paper.
It should show a profit.

Then let us all know your findings.

[Bhagwan]

That Equestian Publishing system that was based on last start winners, was based on starting at the bottom of the list of horses & betting on the first one you come to paying $11.00 or less in the newspaper.

They claimed it made a profit ,after testing it over 350 races ,it was all fabrication.With very long runs of outs.
You would have lost a motsa with buckleys of recovering .

It makes far more sence to start at the top & work down.
With some qualifing rules.
E.G.1-10 TAB Nos. win 90% of races so why would you start at the bottom 10%. It dont add up,it just dont add up.

[hermes]

Bhagwan,

This device - start at the top and look down for the first LSW - has merit it seems. Much more than starting at the bottom. Now I just need to find the ways it will pay.

A numerical cut-off for a start. The results a cut off at #7 will snare the vast majority of both placegetters and winners. Or rather seven is where the money stops. You miss some biggies beyond #6 or #7 but it can't be profitable to chase them race after race. In terms of just the volume of strike numbers you could cut off at #3! That will snare the vast majority of them, but only the lower priced ones. It is worth going down to #6 or #7 to snare the better priced winners.

The distribution of placegetters is noteworthy. Very few outside of saddlecloths 1, 2 3. Over 90% of them have saddlecloths 1, 2 or 3. The distribution of winners is more diverse. No exact figure but I'd say if your lowest numbered last start winner is 1, 2 or 3 there is a solid statistical chance it will place.

And here's something: most placegetters ran second! Very few thirds. I hadn't noticed this before. A proportion of nearly 8/2. A great method of selecting your second placer in quinellas??

I know this is looking like it will pay on win only but I am worried about a statistical fact: fewer last start winners win races than you'd expect. But, more last start winners place than you'd expect. That's why I was chasing placegetters with this method.

Last start winners who win are therefore horses that win two in a row. The stats are against that. The stats say that horses that win a race are more likely to run a place next race than they are to win. Doesn't the stat go:

35% of winners are last start winners. BUT
only 25% of last start winners are next time winners.


In this case we've found a method of locating a high proportion of seconds. In my samples the win strike is healthy, but that defies the stats, doesn't it? In the long term I expect the stat that says last start winners are more likely to place than to win will prevail. Despite my sample.

Maybe we have a quinella system in the making. Something like:

Take the last start winner with lowest saddlecloth number. Selection 1. (Likely to run second).

Selections 2 and 3. Of the four horses with adjacent numbers, take the two with the best last start and/or highest prixewinnings (or placegetter percentage or whatever way you want to distinguish between them).

We do this because, as a general observation, the other placegetters *follow the last start winner* who places. That is, if a last start winner with saddlecloth #1 wins or places, then the other placegetters will have saddlecloth numbers not far away. Usually there is at least one other placegetter one or two numbers away above or below. It is less likely that you get a placegetting last start winner with a low saddlecloth and the other placegetters are 7 and 8 or even 5 or 6. More likely combinations like 1, 2, 3 or 1, 2, 4. or 2, 3 and 5. etc. The bias towards the lower numbers/higher weights.

On the other hand if your placegetting last start winner is in the higher numbers 6, 7, 8, 9 etc. Then so too will be the other placegetters. More often than not.

Of course you get some very diverse distributions, but I note this as a general pattern. Very often one of the horses adjacent to the placegetting last start winner will place too. I colour them in in the form guides with a green marker. You can see the blocks of green. Looked interesting so I did a quick count. Only about 3 in 10 deviate from the pattern. Will need to do more stats obviously.

But maybe when betting quinellas use the lowest numbered last start winner to find the centre of the action. Take it as your centre and look at the numbers around it. To put it another way:

You don't get many placegetting last start winners that stick out like a sore thumb. You find them within a nest of other placegetters. You should be able to find a quinella strategy to take advantage of this.

Anyone find ways to make this pay?

The quest for a viable system (and steady stats) goes on.

Hermes

[hermes]

Twenty-five eligible races on the card tommorow. There's some good bets among the selections. Some days good, some ordinary, some bad. On paper, could be a good one.

Two selections are eliminated. We'll save our money on Warwick Farm R3, #10 Absolute Lure, and even more so Ipswich R5 #9 Tryst. So that's twenty three bets.

For the record they are:

Warwick Farm R1 #2 King of Soul, R2 #3 Lucky Gwen, R4 #4 Dare Du Ciel, R5 #3 Exotikos, R6 #3 Don't Come Yet, R7 #7 Venere, R8 #1 Kurrajong Mist.

Sandown R1 #1 War Abandoned, R2 #1 Stolen Crown, R3 #1 Sacred Hill, R4 #6 Fine & Dapper, R5 #7 Zedati, R6 #5 Power in Motion, R7 #5 Century Gal, R8 #4 Our Target.

Ipswich R6 #4 Graces Roses, R7 #2 Norwhal, R8 #2 Hellie Missed.

Gawler R3 #2 Delete Me Not, R5 #2 Claredon Curse, R6 #1 Kurt, R7 #4 Dutton, R8 #1 Manilow.

We'll try it on paper win only - $5 on the nose. That's 23 x 5 = $115.
Also calculate as per Bhagwan's suggestion - $1 win, $3 place = $92.

And I'd like to try this. Numbers 1, 2, 3 only. $5 to place. That's 14 races = $70.
And let's see if it bends the other way: Numbers 4,5 and 6 only. $5 to win. That's seven races = $35.

Hermes

[hermes]

Some quinella combinations.

Try simple adjacent horses. The two adjacent to the lowest numbered last start winner. Eg. Warwick Farm R1. Selection is #2 King of Soul. So box 3 quinella becomes 1, 2, 3. Race 7 Warwick Farm. Selection is #7 Venere. Quinella becomes 6, 7, 8. Trying to catch the adjacent horse phenomenon.

A more complex possibility:

Take the lowest numbered last start winner. Then the lowest numbered last start second placegetter, then third. If no second, then third, then fourth. And so on.

