Hidden Form
- Thread starter student
- Start date
Thanks Mick & Dave, see if this is what you mean. I've nearly filled notebook 27 (each A4 200 pages) mostly ideas that I haven't checked out yet on a big enough sample of races (200+). One angle is to find a horse thats much better than the OH and market thinks. Several ways of doing that. The bet race is the first time 'prospect' horse runs over a new racing condition (that could be one or more of distance, going, class, £winner, turf, AW, , etc). Sort of obvious pattern is re distance, 6 6 6 6 7. Or all runs over AW then turf. My reasoning is that most punters only like to bet on conditions that the horse is proven on. By the time that happens the price could have gone. Being contrarian, I look for the opposite. My notes regularly show a longshot winner where the new condition is the key. I've noted cases where there are 1,2,3, or 4 new conditions in the same race. I can't predict the pace for a race but I know what pace (as I assess it) the course or the race type favours, so that can be a new condition as well. And there are other conditions? Handicaps or non-handicaps (if you'll parden the expression). That alone gives a massive number of filters or combinations to go on. Most breeding chat is about the sire, so I take the mare. Better still look at the siblings of the mare.
How's that?
How's that?
mick
Sire
Hi
S
student Re your above while i get the thinking i fear i would not feel able to implement as part of my process as i tend to favour risking my wedge on the prospects of repeat form as opposed to improved or unknown. But as i recall your focus is on 2yo and 3yo only races and i can appreciate how this age group and race type would better lend its self to such. Have you investigated simular for the dark side 4yo + Hcaps. ?
Re contrarian thinking if you recall it was your good self who acquainted me with the description many years ago at the time it was something i had long been doing as part of my process which i perceived as attempting " different " thinking and while marketwise its good to disagree with the majority view my take is when doing so you have to know why.
I call the above finding racing reason (s) to back up any stats based data or ideas. A task which is often not possible but on the occasions when you do then this gives extra confidence Re its actual use. I am also pleased to advise that Re the older Handicappers i back they can often have prime conditions and needs met but for many different reasons still be available at bigger odds.........and thank goodness this remains so.
Imo the saving grace of this game is we never stop learning although having been a dedicated student for the past 50yrs i sometimes feel there is an element of injustice involved in having to accept the need. However i have my often used mantra " At least its seldom boring " as a fall back to justify, and even when things are going to plan i constantly question and challenge my own established thinking.
With this in mind during the past year i have made two what for myself are very big changes. One was to overcome a long held bias against backing horses with larger physical weights which for myself means any above 9-4, and the other is even more dramatic in respect that i now only back 5yo. My A4 pads also took a pounding as i researched the pros and cons prior to making these calls.
Perhaps i will post about my thinking and justifications for both on another occasion but suffice to say that so far at least both are proving good decisions on balance. Anyway its great to see you back posting and long may this last.
BWAL
Re contrarian thinking if you recall it was your good self who acquainted me with the description many years ago at the time it was something i had long been doing as part of my process which i perceived as attempting " different " thinking and while marketwise its good to disagree with the majority view my take is when doing so you have to know why.
I call the above finding racing reason (s) to back up any stats based data or ideas. A task which is often not possible but on the occasions when you do then this gives extra confidence Re its actual use. I am also pleased to advise that Re the older Handicappers i back they can often have prime conditions and needs met but for many different reasons still be available at bigger odds.........and thank goodness this remains so.
Imo the saving grace of this game is we never stop learning although having been a dedicated student for the past 50yrs i sometimes feel there is an element of injustice involved in having to accept the need. However i have my often used mantra " At least its seldom boring " as a fall back to justify, and even when things are going to plan i constantly question and challenge my own established thinking.
With this in mind during the past year i have made two what for myself are very big changes. One was to overcome a long held bias against backing horses with larger physical weights which for myself means any above 9-4, and the other is even more dramatic in respect that i now only back 5yo. My A4 pads also took a pounding as i researched the pros and cons prior to making these calls.
