Forum Replies Created
- Posts
Quote: from EC on 11:10 pm on Mar. 24, 2006[br]I know Robert what you are saying..in some races though the first five in the betting can mean 8 or 9 horses.
So does really narrow it down?
It sounds common sense because you are reducing the horses you are looking at..but the ones you don’t look at win rarely anyway. You still have between 5 and 8 horses to study.
maybe we should only study races with 5, 6 or 7 horses running…which is my preferred type of field size<br>
EC,
Well the number 5 is not cast in stone – it just illustrates the principle and fits the average field size. I would actually use my "third of field plus one formula" from my racing tutorial. So your specialisation of 5 – 7 runners would then only concentrate on the first 3 in the BF. An 18 runner field would concentrate on the first 7 in BF etc.
After a sift, I can pass up on many races as not worth betting on, so any time saving filtering them out, multiplies into saving many otherwise totally wasted hours on a busy racing day, that can be put to far more effective use. It all adds up to several weeks saved in a year and time is money.
robert
Quote: from EC on 11:15 pm on Mar. 23, 2006[br]around 80% of winners are in the first five of betting forecast;
that depends on how many runners there are
in 5 runners races = 100% in first 5<br>in 15 runner races = 70%<br>in 18 runner races = 64%
I never understand why people get impressed by this 80% stuff..it’s generalised and a little misleading and doesn’t really help us.
For example..in greyhound racing..100% of the winners come from the first 6 in the betting ..does it make it any easier knowing that though?
(Edited by EC at 11:19 pm on Mar. 23, 2006)<br>
EC,
Well spotted. <br>The more subtle point you may have missed is that if you have an average race of say 10-11 runners then do not spend 50% of your time looking for a horse ranked above 5 in the BF to beat any horse ranked 5 or below.<br>Spend 80% of your time looking at the first 5 and seeing which ones to back and which to lay. That is where the business side comes in making best use of time resources for maximum payback. And this is just one of a range of inputs that do help, not the only one.
robert
My rules that I have learned since starting as a 11 year old clock-bag runner would include:
learn your trade first;<br>if you have a family and a mortgage keep the day job;<br>specialise;<br>simplify to what really matters – the rest is noise;<br>0-2 races per day, in depth, with a reasoned case and value price against each contender;<br>around 80% of winners are in the first five of betting forecast;<br>there is NO useful inside information – you are the best informed specialist in your kind of races;<br>the fastest horse on the day (not yesterday) wins the race.
robert
Topcees,
Yes, this whole issue is tragic.<br>Samaritans do marvellous work on a one to one basis but the industry needs to preempt things from reaching those depths of misery. A funded service that collated all the contributory problem issues and worked with the industry to take the whole working culture from the 19th into the 21st Century might be a positive move on what can be done for potential victims in the future.<br>They do not seem to know at the moment what has hit them and why.
Robert
Hi,
Sadly, suicide is also a life sentence for the victim’s family and immediate colleagues. Just a little bit of help and encouragement at the right time could have / might have prevented that tragedy.
The ages of those mentioned in the article are mature people who have survived some of the worst aspects of the industry but got to an unobserved breaking point.
The bullying culture mocks weaknesses and friendly bonds are broken by the huge staff turnover. Lack of hope, bullying, lack of esteem, long hours with poor sleep, drink, drugs, debt are a sad coctail of possible causes but employers have a legal as well as a moral duty to make the industry safe and fit to work in.
A confidential staff councelling service working with the Samaritans could be set up for the racing staff in Newmarket and other centres as at least a start.
Robert
"Anyone doing form and/or time assessments should have taken this fundamental alteration into account"
Pru,
I expect Timeform would have spotted it??<br>Superform, Racing Post and Sporting Life didn’t and haven’t.<br>Introduction was around 1987 if failing memory serves.
Time ratings only effected if you use weight carried and pounds per physical length but you told me last time they were all interchangeable – not! ;)
Robert<br>
Ok I agree with EC and EW.<br>For outliers you have to decide, before discarding any in a small set, if there is any cause for the very fast or slow time and for Good-Soft this could be because the ground was firmer for fast times or softer for slower times or was the race run slower. (Those who crank out figures by computer miss all this kind of important detail). Often you will not know this too precisely but may suspect some figures – if so discard them. The median time method, which I do not like, has the advantage of taking the average of the 6th and 7th ranked times in your set of 12 and completely avoids the outlier (as duff data) issue.<br>So you could use all data for the average calculation, compare this with the median value and see if any agreement – add some more data and see which way the figures are heading. You mentioned SDs so take all data but use the population SD formula not the sample SD one. A smaller SD will indicate the obvious that your outliers are not too serious, your mean is about right, and vice versa.
Robert
Quote: from empty wallet on 11:52 pm on May 11, 2005[br]Robert or any Speedfreak :biggrin:
Hypothetical
Say if you had past race times between 73.0 and 77.5 (sample size say 12) on Good to Soft
would using  Standard Deviation of those racetimes give a more accurate Going assessment ?
(Edited by empty wallet at 12:02 am on May 12, 2005)<br>
EW,
I would not start at Good-Soft, which is about the most unreliable going for race prediction, but using standard deviations about the mean is the way I go. I won’t give away exactly how many SDs from the mean is most accurate but you are well on the right track. I also agree a lot with your later posts on how inaccurate and basically worthless the standard racing data fodder is – and long may it remain so.