Now, where this yields horses adjacent to the lowest numbered last start winner: box 3 quinella.

The selections tommorow become:

Warwick.

R1 - 1, 2, 4.
R2 - 1, 2, 3.
R6 - 1, 2, 3.

Sandown

R5 - 4, 6, 7.
R8 - 4, 5, 6.


Ipswich

R6 - 2, 3, 4.
R7 - 2, 3, 6.
R8 - 2,3, 6.

Gawler

R3 - 1, 2, 5.
R5 - 1, 2, 9.
R6 - 1, 2, 3.
R8 - 1, 2, 8.

An extension of the idea.

Hermes

[hermes]

Bhagwan wrote:

Then let us all know your findings.

Will try it that way too, Bhagwan. Many thanks for your insights. Will report the findings.

[hermes]

Two punters went to the races today armed with a wad of cash. One to Warwick Farm. One to Sandown. Both had a new selection strategy they heard from a mate who heard from a mate who read it on the Net. Simple system: select the lowest numbered last start winner.

The guy at Warwick farm had a beauty. Six placegetters in seven. Went home a believer.
The guy at Sandown crashed. Two places in eight races. Blamed the heavy track. Went home and kicked the dog.

(Another guy at Ispwich scored two places out of three and another guy at Gawler scored two out of four. Both went home happy but not convinced.)

Total today in twenty two races (one scratching from 23) = twelve placegetters. 54%. Few winners, but King of Soul at $5.80 (Vic TAB) was a nice start to the day.

Warwick Farm

R1 #2 King of Soul - $5.80/$2.60
R2 #3 Lucky Gwen $1.70
R5 #3 Exotikos - $2.60
R6 #3 Don't Come Yet - $4.40
R7 #7 Venere - $4.00
R8 - #1 Kurrajong Mist - $2.00

Sandown

R1 #1 War Abandoned - $1.60
R5 #7 Zedatzi - $1.90

Ipswich

R7 #2 Narwhal - #2.80
R8 #2 Hellie Missed - $1.80

Gawler

R5 #2 Claredon Curse - $1.90
R6 #1 Kurt - #3.90/$2.00


Have to check to see which staking stragegy worked best, but still enough action in this system to keep me interested.

Hermes

[Bhagwan]

The level stakes return for the place were sensational.

Return $29.30
Outlay $22.00
Profit $7.30 =33% POT

If you could do that on win betting you would be brilliant , but on place betting thats sesational, for the day that is .

You were saying its level stakes profit for the place was ordinary, what would be interesting, is to find out the profit possability at level stakes on the past 150+ results with the various filters added.
That had previously mentioned.
E.G. TAB 1,2,3 OR TAB 1-7 etc.

[hermes]

Isolating the TABs 1,2,3.

122 races. 67 placegetters. 54%.

But we're collecting too many low payers.

Average return on that sample = $1.87.

$1.87 at 54% won't go.

Raw figures only and so no money filter.

Bhagwan, the immutability of your morning price $3.30 rule has dawned on me.

Hermes

[hermes]

Bhagwan and co.

Still working on these figures because they bother me. Reviewed the whole thing. What has happened is.

*Did an initial test of the idea on 60 random races, all races, any races extracted from the Sportsman, May, June races. Wow! An amazing resut. Stacks of winners. Even more placegetters. And paying! A $17.80 winner, for instance.
*Expanded this to 180 races. Same time period. Still great guns. Great strikes.
*Tested it out on a batch of forthcoming races. Did really well.

The trouble was my sample of 180+ included several spectacularly good race days, a couple of really good ones, several bad ones but no shockers.

*Tested it on another batch of forthcoming races. Ordinary.
*Did a bigger sample. For this I've had to turn to old copies of the Herald Sun, whatever I can get. Batches from January, Feb, March, April. Some from 2001. All races, any races.

The more I add the more the overall pattern emerges. You strike bad periods. A couple of race days in a row in April, shocking. Not just bad, but way off. One out of eight at Flemington. None out of five, Victoria Park. Then you hit good days and then amazingly good days. But in my initial 180 I didn't strike any of the shockers the larger sample revealed. The pitfalls of a small sample. If you'd used this mini-system on all races any races in my first sample period, May, June, you'd have struck some great days with a healthy strike rate. Some days stacks of winners. Other days, like today, a swag of placegetters. But the larger samples show this is not reliable. The first sample was even, but its not over time. Can do very, very well or very, very badly. (Definition of a "fun" system I suppose.) But over time the bad patches win. Which is what I expected from the outset really.

But there are spectacular days!

The bigger the sample - trying different batches since they're so uneven - and the closer I look the less enthused I am, but then you get days like today that keep me thinking it is worth pursuing. Beware the bad days!

Hermes

[croc]

Another angle. May I suggest that when doing your research, looking at any last start winners in Melbourne that started exactly seven days ago will probably increase your profits. Don't waste your time on any other venues or days to last start.

Ciao

[hermes]

Thanks for the imput on this little system. I now have so many possible lines of investigation I really need a database to work from, so I'm building one. Lowest numbered last start winners. I'II feed in all my samples and be able to do some proper analysis.

Meanwhile, tommorow's card has lots of qualifiers and some strong horses among them. Wednesday was a great day for placegetters. Looking at the card, I'm expecting a strong show from this system tommorow. (But beware the bad days!)

Failing another filter that will push this system into the black, I'II try just plain old judgement. Looking mainly at average prize and average place percent in the context of the other runners in each race. I think the stragegy tommorow should be...

RANDWICK


R1. #3 Mr Attorney - each way.
R2 #1 Honey Ryder - each way.
R3 #9 Ms Bowie - no bet.
R4. #6 Mr Platinum - each way
R5 No qualifers.
R6 #1 Inclusion - place.
R7 #1 Go Ziggy - place.
R8 #5 Freiby - place

MOONEE VALLEY

R1 #1 Spanish Symbol - each way.
R2 #1 St steven - each way.
R3 #4 Midday Matinee - no bet.
R4 #1 Be My Princess - each way
R5 #1 Shaye Spice - place only
R6 #7 Trust Fund - no bet.
R7 #1 Sly Rambler - place.
R8 #8 Runs on Ego - no bet.

