Valiant Thor
-
Posts
4,023 -
Joined
-
Last visited
-
Days Won
15
Posts posted by Valiant Thor
-
-
6.10 Wolves KODIAC PRIDE @ 6/5 B365
-
5.50 Wolves JOES WAY @ 2/1 B365
-
-
seems a bit of a naff day today on the racing front, Ah well here's my usual NR or placed horse for the day
South 1.15 ROBYNDZONE @ 6/4 B365
res
Leic 2.30 MONZA CITY @ 15/8 B365 -
-
37 minutes ago, Alley Cat Glover said:
I assume you were already heavily into maths and statistics before this series of projects or was it all self taught.
Engineer (toolmaker by trade ), then moved in to non destructive diagnostic maintenance ( thats a lot like the condition monitoring system on your car) you fix things before theyre due to break so as not to cause even more damage,
Self taught , I had over 12months having to drag an oxygen cylinder everywhere so I tended to sit down a lot at my computer![:\](https://forum.punterslounge.com/images/smilies/ohwell.gif ":\")
QuoteBut it’s way above my capability. Certainly now.
No it isnt , If you want it bad enough and set your mind to it anything's possible , just do things in baby steps, you dont need to learn everything in one go
-
1 minute ago, Alley Cat Glover said:
Is this a database that you have personally built up over numbers of years for you to do all your research with. Do you do your own inputs or download race results, or do you still Watch the races and give your own ratings. Only asking please go gentle with me. No way do I do that much work, I personally try to watch previous races and try to find an angle which I like, just trying to improve. Some races are a lot easier than others. Seems covering your options by backing 2 or even 3 selections in a race can still be profitable which seems counter intuitive, judging by some very reliable contributors on here.
Its my own ACG
Built up over the years and it started out as a simple project to keep me occupied when I had to spend a long time at home after serious illness but ended up occupying nearly every spare minute I had even after I had recovered
Once you see a light at the end of the tunnel unfortunately it becomes an obsession( or it did with me)
This is how it sort of escalated
Project 1 many years ago ,There were plenty of databases you could purchase but they could only be used the way the seller wanted you to use it (ie tick box for this, tick box for that crap) , I wanted a personalised one that could do things my own way so learnt how to create a database to run the way I wanted it to.
Project 2 was to create an optimised ratings type that I thought would be the most advantageous to my needs
Project 3 was learn to automate ratings so that everything updated as soon as new data was input
Project 4 Build an accurate formula to price up the race according to my ratings and build into rating result
Project 5 was the easiest bit as all the hard work had been done, as soon as I updated the days racing the algorithm picked the top rated horse and price up against the rest of the field (top rated in my ratings opinion is the only 1 Im interested in after all its supposed to be the best option in the race )
Project 6 (Hardest part of all and took me what felt like a lifetime) Learn machine learning and get it to read all the stats created and optimise ratings for the best bets.
Project 7 Get myself set up on a VPS (virtual private server) and get my ML prog running on it and connect it to betfair so it runs 24/7 even when my computer at home is off and has next to no downtime (unlike virgin)
Project 8 Load up with some cash and let it run unattended backing MY top rated for a couple of months at min stake to test for bugs (if certain conditions are met of course)
Overall
I run three systems / methods / algorithms/ ratings whatever you want to call them.
2 of them I tend to back manually for various reasons and 1 runs on its own on a VPS and backs whatever it decides is the optimum bet and meets all its criteria , its built to keep running the stats and checking the prices untill 15 seconds before the off ,then decides if everything is the best it can be , if there's a bet it then triggers one, if not moves on to next race.
I have no input whatsoever though I can check on VPS site to see whats happening but as Ive had it running now for a few year I know its stable enough so just leave it be and update the results after racing.
As for backing 2 or 3 in a race if its statistically viable for a bet then why not (ie If one of them wins and you can make a designated profit in line with your percieved odds chance). If others disagree thats there prerogative
Very rarely watch a race unless theres zip on the telly or Ive nothing better to do (Like a lot of stats players,I know very little about horses, I only play the numbers)
hope thats answered your Q
-
6 minutes ago, The Equaliser said:
I guess I do feel for the poor Dear Deardrie idiots that you mention. It is my quest to help them when I can.
Some geezer named Jesus had the same quest some time ago apparently and look what happened to him. (allegedly)
-
29 minutes ago, The Equaliser said:
Let's now say that you VT is the benchmark of stats that we all should follow
Im not asking nor do I want anyone to follow me in any way shape or form (if you need to follow someone , I suggest join a marching band its a lot less stressful than punting)
31 minutes ago, The Equaliser said:Hence the question on that point is that if you shared your method with other punters
I find as in most things life people who constantly go on about sharing usually have very little to offer in return and just want to take or are too bone idle to put in the time and effort do it themselves, either way not my problem
44 minutes ago, The Equaliser said:would they have made a level stakes profit over the last 5 years. If so how much?