Perhaps i will post about my thinking and justifications for both on another occasion but suffice to say that so far at least both are proving good decisions on balance. Anyway its great to see you back posting and long may this last.
BWAL
Last edited:
Yes Mick, I only cover 2yo and 3yo. The main reason is that a stat'n is usually employed to get max value info from min data (cost time amd money to collate and analyse). So I thought min data looked like 2yos to me, so thats where I started. But I also predicted potential improvement (ratings) because thats what my business clients always wanted (not accurate history). So I just used my business problem solving approach to horse racing. One way of looking at my working a race, is to try to be one race ahead of what of the market thinks the horse will do. It wasn't what I set out to do but what it looks like on reflection. The standout example, was Frankle. After his first race I had a future possible rating that would put him in the first 3 of a 'standard' 2000gns. After his 2nd race I had him winning a 2000gns, and after his 3rd race I had him winning the 2000gns bigger than any 3yo rating I'd ever previously produced. He was off my scales. (I have standard values for every data that appears in the form book). But I was worried about his entry after the gns. The pace during that race was so strong that I reasoned the horse would need bigger than standard recovery days; horse knowledge and stats. A tricky thing to predict?
So my stuff is much easier to work out than your older horses, Mick. I take my hat off to you. I was always fascinated to read your race workings.
So my stuff is much easier to work out than your older horses, Mick. I take my hat off to you. I was always fascinated to read your race workings.
Last edited:
mick
Sire
S student i do not feel that working out any of the stuff yours mine or others is easy but then again when this is the case its arguably not worth having in the first place. Re my bet write ups i still do them on another section of this forum. Forgive the indulgence but below is a recent one which left me feeling chuffed. This type are certainly helpful to the mind as well as the wallet when others backed for simular reasons run a mare.
Wen 25th Kemp 6.0 : Dynakite @ 20/1
An 8 fur 0-55 with 10 run. Kite has the looks of a forlorn hoper and this may prove so here, but there is a case of sorts to be made and a likely tasty price to justify the more obvious risks. This time last year he won two CD Hcaps but he has shown nothing during his eight runs since and this maybe a genuine regression job but i could not make a case for him in any of those runs whereas in this race his rating and profile suggest that he has been placed to win and if the stable are thinking likewise then i am hopeful that he will be running with some positive expectation.
This was not so lto with RI noting " never threatened from off the pace but he didn't have a hard race either " i had to grin because the latter part of this comment lends itself to another interpretation. Kite appears to do best in these double digit size fields and when emulating via 10 run - Hcaps his 7 stall shows best of the bunch. He has won under 9-8 so his physical weight should not prevent and this will be a return to a 0-55 grade where over this CD he is 1-1. As his last win was off OR 62 in a 0-65 then he is certainly well found if returning to anywhere near his best.
I like the rider who has made a solid 10 winners start to the year and had a lto sighter on Kite, i also like the trainer but regretfully can not find any recent positives but longer term he has proved that given the ammunition he can produce one for a punt. This race contains a likely hot pot Fav Bawaader who is attempting a three timer, but now an 8yo he could be worth taking on as both of those wins where 0-50 non Hcaps and he is now 8lbs higher than the first one after which RI noted " It was his trainer's first winner for 3091 days (255 runners under both codes) ".
B365 have opened @ 20/1 and this suits.
Without knowledge of probabilities laws (the language of uncertainty) the risk reward balance can be a real challenge for intuition in investment decision making . To some extent there are aspects that are hidden, like the risk of a long run of losses. Computer simulations of 1000 bets for particular values of win% and av profit per bet can produce some suprising outcomes. One general point is that one should be prepared for a losing run that is more than twice the average losing run. Kelly maths is now used in stock markets as well as by betting syndicates. Its based upon proven probability laws. Whilst no maths can be used to calculate a staking plan to turn a loss making policy into a profitable one, it can assist in boosting profits of a profitable one. Most other discussed staking schemes fail the simulation 1000 bets test. VDW once proposed a staking method that produced disasterous results in the test. Later there was speculation that he was using dutching stakes, splitting a usual bet on one horse to cover several horses when the edge and strike rate of each horse warrants it; much better. Kelly maths showed that one should bet a proportion of the current bank. But full Kelly staking needs nerves of steel. I've used 0.5 x Kelly stakes that gives more stable stakes and bank. It gives slightly less profit but significantly reduces the risk of ruin. Thats for a single investment policy. If one has several policies with differing SR & Edge, its probably best to stake each as indicated by their separate stats.