EAGLE FARM

R1 #7 Kimjed - no bet
R2 #3 Devil - each way
R3 #8 In our Time - no bet
R4 #9 La Philomene - no bet
R5 #1 General Minolta - place
R6 #1 Pittance - each way
R7 #5 Final Shuffle - no bet
R8 #1 Pitterac - no bet

Morphetville

R1 #1 Judanzo - each way
R2 #3 Bellton - each way
R3 #3 Nafir - place
R4 #1 Romalada - each way
R5 #1 Risky Lass - each way
R6 #4 Sure Bet - no bet.
R7 #1 Zip Infatuation - place
R8 No qualifiers.


Hermes

[Merriguy]

Hi Hermes.

You seem to have missed Satashi, ahead of Runs on Ego, MR8.

Good luck today.

[hermes]

As a coincidence I was just looking at that Merriguy. Lowest numbered last start winner in that race is not Runs on Ego but Satashi. Satashi the selection.

Good luck to you today Merriguy, and to all on this forum. Some good races today.

Hermes

[hermes]

Another good day for placegetting lowest numbered last start winners. In terms of placegetters I made 12 out of 19 correct calls and also made lots of correct no bet calls.

Clearly there are plenty of placegetters in the pool of lowest Tabbed LSW. Apply some discrimination and you'll get more than you lose. And some reasonable ones too. Today:

Mr Attorney - $2.70 place. ($5.90 win)
Sly Rambler - $3.20 place ($10.30 the win)
Bellton - $2.80 pl
Nefir - $3.20 pl
Romilada - $2.30 pl
Kimjed - $2.40 pl
Final Shuffle - $3.20 (14.40 the win)

Lowest return of the day was St Steven - $1.20 the place.

Another day showing how this system turns up some decent winners too, but I have no luck finding a device for picking them from the raw pool. Be nice to add winners like Final Shuffle to a system that pulls placegetters like Nafir.

I actually placed some bets on some of these today and came out way ahead. Betting on places only. $62 outlay for $104.60 return. There's potential in this.

(And Geelong beat Collingwood!)

:smile: :smile:

Hermes

[Bhagwan]

Try this with the last start winners.

Only back the runners that have also ran 1,2 or 3 in its 2nd. or 3rd. or both 2nd.& 3rd. last starts.

The strike rate will increase but the average div. will drop but I feel the POT will be stronger.

Check it out.

[hermes]

Thanks Bhagwan. Yes, I have been looking at many factors including second last starts for ways to separate the chaff from the wheat. Doing some calculations to see if your suggestion increases POT.

My current experimental set of filters is working fine on past races so here's tommorows prognosis:

Tommorow's qualifers (wednesday 17th July): lowest tabbed last start winners are listed below.

I have rated their chances of running a place on a scale of 0-4. Zero = no bet. You could bet 2 units on a 2, 3 units on a 3 etc. Or ignore the ratings. No horses rated 4 tommorow but several no bets.

CHELTENHAM

R2 #1 Diver Dave - 1
R4#1 San Sonata - 2
R5 #1 Miss Revic - 1
R7 #9 Mr Vandaam - no bet
R8 #1 Blue Bows - no bet.

GRAFTON

R1 #2 Point Guard - 2
R4 #1 Casual Story - 2
R5 #1 Miss Smugg - 1
R6#1 Sir Redford - 1
R7#2 Stormcat Academy - 3
R8#7 Nikolinis - 2

RANDWICK (Kensington)

R1#1 Al Megdam - 1
R3#6 Covet Thee - 1
R4#3 Acceptive - 1
R4#2 Azzeal - 1
R6#1 Rain Statesman - 1
R7#5 Perry Can Do - no bet
R8#7 Painter's Brush - 1


FLEMINGTON

R1#9 Liston - no bet
R3#3 Lightning Ridge - 1
R4#4 Jacque - 1
R5 #1 Intermagic - no bet
R6#6 Mystic Melody - 1
R8#3 Medori Gift - 2

EAGLE FARM

R4#2 Duel Fuel - 2
R5#6 Interior Trim - no bet
R6#10 Ditty Doo - no bet

Some healthy place strike rates in this lot. Note Nikolonis - 72.7% place average from 11 starts. Casual Story - 71.4% from 7 starts. Sir Redford - 61.1% from 18. Painter's Brush - 66.6% from 9 starts. etc.

Despite the rating I don't actually like Stormcat's chances at Grafton race 7. A very competitive field. A stack of last start winners. Of them I prefer Caissa #5 to win. But we'll leave Stormat the selection, rated 3 to run a place.They say it will overcome the barrier. Not enough, I think.

Go punting to all

Hermes

[TheDuck]

Hi all,

I'm thinking about comments I've read about whether you do win or place wagers. It seems to me this isn't the issue. For example, if this system hits 33% of its guesses then all the 2-1 wins will end up losing money.

I don't know if you've considered this already but it would help if there is an average win percentage. Couldn't you use this as part of your filter?

For example, if this system hits 33% of the time then skip anything less than 3-1. 25% filters to only 4-1. Simple math.

Also, you mention there are bad times and good times. What's different about the bad times? I realize that sounds like and INCREDIBLY naive question, but the pattern is in there somewhere. I've found a neat trick is to figure out how to MISS the picks. As Sherlock Holmes would say, when you eliminate everything else, whatever is left, no matter how incredible, is the truth! Of course, his life was dictated by a writer who could twist facts at will but, hey, what can you do?

This is a great thread! Thanks!

[Equine Investor]

On 2002-07-17 11:40, TheDuck wrote:

For example, if this system hits 33% of the time then skip anything less than 3-1. 25% filters to only 4-1. Simple math.