Robert
EC,
I am with you all the way on this and have been doing same for over 20 years. The going correction calculated contains a measure of correction for wind speed and errors in the average distance the race is run at. You are only trying to establish whether the going forecast is about right or way off. The more way off the better. Jockeys are not reliable pundits on going or pace; going sticks and Turftrax are not reliable (the latter’s circular opinions I was told by them are based on clerk of course opinions) – time and race observation are very reliable.
Robert
ACR1,
The discussion point is about the belief that a profit is an irrefutable proof about having obtained value. The examples I gave are about as Homer Simpson simple as you can get – the practical reality of reliably obtaining value is far more complex – which is possibly why it proves so difficult for many to make it work – I don’t make the rules or make the markets. No brilliant insights have been revealed ( I hope) – no complex mathematics used – no obtuse probability concepts – just a small part of the basics of betting. If you want to learn more on odds-lines theory and practice then join a forum such as SmartSig where this issue has been thoroughly discussed and before-race examples regularly given, as well as "free" software made available to do your own lines.
Robert
Carvillshill,
I think you are missing many of the finer points here. Perhaps you have little knowledge of the basics of probability.<br>You are making fundamental and contradictory errors in your assumptions.<br>Edge is not effected by chance (random events). It is the difference between calculated "true" chance and the price (expressed as a probability) you can obtain, .<br>For value to almost guarantee a profit long term then you would need regularly to get say, 10/1 about a true 6/1 chance or less. This is a rare event in a competitive market and your practical edge would be far less if you have more than a few highly selective bets a month.
If the favourite is a totally false one then you have a far better (but not automatic) chance of getting value within the rest of the field. You still have to know which alternative one is providing the value to be sure – so you still need an odds-line as a reference point. You also need to be sure of how false the favourite actually is.
If you inspect the figures for price bands and winning percents over a year you will see that what has been said is correct. It is obvious that you are ignorant of such data and its huge relevance to basic betting strategy.<br>You also have difficulty comprehending what was clearly said. I stated that in any particular race you can have horses too short or too long. Over a long period these errors (up and down) average out so that the average is almost the same as the odds expressed as a percentage in the lower odds range. Over-round per horse and the short pricing of long odds horse makes the small differences in the annual figures.
Robert
Artemis,
The example given used 100 trials for simplicity. The argument and answer is much the same if you chose 1000 trials. 100 trials in the band of getting 10/1 about a horse you believe to be 8/1 takes an awful long time to happen. I never mentioned the long run. As an optimist I don’t accept the concept of ever being dead or broke. ;)
Carvillshill,
I understood what you said about the long term. Getting value is nothing to do with the long term. My argument tried to disprove your assumption. Each and every race in the simple example provided got value by backing a 8/1 horse at 10/1. Whether you win or lose after a series of races using that criteria is down to chance. The bigger the margin the better the chance of profit, (as per RobNorth earlier post), but the margins are generally very small in practice.
I would not wish golf on anyone, but selection method comes first – the bet follows only if you can get a good price. If your strike rate is 33% then it is no good backing value bets if they average less than 2/1, for example. I did not state you must beat SP to achieve value. I said this was a simple test of whether you are getting value.
SP, in the lower ranges, is proven year after year to be just below the expected strike rate for the long term case. The SP market is efficient, on average, over a year, but in any single race it can have individual horses too short or too long. Again, I did not suggest SP was a "true odds" price. Inverted commas were used throughout.<br>That is why you need to have your own better estimate of a "true price" odds-line before the markets opens. Of course you make the bet when the price offered exceeds your "true price" estimate. This is the whole point of having the "true price" established in the first place.
Robert
<br>
Carvillshill,
With you most of the way on this but winning over time is still not an irrefutable proof of value. It may be evidence of more likely having got value but you can equally get even better value each time and lose long term.
For example, if you regularly get a price of 10/1 for something which the true price (or an opinion of it) is 8/1 then that is value (by definition).
At the "true odds" chance (11%) you should average 11 wins in a 100 – paying out 110 points and losing 89 points of stakes (+21 points).
If it were actually a "true price" of 10/1 (9%) then you would in a hundred races only average 9 wins, pay out 90, losing stakes 91 (-1 point due to approximation).
The point that seems to be missed is that you do not actually get precisely 11 value (8/1) wins in a hundred races. You get a distributed number of wins purely by chance (11% win chance, 89% loss chance, each and every race) that could range from say 11 (+/-) 4. If you get 15 wins you are laughing. If you get 7 wins you are (+70 -93) = 23 points down.
You have undeniably got value every single time, but whether you win or lose overall is down to chance. At this price range, the occasional times you get the 15 winners can cloud your judgement as to what is actually going on. You might not have got any value but 15 winners at 10/1 is a profit. The fact that every race is different and you are in a different price range each time further confuses the issue.
You can even confuse yourself by saying I have got 15 wins in a hundred so the true price is "in fact " 6/1. I have made a bundle but I took 8/1 each time – so I didn’t get value and should not have backed them.
The only simple tests of value is if every bet you make beats SP or "true odds" by a good margin. The final SP can only be estimated and is in the hands of others. Those who get value more reliably know better than others how to assess a race and better estimate its "true odds", which is also in the hands of others – jockeys, trainers etc but there are less of them and their foibles are better known.
Robert