mick
Sire
Deploying KISS principles would not work for myself indeed i feel suspicious of the easy find. But the one exception is staking with my way being a straightforward 1pt win on each. While i could find fault with my own thinking having over the years looked at other staking methods with an open and receptive mind i always end up concluding that for myself level stakes best suit.
As S student rightly says our future results are a Hidden unknown and any backer who has experienced a lengthy run of consecutive losers will carry this as a Monkey on his back. Long term the avg price of my selections win and lose is > 10/1 so it would be fair comment to say that i am more susceptible than most to this grievous experience and perhaps the real question should be how do we deal with it. ?
Because i have fewer bets then they have to be larger in order to create a realistic annual turnover from which to obtain my net profit, and one mental trick i have learnt to play is to think pts not £s and only convert the former to the latter at each years end. When experiencing a run of losers i have found that viewing this in the longer term is also helpful.
As an example if your YTD Lsp is 28pts and you then experience a rotten run of 16 consecutive losers you still have 12 chances remaining to sort the situation without losing anything. ? I feel the most important topic S student mentions above is to prepare for the worst and there are two parts involved. Firstly having ample + working capital and secondly dealing with the mental aspects which can kick in and affect your confidence and judgement. With perhaps another being our personal Risk - Reward mentality. ?
Both of the above are worthy standalone topics and as ever i will find far more interest in reading other members thoughts and experiences Re both than just repeating my own waffle on same. If you have any thoughts on all matters Hidden then please do share them as i feel this threads title opens the door for some very interesting discussions.
Having said that with this being the public part of the forum which can be viewed by non members i can appreciate a reluctance from some to offer full transparency i feel the same and those who are currently experiencing problems Re obtaining advertised odds will understand why, but there is still room to make mention of other Hidden aspects without potentially sabotaging yourself. !
As S student rightly says our future results are a Hidden unknown and any backer who has experienced a lengthy run of consecutive losers will carry this as a Monkey on his back. Long term the avg price of my selections win and lose is > 10/1 so it would be fair comment to say that i am more susceptible than most to this grievous experience and perhaps the real question should be how do we deal with it. ?
Because i have fewer bets then they have to be larger in order to create a realistic annual turnover from which to obtain my net profit, and one mental trick i have learnt to play is to think pts not £s and only convert the former to the latter at each years end. When experiencing a run of losers i have found that viewing this in the longer term is also helpful.
As an example if your YTD Lsp is 28pts and you then experience a rotten run of 16 consecutive losers you still have 12 chances remaining to sort the situation without losing anything. ? I feel the most important topic S student mentions above is to prepare for the worst and there are two parts involved. Firstly having ample + working capital and secondly dealing with the mental aspects which can kick in and affect your confidence and judgement. With perhaps another being our personal Risk - Reward mentality. ?
Both of the above are worthy standalone topics and as ever i will find far more interest in reading other members thoughts and experiences Re both than just repeating my own waffle on same. If you have any thoughts on all matters Hidden then please do share them as i feel this threads title opens the door for some very interesting discussions.
Having said that with this being the public part of the forum which can be viewed by non members i can appreciate a reluctance from some to offer full transparency i feel the same and those who are currently experiencing problems Re obtaining advertised odds will understand why, but there is still room to make mention of other Hidden aspects without potentially sabotaging yourself. !