This is a great thread! Thanks!


TheDuck, I have found that trimming down selections to only include the longer prices, in my systems expands the run of outs, so it actually ends up evening out the same. For instance, if you eliminate anything under 3/1 then the strike rate decreases significantly. You might like to check it out yourself with some of your selection methods and let us know if this is true for you.

[hermes]

A bad day for lowest tabbed last start winners. Off to a reasonably good start with a win, Al Megdam, followed by two placegetters at Randwick. Casual Story at Grafton. But Flemington - wipe out. Worse as the day went on.

If you ventured to Cheltenham, though, you might have picked up Diver Dave in R2 ($4.50 the win, $2 the place) and San Sonata in R4, $3.30 win, $1.40 p. And a second, Miss Revic, $1.80. which might have ameliorated some of the damage at other venues.

Vis-a-vis last Sat. I'm personally $1.10 up. Outlayed $34 on this today. Return = $14.80. Miserable. Saturday and today illustrate the good and bad of this method.

Back to the filters, although no filters would have helped today. Yes, EI I'm finding a price filter tend to impact too much as you say.

A days-since-last-start filter has definite merit. The idea is to snare horses at peak or just dropping off. How long ago they won that last start is obviously relevant, and over past races I am finding it so. A last start filter of 18 or so days will certainly help tip the scales towards profit and eliminate bad bets.

Always happy for contributions to this thread...

Hermes

[TheDuck]

Hey Hermes,

How's that database coming along? I thought it might be good for all of us to continue our research from the same place. If I can help I will. These computer thingies are my job so I have some facility with the annoying things. :smile:

[hermes]

Been very busy. Will talk about the database soon the Duck. Coming along.

Today:

Some strong qualifers in the lowest tabbed last start winners. With some judious betting you should be able to squeeze a profit from placegetters, but I'd expect a couple of winners from this lot too. A few favourites there. Again, my ratings from zero to 4 on their prospects to place. Zero = no bet.

ROSEHILL

R1 - #2 Thorn Park - 1
R2 - #1 Cariboo - 3
R4 -#4 Coablo - 1
R5 # 2 Interbank - 1
R6 - #11 Eartear - 1
R8 - #2 Dashiki - 3 (scratched)

CAULFIELD

R1 - #1 Chowpie - 2
R2 - #1 Romantic Sea - 2 (a question about barrier)
R3 - #1 Palais - 1 (a question about barrier)
R4 - #4 Quatyman - no bet
R5 - #2 Rubitano - 3
R6 - #2 Mr Gold Flyer - 2
R8 - #1 Out of Options - 2

DOOMBEN

R1 - #3 Centona - no bet
R2 - #2 Kymessa - 1 (scratched)
R3 - #2 Craiglea Bells - 1
R4 - #2 All That Glitter - 1 (a question about barrier)
R5 - #3 Linea - 2
R6 - #7 Gonzo - 1
R7 - #2 Jillina Blue - 2
R8 - #9 Icy Grip - no bet

CHELTENHAM

R1 - #4 Just Vic - no bet
R2 - #2 Terminus Prince - 2
R4 - #5 Kastabon - 1
R6 - #3 Shocks - 2
R7 - #1 To Pay Up - no bet
R8 - #6 Addition - no bet

[Bhagwan]

Try This

The first last start winner you come to between TAB 1-7
that is also starting between Barriers 1-7

Most races have 14 & less runners ,we are covering at least half of the inside field in the majority of cases.
We are assuming the out side half has a less of a chance of getting up.

Stats show Barriers 1-7 win 55.2%
1-8 61%
1-9 66.7%
1-10 71.7%
1-11 76.9%
1-12 83.2%
1-13 88.5%

Go over your past results & see if barrier has any impact on the overall figure.

[thevig]

Bhagwan: There are lies,damned lies and statistics. Most winners come from barriers 1-10 say because there are fewer races with larger fields. The influence of the barrier depends on the length of the race and the shape of the course. Barrier 14 in a field of 14 over 1400m at Caufield is very different from Barrier 14 in field of 14 over 1200m or 2400m at Flemington. This is one of the disadvantages of using a mechanical system. The appropriate statistic is the strike rate (%wins from starts)for a particular barrier over a given distance at a particular course. Of course with luck in running anything is possible. Remember Gurner's Lane?

[TheDuck]

Good point about statistics. There was one where it was proven with statistics that an increase in churches resulted in more crime. The real point was there was an increase in population that resulted in more churches and crime. Percentage-wise, everything was the same.

I am curious about this database as I would love to run a statistical method called 'factor analysis' over it. What this does is statistically determines which factors appear to most affect the variable(s) being analyzed. Since this is a (more) mechanical approach then we'll let the mechanics tell us what we should be looking at -- from a mathematical standpoint anyway. We then switch that around and find out what ranges and combinations of factors (clusters) present the best opportunity.

It's more involved than that but the poor suckers (ok, like myself) who are into this will understand the approach. We could then test it and see if it actually works.

Who said math isn't fun (besides my kids!)?

-Duck

[hermes]

The hold up with the database The Duck is my learning curve. You may be distressed to know I'm working on a Mac. A G3. The software I'm using is Filemaker Pro. On most things Mac you can now get the standard MS software, eg. MS Exel for spreadsheets etc. but I'm told Filemaker Pro is the native Mac database builder. Not too hard to learn but then I'm making it quite complex. This is the third attempt at starting from scratch. Think I've got it right this time. Just a mattrer of adding more data. It will be flash when its done.

Any assistance or advice appreciated. I'm very much a novice with this sort of software. Usually work with images, not numbers and code.

Hermes

[hermes]

WEDNESDAY SELECTIONS

Last Wednesdays races were bad for this system. On paper tommorow's won't be much chop either. Not a strong group of runners in this category for Wednesday 24th July 2002. I anticipate a wipe out at Sandown with big fields and little talent among the lowest tabbed last start winners. So I've tightened the parameters of my ratings, especially for runners in large fields. (Will probably cut out some good returns in doing so.)