You'd have made a profit @ level stakes over the last 20yr , How much Ive made has nothing to do with my wife never mind some random poster on a message board.
As I am not trying to sell anything or prove anything to anyone whatsoever then I find the question particularly bad etiquette and usually the sort of thing akin to the dickswingers who've had a winner and feel the need to tell every-man and his dog. If you want that sort of crap go to BF forum full of them57 minutes ago, The Equaliser said:Now contrast this with the all too many private services out there that are fleecing punters on the basis that there wonderful rating services
That's the problem of the people who sign up to these clowns not mine , Im not a Dear Deardrie for the idiots who lose money to these shysters its there money to do with as they wish.
-
-
44 minutes ago, The Equaliser said:
Yes, I get that but what is the point of using stats or top rated selections if one is still having to plough through all the other horses in the race before coming to decision as whether to take notice of them. A waste of time and expense. If one is going to do that then much better just to focus on the leading market contenders and work from there
If you keep dissing stats and say there's no point then why keep asking.......... ?
Q1
QuoteYes, I get that but what is the point of using stats or top rated selections if one is still having to plough through all the other horses in the race before coming to decision as whether to take notice of them.
As I stated before (so tedious having to repeat oneself) I have a ML program linked to my Access DB which does it all for me in minutes.
Q 2
QuoteIf one is going to do that then much better just to focus on the leading market contenders and work from there
Its because MY market leader from MY stats MAY NOT be the newspaper / racing post / tipster market leader which are widely followed and usually over-bet by the sheep, and therein lies the opportunity for profit
-
5 minutes ago, The Equaliser said:
Haha, I much prefer your previous reply that people should use statistics sensibly. E.g It may correct to say that 33% of favourites win on the AW over the summer months and in most years that they show a profit. However, placing a bet blindly on the favourite will not necessarily yield a long term summer profit. And, suppose someone thinks that getting 4/1 plus about a favourite is bound to make them a profit in the long run because the dumb bookies are offering 20% odds about a 33% chance is still wrong. This is because mixed up in the 33% winners figures are odds on shots and even money all the way through to 2/1 (33%) shots. So one has to remove all those winners and look at the stats again to find a value bet.
I think if I were to pin my hope on such stats I would look at the favourite, But not look at any of the opposition horses. Then, I would consider all aspect of this favourite under today's conditions e.g recent form, reliability, whether race fit, if the going is OK, the Mark OK, the Distance OK, the Course OK, the Class OK, the Surface OK, Under the right racing code and whether the trainer/Jockey is in good form as well. I would apply the same approach to using ratings methods
However, this is not my approach to choosing betting selections. I prefer to focus on either horses at the front end of the racing market and use expert opinion both for the well fancied horses and also for horse selections around 10/1 plus in the betting.
I don't like the idea of my fellow punters placing bets on stats and/or top ratings, in this regard, I agree that they should work at tuning their harps
Do 2nd , 3rd , 4th favs etc not have stats ?
I stated stats should be used correctly , which would incorporate all available to create the best overall stat for top rated/fav whatever you wish to call it. -
23 hours ago, The Equaliser said:
Hence, what sort of advice would you suggest to punters for them to carry out in addition to pure stats/ratings before putting their hard earned cash on a horse?
A harp player spend 90% of the time tuning their harp
and only 10% of the time making musicTherein lies your answer,put the work in reap the rewards
-
5.30 Winds DIVA KAREEM @ 3/1 B365
RES
1.10 Salis KODIAC PRIDE @ 11/10 B365 -
1.15 Ascot ATASER @ 4/1 B365
-
2.55 Sand MON CHOIX @ 5/4 B365
-
Certainly getting your moneys worth with this fella,seems to be running every other week
-
5.50 Strat DEMON D'AUNAU @ 11/4 B365
RES
5.40 Bath LISMORE @ 11/10 B365 -
1 hour ago, The Equaliser said:
I kind of agree with Trotter in as much as I don't like applying statistical data to a specific event such as the 6.40 at Kempton. However, what you say seems to make good sense about using statistics intelligently. Many people know that tossing a fair coin thousands of times will lead to results tending towards 50% heads and tails. What they don't seem to appreciate is that the numerical difference between either heads or tails grows larger as further spins are added . If you have written anything on using horse racing statistics intelligently I will be please to receive a copy. Then maybe I could go on the Adrian Massey website and make some good use of it.