Last edited:
A masterpiece on the subject Mick. Losing runs can be the extreme events that there is little data to go on (hopefully). At one time I was thinking of having 2 banks' a normal working one, and an emergency just to keep one afloat provided you hadn't detected an explanation that the selection method had changed for the worst.
When I was tipping, first in the factory, then online before the web, I iook comfort in selecting races where the win probs and odds showed value in several horses. It usually ensured that the odds on each horse would give a good chance of a good return. To my relief it worked for a whole flat season in those days (no aw) and I retired satisfied when they wanted me to repeat it next year and I passed.
I wonder whether you'd ever time to run a paper staking trial on your past selections. Bet 2.5% of cuurent bank on all selections. The 1000bet simulations suggest that it would gibe better growth than level stakes. Of course this is a sig change; having to calculate a bank but one can just set it at 100pts, will not change the validity of the trial.
Just a thought
When I was tipping, first in the factory, then online before the web, I iook comfort in selecting races where the win probs and odds showed value in several horses. It usually ensured that the odds on each horse would give a good chance of a good return. To my relief it worked for a whole flat season in those days (no aw) and I retired satisfied when they wanted me to repeat it next year and I passed.
I wonder whether you'd ever time to run a paper staking trial on your past selections. Bet 2.5% of cuurent bank on all selections. The 1000bet simulations suggest that it would gibe better growth than level stakes. Of course this is a sig change; having to calculate a bank but one can just set it at 100pts, will not change the validity of the trial.
Just a thought
mick
Sire
Hi
S
student Re your above in the past i have run simulations of various staking plans against my own past actual bets and when judged against my level stakes benchmark some have shown positive but comfort zones also play a part and my own where and are to remain as i am. Of course the journey involved in a very slow safety first reinvestment of all profits on a year to year as opposed to bet to bet bases which i deployed in order to reach the unit stake which made things worthwhile and one i have now remained on for a very long time is another story and one which took me many years to complete.
These days i anticipate having less than 100 bets per year and work off a 100pts bank with the same amount held as a reserve, which i hope will never be needed but regretfully.........who knows.
I judge my own worth via Roi but here's a thought if a backer deploys a 1pt win for the whole year with his 4th bet winning @ 8/1 and as a result never having to put his own money at risk during the rest of the year which ends with a 50pts Lsp what Fig would best describe this. ? I have done so and better previously and taking the above example seen as a max 4pt risk producing an annual 50pts gain i suspect the City Boys would not believe. ?
These days i anticipate having less than 100 bets per year and work off a 100pts bank with the same amount held as a reserve, which i hope will never be needed but regretfully.........who knows.
I judge my own worth via Roi but here's a thought if a backer deploys a 1pt win for the whole year with his 4th bet winning @ 8/1 and as a result never having to put his own money at risk during the rest of the year which ends with a 50pts Lsp what Fig would best describe this. ? I have done so and better previously and taking the above example seen as a max 4pt risk producing an annual 50pts gain i suspect the City Boys would not believe. ?
Last edited:
ARAZI91
Mare
Absolute "True" Kelly is a much misunderstood concept in A v k type markets (multi-competitor sports such as horse racing), no matter what divisor you use, full, half, quarter etc, reason being that when John L. Kelly worked on that paper between 1956-58 he only used an A v b type market scenario. It is a myth that serious big Kelly type bettors such as Benter, Woods, Zeljko/Walsh, Warren Buffett, Ed Thorpe, Charlie Munger, Bill Gross and probably the greatest of them all in any market type scenario - James Simons - bet only "overlays" determined by "perceived edge obtained". They also bet a proportion of slightly optimal "underlays" (usually ranked by "EV" in market order) based on their own personal risk/reward ratio. This retains "skin in the game", maximises "churn" in a turnover sense and negates some of the much maligned Risk Of Ruin (ROR) concept.