Both Pro Gold and Hulavie are borderline selections at Sandown. Won't be surprised if neither run a place, or if we get nothing from Sandown at all. Only La Banca in R8 at is a confident bet.

By my ratings Clarendon's Curse is an outstanding prospect to run a place. If either La Banca or Clarendon's Curse is at a good price, worth a win bet.

CANTERBURY

R1 #2 Vital Agreement - 1
R2 #1 Patch Adams - 1
R4 #1 Secret Gift - 2
R5 #3 Midnight Sports - 1
R6 #1 Melissa Mel - no bet
R7 #5 No Integrity - no bet

SANDOWN

R1 #5 Miss Megari - no bet
R2 #1 C'este Le Reve - 1
R3 #3 Pro Gold - 1
R4 #1 Hulavie - 1
R5 #8 Meet the Stars - no bet
R6 #4 Allez Jane - no bet
R7 #4 Prime Target - no bet
R8 #2 La Banca - 3

DOOMBEN

R3 #2 Scooters - 1
R6 #4 Disco Girl - no bet
R7 #2 Just Devine - 2
R8 #7 Rubys Jester - 1


BALAKLAVA

R2 #7 Upmarket Star - 1
R6 #4 So Late - 2
R8 #2 Claredons Curse - 5

----------------------------------------------
TOTAL = 23 units

Over three race days now this system - filtered lowest tabbed last start winners - is a whisker ahead. Outlay $97, return $98.50.

[TheDuck]

Ah, a Mac. Yes, excellent and popular graphics platform. However, I'm a number cruncher and use 'the other kind'.

Send whatever whenever in any old format and I'll figure it out. I can even send it back to you in whatever format you're looking for.

But there's no rush. I was just offering to help with the crunching and analysis. I was also curious to see how similar or different Australia racing is from Canuck and Yankee racing. It's interesting when the DRF has horses from the other side of the planet. Some sleepers clean up, some really strong entries have to be dragged across the finish line by their tails! It seems your ponies are just as finicky as ours.

-Duck

[Equine Investor]

TheDuck, how healthy is "Yankee and Canuck" racing at the moment? In Australia it is really flourishing with prizemoney being increased each year. The only problem is attendance figures on non feature race days has fallen away.

I was wondering if it is the same overseas, as I have heard that racing is not doing so well in the States, but doing well in the U.K. and Pacific Rim.

[Bhagwan]

Try This

Only back last start winners TAB 1-7
With a 25% career win strike rate.

Test it over past 150+ races & see if this has any overall positive impact on the stats.

[TheDuck]

Hi EI,

I must say that U.S. tracks in general aren't looking very sharp with a few traditional ones still holding their ground.

The one near Toronto that I go to is doing quite well, however (Woodbine). It is a first class facility and has recognized that it's a business that needs to attract customers. They have slot machines now which brings in a few more stragglers and helps with the upkeep. I can't believe how full the parking lot gets on a Wednesday compared to 10 years ago when those areas weren't even parking lots!

Those institutions who recognize they are there for entertainment based in tradition -- I think -- are stronger than ever. If you were at the track at Woodbine you would be impressed. Everbody (including the many that come as whole families) is having a good time with very little 'scared' expressions revealing that the punter has gone beyond his means. We also have a group of Jamaican regulars that are absolutely hysterical.

So in summary, yes it has fallen off a bit. But I think that has made the others stronger. When our caller comes out in his pristine red outfit, lifts the higly polished horn, and plays a bit of 'Sesame Street' before the call to post you can see the crowd just loves it.

So how about Australia? And what's this about dog racing? You don't see much (if any?) of that in Canada.

-Duck

[Equine Investor]

TheDuck, sounds like I'd love Woodbine.
Dirt track?

The trumpeter story made me laugh. We do have family days where the kids have face painting, pony rides, clowns etc. Basically though I would say that racing is more punter orientated.

Yes, we have Greyhound racing in every State of Australia. There are 8 greyhounds in each race (on average). They race over 340 metres to 728 metres, the most common distance being 511 metres, and chase a fake bunny (lure).
Sometimes it's hilarious when the electronics slow down the lure and the Greyhounds catch it. They call it a "no race." There's even a few hurdle races for the dogs and that's a great spectacle!


<font size=-1>[ This Message was edited by: Equine Investor on 2002-07-25 02:36 ]</font>

[thevig]

Hi Quacker: Ziemba, a professor of economics and mathematical guru spent a summer vacation trying to apply factor analysis to racetrack data in attempting to find the winner. Found that the more complex type of analysis did not work as well as simpler model with fewer variables. As far as I know he discontinued using factor analysis. Has written about another method he investigated in "Beat the Racetrack".

[Placegetter]

On 2002-07-25 02:33, Equine Investor wrote:
Sometimes it's hilarious when the electronics slow down the lure and the Greyhounds catch it. They call it a "no race."


Hi guys,

Just dropped in for old time's sake and was immediately reminded of this story.

When I lived in Darwin I decided to go to the dogs for the very first time (and only time so far). It was Darwin Cup, biggest race of the year ($10,000!!).

There was no secret about the favourite, it was backed into about $1.40 from memory. Come race time, the lure broke down and the favourite, which was leading by about 6 lengths, caught the lure. The look of utter disappointment on the dogs face was priceless when it realised it had been conned for it's whole career. Scarred for life I'd say!

Good luck!

Placegetter

[Equine Investor]

Placegetter,

Great to see you popping in to say hi.

Yes, catching the lure can actually spoil a greyhound for life, they are not as silly as one might think. They do learn that it isn't real if they catch it and sometimes remember. Some dogs have to be retrained again...in the bush (if you get my drift) to get back the impulse to chase!

[hermes]

Good to hear from you placegetter!----

Wednesday was an ordinary day. After scartchings, $21 on for $16.30 return. Running total now: outlay of $118 for a return of $114.80.