No I havent written anything on using statistics, there are a lot bettter informed people than myself on the subject, what I would say is only learn what you need to know that concerns your method of betting the rest is irrelevant and take no notice from books written from ex-punters or racing media "experts" ,they are nothing better than failed tipsters, as they say those who can do those who cant write a book about it .
Books by academics on statistics are a lot more informative , one I would recommend is Winning at the Races by William Quirin PhD deals with impact values which is something Im interested in.I have a question for you? I was looking for a decent "lay" bet earlier and came across the 6.35 at Lingfield. There was a horse in it called O'Reilly's Pass. For quite a large part of the afternoon it was quoted as around 9/4 to 5/2 favourite. The only information that I could glean was something like "hasn't done much up to now but should improve now handicapping". There were other horses in the race that had showed some good form. Having looked at later races I decided that it would be worth laying it as it seemed to be my best option. Anyway, my question is how does anyone work out that if sending a horse handicapping should actually show such dramatic improvement so as to make it favourite over the other horses?
I dont do laying I think it should be banned to punters, but I cant see why OP should be fav , its a pretty naff race and OP's hcp mark may be considered a bit low for his first run in a hcp and bookies are always wary of that as its an unknown to them (three crap runs low handicap mark, good price , money down, BOOM gamble landed) and if any money from connections has come in it wouldnt take much to shift the price.
-
38 minutes ago, Trotter said:
Not sure how that helps with trying to find the winner of the 6.40 at Kempton ...... does that belong in the third of races that favourites will win or the two-thirds that they won't ? ....... after the event you can put it in it's category but not before the race
Its all about using the correct statistics which most don't
If the race is a 3yo+ 0-60hc you use all available stats from that type of race as all horses then should have a similar makeup otherwise they shouldn't be in that type of race,
I personally then try to break my selections down to a coinflip 50/50 chance so say the past stats, eg... say 1st & 2nd favs win 50% between them against the field Ive already reduced the field to 2 by doing nothing with the knowledge backing all selections these stats bring up Im at least going to win 50% of the time, if I can dutch the 2 for over evens the pair then its a no brainer back the 2, If not price match those 2 against each other to see if there's a bet or not.
Luckily for me I have a machine learning program linked to Access that sorts out the best bets for the day then all thats needed is the books prices to see if any are worth backing.
In the database I have 149K flat races
Which is over 1.5 million horses raced against each other
-
5 hours ago, Trotter said:
I'm not really a fan of those kind of stats
My view is always that those 5 years worth of favourites that the stats are based on are not running in the 6.40 at Kempton tonight !
But overall I think that about roughly a third of favourites ? ....... which means two-thirds don't
Just a 1 question
if you were following a tipster which would you say is more statistically significant
tipster 1 £100 profit from 1 race
tipster 2 £100 profit from 500 races.@BillyHills once asked me why I disliked 10yr trends, its because in effect they are not 10yr trends but 10 race trends and therefore just chaotic events akin to your 6.40 @ Kempton
But whenever a large group of chaotic events come together and are placed in order of their regularity, they are then no longer chaotic but become the norm, without this planes wouldn't fly, engines wouldn't work, life wouldn't exist (unless you believe in some monotheistic deity that is ) -
12 hours ago, Valiant Thor said:
3.40 Ayr GREAT COLACI @ 11/2 B365
I see my penchant for picking NR's strikes again (better than coming the usual 2nd I suppose
)
RES~1
5.20 Sligo LOUDEST WHISPER @ 11/4 Coral
As this is a race reserve horse and has no designated jockey atm I better put another res in for my res
RES~2
5.50 Sligo ROYAL AIDE @ 5/2 B365 -
As the majority of races fall into10-11-12 runner races Most rating or ranking systems given enough data the combined average success rate will fall within the ballpark figure %ages below using this formula LOG10(1+1/d).....d = rank position
RANK 1 2 3 4 5 6 7 8 9 10 11 12 %age 30% 18% 12% 10% 8% 7% 6% 5% 5% 4% 4% 3% this goes for RP ratings ~ forecast odds ~ fav win rates ~ Top rated in handicaps ~ etc etc
saves a lot of back checking only to come to a similar statistical conclusion.Note for the usual pedants ....
Ballpark means within a reaonable margin of error so dont be posting I got rank 1 as 31.5 % where you state 30%...
-
Naps - Thursday August 13th
in NAP of the Day Competition
Posted
6.30 Bath BAL MAL @ 13/8 B365