The original work that preceded much of Kelly's work at the Bell Labs industry in that time period was based on Claude Shannon's work on "information theory" and was based on a system that analysed information over networks. By far the biggest "driver" in the use of "True" Kelly type betting and it's related variations constructed over the years are a heavy reliance of the accuracy of the pre-outcome probabilities used in forecasting. Authors like Nick Mordin and others have over the years in print tried to simplify "Kelly" but have remained stuck to the single outcome scenario of the A v b type market example used in Kelly's original paper hence causing some of the confusion.
So far it is the only known system of staking in any risk related venture to offer full mathematical proofing. For example, one major error still used in daily publications of stock exchange pricing such as Nasdaq, MorningStar etc is the use of this part of the formula - Kelly % = W – [(1 – W) / R], where W is the win probability and R is the ratio between profit and loss in the scenario. This is simply incorrect and does not account for "volatility" (which is the magnitude of potential profit/loss) but only their ratio to each other and only works in a market where bets are "binary". Changing that very small part of the formula to - Kelly % = W/A – (1 – W)/B, where W is the win probability, B is the profit in the event of a win, and A is the potential loss, minimizes the "downside scenario" of the formula. This is because in stock exchange pricing the downside-scenario probability must be set to the probability of a total capital loss, not the much larger probability of some loss.
Any better
mick ?? - the 3 paragraphs are connected in a way.
Last edited:
Hi ARAZI91, thanks for that posting. I only came across a ref to use of Kelly on stock exchanging whilst trying to see how the topic of 'staking' was being discussed and used now. My personal experience is restricted to horseracing and was focused on the estimate of win probabilities based upon the error distribution on ratings that I measured, after modelling their life histories.
Having modelled many processes in industry/business, racehorse performance seemed much more volatile and I was interested to see how things worked out on my computer assisted 'decision support system', particularly on 2yos in about 2003. I rated every 2yo race for one flat season.
I collated the actual win% for each band of predicted win% and was astonished that the agreement was so good except for 0-5% and >85% ranges. I thought 0-5% was possibly due to the extra difficulty of predicting 2yo debutants and the >85% was down to small sample size and possibly different mix of race types? In that method, the max ROI measure was in the 30%-35% win range.
It got me thinking about the effort that goes into breeding, preparing, training and riding to get the best end results. I ran the same DSS for another 2 years before Mick persuaded me to go over to the dark side and I extended that approach to 3yo handicaps. Then I discovered that hidden form was the most exciting.
Management consultancy often involves risk assessment and I was often asked to decide on what should be done. But I would argue that I wasn't being paid to make those decisions; they were. Invariable I would have to provide a visual scale of risk with likely consequences to help them make a decision. The language of probability still seems relatively rare in UK, less so in USA?
PS Including 'underlays' can have an interesting effect.
Having modelled many processes in industry/business, racehorse performance seemed much more volatile and I was interested to see how things worked out on my computer assisted 'decision support system', particularly on 2yos in about 2003. I rated every 2yo race for one flat season.
I collated the actual win% for each band of predicted win% and was astonished that the agreement was so good except for 0-5% and >85% ranges. I thought 0-5% was possibly due to the extra difficulty of predicting 2yo debutants and the >85% was down to small sample size and possibly different mix of race types? In that method, the max ROI measure was in the 30%-35% win range.
It got me thinking about the effort that goes into breeding, preparing, training and riding to get the best end results. I ran the same DSS for another 2 years before Mick persuaded me to go over to the dark side and I extended that approach to 3yo handicaps. Then I discovered that hidden form was the most exciting.
Management consultancy often involves risk assessment and I was often asked to decide on what should be done. But I would argue that I wasn't being paid to make those decisions; they were. Invariable I would have to provide a visual scale of risk with likely consequences to help them make a decision. The language of probability still seems relatively rare in UK, less so in USA?
PS Including 'underlays' can have an interesting effect.
ARAZI91
Mare
Just another thought S student -have long retained the notion that "paper trialing" a selection method whilst has value ,i feel a small bank type scenario provides better outcomes - as long as you can afford to lose this "small bank", you'll retain "skin in the game" where each selection simply because there is monetary value behind it will probably have more level of study behind it, and the lessons learned in that process can be more positive going forward than just simply logging selections on paper.
