- This topic has 336 replies, 13 voices, and was last updated 19 years, 5 months ago by
.
-
Posts
-
I have had to delete the details of this method to protect it from plagiarism. Sorry.Edited by Artemis at 12:27 pm on July 18, 2005)<br>
(Edited by Artemis at 1:35 pm on Aug. 2, 2005)
Artemis … that is a fantastic post, one of the best I have read on here for a very long time, well done … <br>:clap:
I have thought for a while now that punters follow a road of evolution in the way they bet and view horse racing in general before they start making a profit. And between each step there are dead ends and frustration before you make that next quantum leap onto the next rung of the ladder. I think this is approach is near the top.
I have a few questions I would like to ask you before I share some similar findings with you that you may find interesting.
I have put up the Question as a Q and the reasoning behind the question, to avoid any mis-understanding.
Q. Why do you use poundage as your baseline? <br>Reason: I personally find the use of poundage a bit cumbersome and don’t really understand how it works.
Q. Why do you use Speed Ratings and Collateral Ratings?<br>Reason:  Horses that haven’t run for a while or are changing class may have ratings that are unrelated to the current state of play. Would you not be better off playing up one and ignoring the other, instead of playing one off against the other and getting a generalised rating?
Q. What is your final rating figure measured in; points or percentages?<br>Reason: Interest.
Q. Do you think your ratings are over complicated?<br>Reason: Adrian Massey.
My main criticism, and don’t take it as a downer because, it isn’t mate, is that there is nothing in the whole process laid out above that is ‘unique’ to you, it’s all out of the racing post! I think to move into the high teens percentage profit, you have to be unique.
Here’s some information, in turn for your great post and an example of what I mean …
Adrian Massey.<br>I looked into his ratings sometime ago and worked out how he comes to his rating figures. I swore at the time that I wouldn’t post this up, but since he’s gone all pointy headed and has restricted access to his site, I reckon he’s got it coming to him sometime, so here it is! I’m working from memory here so this might not be the exact way that the program works, but this will give you a flavour of what he is all about and what I mean about being unique.
He rates HCP races different to Non-HCP races.
HCP Races.<br>Look at the last two races and work out a reverse form rating. Look at the finishing position and deduct this from the number of runners. <br>Example; 2nd out of 14 gets 12 ….. 7th out of 14 gets 7 ….<br>Do this for the last to races for every runner, add them together and divide by two. This is good, because the form figures are an interpretation of normal form figures which are in the public view, but they have more meaning and depth and are hidden away.
Then he adds points, run in last 31 days +4, significant draw percentile +2. (This bit may be wrong)
Then add all of the ratings figure together and divide each individual rating by the sum and convert the points into percentages.
To finish it off, he looks at the class of the race last run and loads the individual scores for each horse.<br>If it’s going in the same class as the last one he multiplies the figure by one. If it’s going down in class he multiplies it by 1.1 for one class 1.2 for two classes and 1.3 for three classes. If it’s going up then it’s 0.9 for 1, 0.8 for 2 and 0.7 for three.
This isn’t how he does it in real life obviously, a program is doing it for him and it will probably work on percentile figures from the off. But doing it this you get the same ratings, more or less, as he does.
Anyway, I enjoyed writing that and look forward to your responses.
Ps. I don’t give to charity, they can take it out of my Council Tax … :biggrin: <br>
dave,
I am absorbing your questions and comments and the stuff about adrian massey(whom I’ve corresponded with in the past). Will reply soon.
An excellent article Artemis.  As you say, it does look like quite a bit of hard work to get the figures out.  How long does it take to rate an average race?
Is there an electronic form of the data that could be used, and therefore automate your ratings?
lookingforawinner,
On average, about 10 minutes. You are really only interested in horses with a recent good Topspeed rating and can safely eliminate any horse with an X or? against its name in the recent form column of Postdata.<br>A horse with no recent Topspeed figure can only get about 4 points max from its RPR, so would have to pick up at least 6 points from the rest of the procedure.<br>This would need the horse to be ridden by a top jockey and have a really good profile and then be able to get within three points of your top rated horse.<br>I always try to get my top rated horse’s points first, and it soon becomes obvious(with a little practise), if a horse with no recent form can get anywhere near it.