A full swag for tommorow.

Randwick

R1 #3 Inclusion = 3
R2 #1 Al Megdam = 2
R3 #2 Dashiki = 1
R4 #4 Play Mistress = 1
R5 #3 Conzeal = 1
R7 #6 Shantey = no bet
R8 #6 Just Imagine = no bet

Moonee Valley

R1 #2 Tycoon Ruler = 1
R2 #5 Dornach = 1
R3 #1 Steel King = 2
R4 #7 Sterling Knight = no bet
R5 #1 Bush Padre = 1
R6 #2 St Steven = 3
R7 #4 Green Pick = 2
R8 #2 My Lavina = 1

Eagle Farm

R1 #2 British Buska = no bet
R3 #1 Dancing On = 3
R4 #2 Sheer Devotion = no bet (barrier)
R6 #1 Smart Chariot = 2
R7 #1 Kenconcarne = 1
R8 #4 Bishop Bill = no bet (barrier)

Cheltenham

R1 #4 Social Glow = 3
R2 #1 Frenzel Rhomb = no bet
R3 #4 Shoppaholic = 1
R5 #2 Glenwest = no bet
R6 #2 Ikaros = 3
R7 #5 Onyx River = no bet
R8 #3 Lawful Poker = 2

[hermes]

Lots of action today and some good winners and placegetters again. Lowest tabbed last start winners were in the thick of it at all tracks bar Eagle Farm, one out of three. After scratchings, seventeen selections, eleven ran a place. Betting level stakes for places, a return of $22.50 for $17 outlay. Tycoon Ruler was the best result for the day.

**********************

I've watched this category of runners closely the last few weeks matching the live tests to past results. Live tests make past results much more useful. I'm happy that my ratings filter succesfully removes the dross. These are simple ratings based on the commonly considered factors like average prizewinnings, place percentage, days since last start. Nothing mysterious. If nothing else, they identify the no bets only missing the occasional strike.

With this no bet filter applied to the lowest tabbed last start winners the overall results on level stakes are quite OK for placebetting over a fair sample. I have full prices on 360 races:

360 races
192 place strikes = 53.3%
Return on places = $372
Average return = $1.94
POT = 3.33% for places

For win bets it shows just under break even:

76 win strikes = 21%
Return on wins = $357.60

This is not very exciting - except that getting anything to show POT on places over 360 races is a victory for me - BUT my rating system has revealed a way to improve on this dramatically. It happens that the runners I rate as 2, 3 or 4 - the better prospects to place - have a good strike rate but a very poor average return that is draining potential profit. Thus:

Rated 2, 3 or 4.

160 races
92 place strikes
40 win strikes
Return on places = $153.60
Average Return = $1.67
Return on wins = $120.40

This is why I've been having trouble squeezing out a profit from a good strike rate. In this sub-group we have an excellent strike rate but a loss on turnover. Too many lowest tabbed last start winners pay too little. The problem with this overall category of runners I realise is that it includes too many favourites - there is too much poor value in the category as a whole.

But the figures for those runners I rate as marginal - the rating 1s - show very good results in terms of value and here I think I have at last pinpointed the runners that attracted me to lowest numbered LSWs in the first place. I've isolated the action. Look at these figures:

Rated 1

200 races
100 place strikes = 50%
36 win strikes = 18%
Return on places = $218.40
Average place return = $2.10
POT on places = 9.20%

Return on wins = $237.20
Average win return = $6.58

Here are the value runners in this category! The marginal prospects - as opposed to the more certain bets - have about a 50% place strike rate, and a low win strike rate, but give healthy average returns and a good POT. An extraordinary average win return. I had to check it again. Comes to $6.58. Mostly $7 plus winners. Tycoon Ruler was today's example at $11. Eartear on 20th July = $7.80. Jestica - $8.40. Sly Rambler - $10.30. Final Shuffle - $14.40. Not many of them - 18% strike for wins in this sub-group - but this band pays well when they come home. And they come home often enough to give a return. Over 200 recent races.

So to extract value from this group you have to eliminate the dross, ignore the sure things and exploit the middle band. I know this is what various people on this forum have been telling me in different ways, but now I've worked it out for myself.

Who are the rated 1 runners? Conzeal was a good example today. Rated down because its last start win was more than 14 days ago and was not among the top three prizewinners, amongst other things against it. But not dross. Started fourth favourite. Paid $4 for the place. Tycoon Ruler is a perfect example too. St. Steven was an example of rated 2,3 or 4. Nothing against it on paper. A dead cert but not worth it. Shantey R7 Randwick today is an example of a horse with too much against it to be worth the bet. Started about 7th or 8th favourite, 26/1. No bet.

You can obviously achieve much the same filtering by observing the market. The rated 1 runners are usually your lowest tabbed LSWs starting at third or fourth favourite, or if second or first then usually at a good price.

But I'm happy that my ratings work without recourse to the market. Except of course that my ratings are exactly wrong. Don't bet by my ratings. Bet the opposite. If its only rated 1, back it!

In conclusion:

On further analysis, there are three types of runner in the category lowest tabbed last start winner and success or failure depends upon distinguishing between them.

1. Dross. Not likely to place at all. The category includes quite a few cases where the fact a horse is the lowest numbered LSW is meaningless in the context of the race - the Gungadin factor. They have very little chance. Often TAB numbers higher than 5. (Of course, sometimes these will win. Oh well.)

2. Marginal prospects. A fifty/fifty chance of placing and much less of winning but at a good price. Not standout selections and often borderline with the dross category. Or often strong on paper but with a black mark like poor prizewinning average. Not usually TABS 1, 2 or 3.

3. Heavily backed good prospects. A strong chance of placing and a good chance of winning, but at no value. There is a high proportion of heavily backed favourites among the lowest tabbed LSWs. Everyone likes a winner. This drains the value out of the category. Often TABs 1, 2, 3.

Might not be news to experienced punters, but its all news to me.