A bet should mean something to the bettor, a commitment - in most cases - financial. How high or high low is ALL relative to each punter -meaning one punters £1.50 stake is as RELATIVE to another punters £750 stake.
Last edited:
mick
Sire
ARAZI91 Although i have not felt the need i think your above comment is valid. If while auditing my past bets i glean a possible strength or need i am very aware of the stakes deployed and the amounts won or lost when contemplating making a change based on same ( i made mention of two in my post #5 above ) and perhaps this could be seen as same difference. ?
ARAZI91
Mare
Yes, this is actually a fundamental of "True" Kelly Criterion principles -ranking a race by EV is the most optimal way to do this and your looking to bet only a slight -EV proportion of "unders" .......as said it retains "skin in the game", increases "churn" and counteracts Risk Of Ruin to varying degrees.
Many would be asked on "Kelly" and they would give the answer "it is a maximisation/minimisation staking technique based on perceived edge obtained"- there is a lot more to it than that.
ARAZI91
Mare
Agree
mickArazi91, couldn't agree more. I studied losers of real money far more than winners. But I study everything hard; its what I do. Been there done that. But much of my life has been modelling situations and passing on ideas and tools and techniques to my students who then went on to land and successfully delivered solutions for some of the biggest companies. Thats the biggest kick I've ever got. It showed that there must have been some science there; the ultimate test that others get the same results.
The best example was a student who recommended over 50 improvements in his final year project. This leading UK company wrote a letter to me. Their typical reports from International consultants comprised 3 to 4 recommendations without detailed instructions of how to achieve them (they usually do this hoping to get another contract to do the detail). This company had checked over half the students proposed actions to date and all had been successfully implemented. Thats to show how bad things are out there. Not how good I am. I'm dead average, just work hard.
As I've said elsewhere, having had difficulties getting money on, it took the shine off doing that. I'm 'rich' but time poor. Story of my first attempt to have a bet in my local town here in north wales. Tried to play dumb, but asked if any rules other than in their rule book. He asked me a few questions, were I lived, how long, my job, then asked how many bets would be handicaps, I answered to one decimal place! Lol. Just lost concentration? That was it, he reckoned I was a 'jobber' for a racing stable.
But my daughter ex-show jumper trains my granddaughter and her ponies in dressage (11th in UK age group - we all work hard) wants to start betting again. She had a year off between degree and PhD (made hundreds?? after her first bet/£10 winner). She's going to have some free time in the summer. So, I am providing her with the ratings and expected values. I'll let you know how we get on if you wish; in the meantime keep going. No luck, just hard work.
Bye for now.
The best example was a student who recommended over 50 improvements in his final year project. This leading UK company wrote a letter to me. Their typical reports from International consultants comprised 3 to 4 recommendations without detailed instructions of how to achieve them (they usually do this hoping to get another contract to do the detail). This company had checked over half the students proposed actions to date and all had been successfully implemented. Thats to show how bad things are out there. Not how good I am. I'm dead average, just work hard.
As I've said elsewhere, having had difficulties getting money on, it took the shine off doing that. I'm 'rich' but time poor. Story of my first attempt to have a bet in my local town here in north wales. Tried to play dumb, but asked if any rules other than in their rule book. He asked me a few questions, were I lived, how long, my job, then asked how many bets would be handicaps, I answered to one decimal place! Lol. Just lost concentration? That was it, he reckoned I was a 'jobber' for a racing stable.
But my daughter ex-show jumper trains my granddaughter and her ponies in dressage (11th in UK age group - we all work hard) wants to start betting again. She had a year off between degree and PhD (made hundreds?? after her first bet/£10 winner). She's going to have some free time in the summer. So, I am providing her with the ratings and expected values. I'll let you know how we get on if you wish; in the meantime keep going. No luck, just hard work.
Bye for now.