You can set up a spreadsheet and import the data from the RP, but it may cost. It’s easier to use the newspaper as your ‘scribbling -pad’. I work out the first part of the rating in the Postdata section, then transfer this figure to the racecard. I write the jockey’s points and the draw points straight on to the racecard, and the hose’s profile/today’s jockey points beside the horse’s lifetime form. Finally, I add them up and that’s it.
dave,<br>There is no short answer to the weight question – I presume that’s what you mean by poundage. It merits a lengthy chapter in any decent book on handicapping. My own view, and that of many others, is that adding weight slows horses down whilst taking weight off may not enable horses to run appreciably faster.
All the ratings systems I use are based on weight and I don’t have access to alternative models, although it might be interesting to look at any that disregard weight. I know Adrian Massey’s ratings seem to ignore weight, but I don’t really understand them.
I have been working with the 0-140 scale and the WFA scale for a long time, and I am comfortable with it and confident that it is somewhere close to the truth.
I prefer speed ratings over collateral form ratings and consequently, I give them more weight. A good speed figure is proof positive of a decent performance whereas a decent form rating may be suspect. The reason I mix them or average them is merely to moderate unusually large ratings. Timeform ratings are essentially collateral ratings moderated for speed and by the skills of their racereaders. I would probably use them if I could afford them. The RP costs me about £500 p.a.
The final rating is really points, but a point is the same as a lb in the ratings part of the method. In the second part, I am still thinking in terms of 1lb=1point, but it can only be intuitive. Different people will rate jockeys/draw and other factors slightly differently.<br>The end result is a top rating that is on average about 8 points higher than the RPR top rating because of the ‘intuitive’ points added by me.
Yes, the ratings are complicated to other people, and I apologise for that. Many people will look at them and think that they’re not worth the bother. However, I know MY version of them inside out and it has certainly given me confidence and saved me a lot of time going through form. It is also quite pleasurable to find the odd big priced winner and not have to blame other people for my losers.
It is very difficult to get a unique approach in today’s ‘number – crunching’ approach to form. Every detail is examined inside out. Many people are running models that analyse form to the Nth degree using statistical models and neural networks. They inevitably come up with the same horses which are consequently shortened in price and end up poor value. I have found that most of my top-rated horses shorten up from the morning odds, probably because I’m using the same raw data as more sophisticated models. although I’m not using as many variables as they do.
It is possible to go on and create a program that will calculate odds and produce a tissue , but it requires more work taking you into the realm of professional betting.
With regard to Adrian Massey and his methods, he does seem to do very well, so he must be doing something right. His approach seems to be that of an impartial observer, unencumbered by long years of hanging around betting offices and racetracks. He has looked at racing very logically, using only facts. No hype, no hard luck stories, no weight to worry about. Just bare results. I certainly admire his methodical approach and his skill with computers. I’m surprised he doesn’t charge for his services. I think I would if I could present my work in the same way that he does.
Thanks for the quick and honest reply Artemis, very interesting! I like the bit about, not having anyone else to blame if you lose, how important is that?  ;)
There’s no need to apologise for your system being complicated, it is your system after all! And when you write something like that down it always looks worse than it is.
The Exit Plan system that I have posted up on here, is the exact same kind of system. It’s not as complicated as your system and doesn’t use weight, but the principal is the same and I do it all in my head. You will be the same, you’ll look at a race and KNOW which one the top rated, I bet? I do with mine! I’ve been doing it that long, I can’t remember how I came up with the rating in the first place .. :biggrin: what a fool …
This approach and the way you are going is spot on, imo! I like the way you moderate the Speed figure with the Collateral Form figures, that’s a good idea.
You mentioned above that the unique approach is difficult to get right …
Do you remember the 5/4 system that was posted up on here a while ago. It basically took, 5 top ratings and if they were all clear and all ‘agreed’ then it was a bet? Everything was fine with the system for a while and then it started to wobble, so out went the odds on shots, then it got even wobblier and out went the HCP races. Then it wobbled itself to bits and out went the whole system, filtered and back-fitted into obscurity! The problem with it being, that it picked the very best Speed and Form rated horses on the day, everyday. If picking winners was as easy as that there wouldn’t be any bookies!
I think the more criteria you include in any rating the more your rating will gravitate towards the favourite and away from value.
The Massey ratings are reverse form ratings, in HCP races. He can’t charge for what he does on his site because he downloads his race cards for free, off the telly! If he were to start charging people for access to the site he would find himself having to pay for information he’s getting for free at the minute.