Maybe there are these three types of runners in whatever category of horse we look at . Anyway target the middle band. The shortcoming of the middle band will be more and longer runs of outs, not broken up by the sure things, and long waits between drinks for winners. But it should give POT.

In any case, should my figures once again level out over larger samples as promising figures have a tendency to do, I'm quietly confident that if I concentrate on those I rank as the marginal prospects it is near enough to profit at level stakes to warrant a staking plan. Lowest tabbed LSWs are a rich category of runner. Intelligent selection can isolate the band of value. Then intelligent staking can turn over profit. Open to suggestions as always.

Bhagwan, you wrote many posts back of betting 1,2,6,18. With six banks of $28? Can you explain further?

Hermes

[Bhagwan]

That 1-2-6-18 places staking plan came from E.I.
1st. bet 1
2nd. bet 2
3rd. bet 6
4th. bet 18
Total 28
You stop, then start again as soon as you strike a place getter.
You should have 6 banks of 28=168 .

You will win on some runs & loose on others. But that does`nt matter, its the overall profit we are after , this is called sequence betting.

Have you thought about ignoring TAB No.1 all together & starting at TAB No.2 & down.
Quite often No.1 is at low value, if ever they get up .Its rare that they pay well therefore poor value & best ignored all together.
You will probably find that you have missed some very good payers because your selection was No.1 in that race & the other got up instead.

[becareful]

Hermes,

Congratulations on sticking with it and coming up with some useful filters. You seem to have come to a similar conclusion as me (although by a different route and using different selection criteria).

There is rarely value in the short priced horses simply because they are too obvious and everyone can see they are a good chance of winning - they do have a good strike rate but the poor dividends are not enough to make a profit on. The true longshots ($20 and up) generally deserve their status and rarely manage to win - you can occassionally get some great priced winners from this group but it is generally not worth spending time analysing them all. The real value is in the mid-ground - those horses that have the class to win the race but are out of favour with the punters for one reason or another. If you can successfully pick your winners from this group then you have a much better chance of developing a profitable system.

Anyway keep up the good work and best of luck.

[TheDuck]

Still laughing over the dog racing stories! I had a dog that chased birds for 11 years then finally caught one. Just about scared him half to death.

To answer EI, Woodbine is dirt and turf. It's claim to fame is it's the only track in North America that runs thoroughbred and harness (standardbred) racing in the same day.

To respond to thevig, thanks for the lead. I only expect to use factor analysis to determine useful factors, not continue with that each race. I also want to do an analysis on 'soft facts' such as comments. One of my favourite books is Marketing Analysis: Methodological Foundations (a big, heavy statistics book -- isn't that sad?). It has some great stuff I have used successfully in other projects (this stuff is my day job too, isn't that even MORE sad?).

Then I expect to do some clustering to create deterministic functions that I can turn into rules -- or whatever. And if it all doesn't work I'll try something else. I enjoy the journey as much as the outcome.

If you come across anything else like this I would certainly look forward to those leads as well. Thanks!

Duck

[Equine Investor]

On 2002-07-28 09:21, becareful wrote:
There is rarely value in the short priced horses simply because they are too obvious and everyone can see they are a good chance of winning - they do have a good strike rate but the poor dividends are not enough to make a profit on. The true longshots ($20 and up) generally deserve their status and rarely manage to win - you can occassionally get some great priced winners from this group but it is generally not worth spending time analysing them all. The real value is in the mid-ground - those horses that have the class to win the race but are out of favour with the punters for one reason or another. If you can successfully pick your winners from this group then you have a much better chance of developing a profitable system.

Anyway keep up the good work and best of luck.




Absolutely cannot agree MORE!

Sometimes your selection may be the favourite or the outsider, but the guts of then should be somewhat midrange to ensure your not chasing your tail on shortpriced losers. The correct mix of winners is very important...eg strike rate / average dividend.

[hermes]

Some solid results in this category last Sat. Master Pom $5.80 the win, Pedro Girl $5.10, Sammuka $5.60. And the usual spray of placegetters.

Tommorow, Wednesday, 7th Aug. among the lowest tabbed last start winners we have (with placegetter ratings on scale zero to four):

Canterbury

R4 #4 Delightful Success - 1
R5 #7 Ramadee - 0 - no bet
R6 #6 Back in Style - 1
R8 #2 Covina - 3
R9 #1 Blazing Arch - 1

Sandown

R1 #1 Living End - 3
R2 #2 Tarcoola Diamond - 1
R3 #1 Step Ahead - 2
R4 #2 My Lavinia - 2
R5 #4 Smoking Barrel - 1
R6 #11 Vocals - 0 - no bet
R7 #2 Skewiff - 3
R8 #1 Piermont - 4

Doomben

R1 #3 I'm Fighting - 1
R4 #1 Victory Ranger - 4
R5 #2 Carissma - 2
R6 #3 Foxmore - 3
R7 #2 Nattie - 2
R8 #5 Huey - 1

Gawler

R5 #1 So Say All of Us - 4
R6 #2 San Sonata - 4
R7 #3 Cullen Bay - 2

Skip Ramadee and Vocals.

I anticipate quite a few placegetters in this lot but probably not much value. Should be quite a few faves and second faves among them. Too many.

Bet level stakes to place on all but the zeros or on rated 1 and 2 only, or as you wish. Rated 4s have good strike rate but poor returns. Rated 1s a lower strike but for a better proportion of return.

In another bank bet each way $1/$4 on the rated 1s to set and forget or better still on those showing $5+ with a few minutes to jump.

Happy punting

Hermes

[hermes]

A typical performance today. After scratchings $18 outlay at level stakes, $15.30 return. But again the value was in my rated 1s: outlay at $1 win, $4 place on the rated 1s = $35, return = $42.90. As per my research, if you concentrate on the middle band of runners in the category, you'll win. Today, three winners and a placegetter out of seven. Today's best: Back In Style: $7.70 win. $3.50 place.