One thing that did take me back a bit there Artemis was you thinking that making up a tissue was difficult, it’s actually very easy. I learnt how to do it myself on here a good while ago, I’ll post up how you do it later on, bit pushed for time right now!
Dave. <br>
Hi Dave,
I don’t have any problem making up a tissue because I used to be a bookmaker. What I meant was a tissue that could be a useful tool for betting purposes, highlighting any ricks made by the professionals. I usually have a pretty fair idea of how the market will go and really should make more use of this knowledge.
Your’e right about finding the top rated in the first part. It usually takes about a minute, but it takes a little longer to see if the horse stays at the top after applying the extra points.
Systems Wobble is a good subject. As you say, very few systems based on a few basic facts endure for a long time. The relentless percentage working against the backer eventually gains the upper hand despite the frantic introduction of refinements to the system. If anyone has come up with the holy grail, they are keeping quiet about it. I would.
Too many criteria do tend towards shorter odds where random chance and chaotic events can wipe out the small profits.
One final point. The best computerised systems do factor in the view of the market as close to the off time as possible. The weight of money (itself influenced by computer models) is still a pretty important factor.
Some good points there again Artemis, sorry about the misunderstanding about making up a tissue, it did make me wonder!
I know what you mean about a tissue with any value as a betting tool. Recently, I’ve started making up a Backer’s Tissue which is prooving to be quite a success. Working to an 85% book makes a big difference, I’m going to start doing a Layer’s Tissue next and see how that works out. It’s still early days with that, so I’m not fully confident in the process yet.
What I do, at the moment, is use a series of tables and judgements to make up my ratings. Once the ratings are done I make them into a book and back everything that I think will be good value, between the price offered and the price on my tissue.
Have you ever considered this approach?
brief reply,dave.
The problem is that this is only a hobby for me,being retired.
If you want to put things on a business footing, you have to start talking serious money.
Like most people, I have hostages to fortune in the shape of family, house, commitments, need for holidays etc….etc…… I stay inside my comfort zone.
Now if I had a spare £50,000 or a sponsor/patron, all of the things you suggest are feasible. Unless you are prepared to become a professional, it’s not worth the extra work and the aggravation/risk. I suppose this scenario applies to most of us who tinker around with systems/methodical betting.<br>
I’m only a small fish in a big tank Artemis and don’t bet to big stakes at all !! I have done in the past but found the whole experience a disaster, something I wouldn’t want to put myself and my family through again!
My stakes now fall between, £1 and £30 generally, £5 a point being about average … :biggrin: I’m happy doing that because there’s no pressure and I can just enjoy what I am doing. I also don’t expect to win. Like you said in an earlier post the bookie will get you in the end, or the exchange layer come to that, fair enough!
I do all of the things I am suggesting to small stakes. I can get a very, very long run out of a relatively small betting bank !
When I get some time I will rate, price up and post up a race on here and let you see what I mean … it’s actually very easy once you’ve done it a few times.
All pleasure and no pain …. Promise …. :biggrin: :biggrin:
Hello Artemis-a fine read.It looks like it gives the individual the scope to weight positive factors how he sees fit. The draw at certain courses for example todays meeting at Salisbury has a much greater influence on the outcome than the draw at Yarmouth it can be weighted accordingly.A comprehensive approach and you have the satisfaction of knowing that it is your own calculations that will determine success or otherwise.<br>What as punters can we reasonably expect I feel the approach will guide the user to a calculated selection.I wonder what percentage of qualifiers would place I have a feeling it could be quite high maybe lending itself to the place markets.<br>Enjoyed reading your post and the following discussions its a pleasure to have a thread that concentrates on approaches to winner finding.Thanks<br>I’ll be doing some dummy runs on D Handicaps and above weighting as I see fit.This will provide an interesting challenge
kersly ÂÂÂ
(Edited by kersly at 12:02 pm on Sep. 14, 2004)
Artemis – Thanks for your reply. When I was young and keen I had the RP delivered everyday, and as you say, it became a scribble pad. Back then I would use either the Fineform (variation of) or C/D systems during the week, but on Saturday’s I would spend a good 4 or 5 hours studying the form on 1 or 2 races. The former ticked over with probably a small loss, whereas the latter made a reasonable profit. But the profit was never enough to justify all the time spent. This was because my betting kitty was small, and I couldn’t stomach losing a £20 bet (or more).
A wife and family now mean I have very little time to devote to the game, so I have looked at systems based on data from various sources. I currently use the SSR product from Racing Sense that uses the Solidus ratings. I don’t use the ratings as such, but the database is great to devise my own systems. There are timings for each horse, so I could in theory develop a speed rating and go on from there.
dave jay – I would be very interested to see an example race with ratings and prices.<br>
Here’s one I prepared earlier … :biggrin:
Yarmouth 4.50<br>6F Conditions Stakes <br>9 runners (Class C)
1 Russian Valour<br>2 Atavus<br>3 Boleyn Castle<br>4 Colonel Cotton – 6/1 paper tissue – 6/1 my tissue<br>5 Country Reel – 15/8 paper tissue – 6/1 my tissue<br>6 Whitbarrow – 11/2 paper tissue – 13/2 my tissue<br>7 Flushing Meadows – 5/1 paper tissue – 8/1 my tissue<br>8 Forest Rail<br>9 Dowager
I would normally just scribble all of the paper so I’ll just quote the horses numbers below and refer back to the top at the end. Using the Daily Mirror – add each horses rating as a basic rating. No.7 has no rating (lowest rating =20 highest = 48) so I’ll give it 32.
1. 32 + 5<br>2. 41+ 5<br>3. 38 + 5<br>4. 47 + 10 + 5<br>5. 48 + 15<br>6. 42 + 10 + 5<br>7. 32 + 15<br>8. 20 + 5<br>9. 47 + 5
Add <br>10 points for every horse in the first 4 of the betting and a further 5 for the first and second favourite.<br>5 points for every horse that has run in the last 7 days.
Add totals
1. 37 / 432 =  8.56% – odds 11/1<br>2. 46 / 423 = 10.65% – odds 9/1<br>3. 43 / 423 = 9.95% – odds 9/1<br>4. 62 / 423 = 14.35% – odds 6/1<br>5. 63 / 423 = 14.58% – odds 6/1<br>6. 57 / 423 = 13.19% – odds 13/2<br>7. 47 / 423 = 10.88% – odds 8/1<br>8. 25 / 423 =  5.79% – odds 17/1<br>9. 52 / 423 = 12.04% – odds 7/1
Add all of these together =  432 .. divide each horses total by this number.
That’s taken me 10 minutes to make up the ratings and the tissue. I just use a spreadsheet to do the Maths and round the odds off as I see fit.
As there are only four runners quoted in my paper tissue and I don’t normally bet in the outsider zone. So the horse that seems to be fair value is;<br>4 Colonel Cotton – 6/1 paper tissue – 6/1 my tissue
BTW .. these aren’t the criteria I use, just an example of making a book up.
There are loads of ways to do this. Some people group all of the outsiders together and give them a collective percentage of say 15% and then make a 100% book up on the rest.
The problem with doing ratings, in general, is that if you use existing ratings or refined data as a base for your own ratings and tissue, you end up with a tissue with 7/1 favourites and 14/1 outsiders, as above. This is because the numbers are too close together.
In my experience a good spread of numbers is around 5 to 25 with some horses scoring 5 and others 25, for a full show tissue.
I also try and work with bands when looking for value, like not stepping outside of a limit, say I have priced a horse up at 8/1 and have it as my third favourite. But on the paper tissue it’s the sixth favourite at 25/1 I would probably leave well alone.
I hope that has made a bit of sense!
Dave. <br>
dave,
The spread of the prices on your tissue look a bit tight.<br>I understand what you mean by getting such a spread using only ratings.<br>I used to get round the problem by setting a base figure. In your Daily Mirror example, you could raise the base line of the Spotform Ratings from zero to say 20 or maybe higher. The points awarded for the other things then become more significant, so you have to be careful.
I don’t think you get a very good tissue on ratings/positive points alone – you need to bring in some other information such as how well tipped up a horse is, and how the horse is priced up on the RP tissue.
The BRIMARDON system advertised in the racing press actually works out tissue prices for each race, and then advises people to bet where there are large favourable discrepancies. They do seem to find some big priced winners, but they do every race and may advise more than one horse per race, so their net is cast very wide.
dave jay,
Thanks for the well worked example. Using this method, how many bets do you make a day?
Do you compare with the bookmakers odds, the paper’s odds or the odds available on Betfair?
If I understand what you are saying, you are not betting on which horse you think will win, but on the horse which appears to be good value (and within your limit).<br>
- You must be logged in to reply to this topic.