Hope you're ahead

Hermes

[hermes]

1. Take the lowest numbered last start winner, all races, any races.
2. Eliminate any rank outsiders but leave long shots with even a slim chance.
3. Eliminate all TAB number 1's.
4. Bet level stakes to win.

Strike rate 21%
POT = 6%

Places gets to break even.

Better still, eliminate all TAB numbers 1, 2, 3. Lower strike rate, less action, better POT. Also shows small but steady POT on places.

Or, bet to win by TAB number, so $4 for TAB 4 etc. Assumes that the lower the TAB the better the return, on average. (Actually i added the wrong columns in the spread sheet to find....) Yields 13% POT.

But everything depends on rule 2. distinguishing between a no-hoper and a slim chance.

Hermes

[hermes]

Otherwise:

Take all the lowest tabbed last start winners. Bet level stakes to win on any showing between $5 and $18 to win a minute to jump. You'll come out ahead. Works on place bets also but POT is much better on wins.

Hermes

[BlueTyson]

On 2002-08-12 01:26, hermes wrote:

1. Take the lowest numbered last start winner, all races, any races.
2. Eliminate any rank outsiders but leave long shots with even a slim chance.
3. Eliminate all TAB number 1's.
4. Bet level stakes to win.

Strike rate 21%
POT = 6%

Places gets to break even.

-- Very interesting thread Hermes, thanks. So this is over several hundred races? Rank outsider = 20+ or worse?


Better still, eliminate all TAB numbers 1, 2, 3. Lower strike rate, less action, better POT. Also shows small but steady POT on places.

Or, bet to win by TAB number, so $4 for TAB 4 etc. Assumes that the lower the TAB the better the return, on average. (Actually i added the wrong columns in the spread sheet to find....) Yields 13% POT.

-- Do you mean here $4 on 4, 3 on 3 etc.

But everything depends on rule 2. distinguishing between a no-hoper and a slim chance.

Hermes




Also, in your sample there, when you say eliminate the TAB 1's, if you have a 1 and an 11 say in a race, do you then back the 11, or throw out the race altogether?

Nice work, keep it up.

[BlueTyson]

Couple of other quick questions - out of interest what is the POT for all last start winners?

How about those that ran recently if you have that?

Thanks,

BT

[TheDuck]

Speaking of favourites...

We have heard that 30% of favourites finish first. Does anyone know what the morning line looked like for the 30% and the 70%? What the final odds looked like? Quartiles or a box plot would be perfect -- or asking too much, sometimes hard to tell the difference. :smile:

[hermes]

Blue Tyson wrote:

Also, in your sample there, when you say eliminate the TAB 1's, if you have a 1 and an 11 say in a race,
do you then back the 11, or throw out the race altogether?


Just to clarify. In this method only consider the last start winner with the lowest TAB number. If that is TAB 1, eliminate and move on to next race. You don't then go down to the next last start winner in the same race, if any. Although you could. I haven't tried it.

Lowest tabbed last start winners yield lots of placeghetters and a good volume of winners too but the category as a whole needs further filtering to make it pay.Too many pay too little. You can locate the band that does pay in several ways. Removing TABs 1,2 and 3 helps, for instance, coz the poor payers tend to be those numbers (and usually favourites). You get much the same effect by looking at a market filter: eliminate anything under $5.

But it so happens, as I found by accident, that yes if you bet $3 on a qualifier carrying saddlecloth 3, $4 on 4, $5 on 5 etc. you get much the same effect - slightly better! Especially if number 11s like Upmarket Star win for 16/1. Just an idea.

Hermes

[BlueTyson]

On 2002-08-14 12:07, hermes wrote:
Blue Tyson wrote:

Also, in your sample there, when you say eliminate the TAB 1's, if you have a 1 and an 11 say in a race,
do you then back the 11, or throw out the race altogether?


Just to clarify. In this method only consider the last start winner with the lowest TAB number. If that is TAB 1, eliminate and move on to next race. You don't then go down to the next last start winner in the same race, if any. Although you could. I haven't tried it.

Lowest tabbed last start winners yield lots of placeghetters and a good volume of winners too but the category as a whole needs further filtering to make it pay.Too many pay too little. You can locate the band that does pay in several ways. Removing TABs 1,2 and 3 helps, for instance, coz the poor payers tend to be those numbers (and usually favourites). You get much the same effect by looking at a market filter: eliminate anything under $5.

But it so happens, as I found by accident, that yes if you bet $3 on a qualifier carrying saddlecloth 3, $4 on 4, $5 on 5 etc. you get much the same effect - slightly better! Especially if number 11s like Upmarket Star win for 16/1. Just an idea.

Hermes

[Bhagwan]

Hermes;

You might like to try the 8 Bet Place Staking Plan as mentioned in in the systems site.

Check it out,I think you will find it will perform like a train on your past selections.

Let us know what your percentage on turnover works out to be .Based on past results workout.

[croft]

TheDuck,
I lived in Toronto for 13 months in 1989. While there I went to a couple of race meetings at Woodbine and also Greenwood (I think that was the name). Anyway I believe the Greenwood track is closed. It was the first racetrack where I had seen night racing although it was not the best racetrack around.
I also spend a number of nights at the skydome watching the Bluejays and at Maple Leaf Gardens watching the Maple Leafs.
I lived in East York which you would know is not far from the city centre.

[hermes]

Haven't encountered a card like tommorow's before. Last start winners everywhere and most of them will be short priced. So its really a day to watch the market, but on paper:

rosehill

R1 #3 Takas
R4 #1 Go Errol
R7 #2 Master Pom
R8 #5 Skiddaw

Moonee V.

R1 #8 American Graffiti
R2 #4 Titanic Jack
R5 #2 Marstic

Eagle F,

R2 #2 Laurenil Impulse
R7 #10 Fairyman
R8 #3 Booboo

Vic Park

R7 #8 Cullen Bay
R1 #2 Miss Revic

Good punting

Hermes