Jump to content
** March Poker League Result : =1st Bridscott, =1st Like2Fish, 3rd avongirl **
Download the PL Forum App!
** Cheltenham Tipster Competition Result : 1st Old codger, 2nd sirspread, 3rd Bathtime For Rupert **

Valiant Thor

Regular Members
 View Profile  See their activity

  • Posts

    4,023

  • Joined

  • Last visited

  • Days Won

    15

 Content Type 

Posts posted by Valiant Thor

    Racing Chat- Saturday 14th August

    in At The Races - Racing Forum

    Posted · Edited by Valiant Thor
    spelling

    On 8/14/2021 at 7:55 PM, MCLARKE said:

    With odds of 33/1 you could expect a losing streak of 846. I'm not sure I'd have the stomach for that.

    You may want to re think your calculation there , its a tad out to say the least.

    Even if he had 1,000,000 bets @ a true 33/1 ( 0.029%) there would be an approx losing run of 463

     

    So for 1 true 33/1 bet a day for a year a losing run of around 200 would be expected.

    As 33/1 shots only win around 0.018% ( 54/1) ish then a losing run of around 325 would be expected .

    Still a lot in terms of bank required unless your playing for pennies ?

     

    RACING CHAT - Thursday 10th December

    in At The Races - Racing Forum

    Posted · Edited by Valiant Thor

    2 hours ago, MCLARKE said:

    Unfortunately it's not.

    It's a lot easier to make money on the stock market than with horse racing. The average investor on the stock market will make money, the average punter on horse racing will lose money.

    The rule that works well on the stock market is the 200 day rule. When the share price is above the 200 day moving average then it is time to buy, when it is below then it is time to sell.

    The problem with trainers is what period to use, x days or weeks or x number of runners.

    I might do a bit of research unless somebody already has an answer.

    I think you'd want to re think that, your buying @ approaching peak , the majority of profit has gone.
    Ok if your a trader playing with other peoples money and taking of a cut per trade, but as an investor its a big no no as the next step is on the way down do the math , where can a stock go from peak certainly not upwards.
    They better take the Shiller list off all the trading markets then if thats the case as they've been doing the PE ratios wrong since the 30's.

    To quote Warren Buffet

    Quote

    The first rule to making money is not to lose money

    Exit when the PE's are high and enter when the PE's are low
    Im led believe he has a good idea what hes doing.

     

    RACING CHAT - Thursday 10th December

    in At The Races - Racing Forum

    Posted

    12 hours ago, MCLARKE said:

    Another 2 losing favourites. 

    I should perhaps look at his current form as well as his past record.

    How do we define that a trainer is in / out of form ?

    You say you make money on the stock-market, then why are you asking the above question,the answer should be blindingly obvious.
    I suggest reading the Neiderhoffer / Osbourne paper "The Magna Carta of High Frequency Trading"
    Then use Robert Shillers 10yr moving average on the trainers PE value and therein lies your answer ;)
    At the moment you are in effect buying-in  @ at the high PE ratio which is a no no :ok

    10 ways to improve your punting

    in At The Races - Racing Forum

    Posted

    10 hours ago, The Equaliser said:

    Further, you will see that a horse called Highland Avenue got beat by a neck at 1 to 4 in the 4.55 Kempton yesterday (previous price 13/18 PL). It was expected to run well and win however I don’t think it was a true 1 to 4 shot.  Looking at the table you will see that 1 to 4 shots are supposed to only lose 4 times in a 1,000 bets.  I am pretty sure that I have seen much more of these short-priced favourites turned over in the last 1,000 results. In my opinion, the bookies have deliberately shortened the price to mitigate their liabilities. This is the main reason why I don’t trust any odds offered at shorter than 4/6 ( I don’t even like these that much).

     

    1/4 shots have an implied probability of 80% therefore should  lose approx 20% of the time
    In 1000 bets there will be approx 200 losers not 4
    Its the 200 losers that have an Implied probability losing sequence ratio of 4 (4.29) in 1000 bets
     

    10 hours ago, The Equaliser said:

    Finally, @The Brigadiersaid something like that if he fancies a long shot he has found it best to put the same amount of money on that he would do so on a horse at a skinny price. 

    I believe that's the same advice advocated by Dave Nevison (Failed Gambler,failed tipster) in one of his books who now sells his wares on Racing TV as some sort of expert mustn't be paying him that much as he looks like a tramp whose just slept in a skip when Ive seen him on tv.
    Mind you John McCririck made a decent living out of talking crap and he was a failed bookie as well so couldn't do it as a pundit a poacher or a gamekeeper.

     

    GENERAL RACING CHIT CHAT

    in At The Races - Racing Forum

    Posted

    2 minutes ago, Trotter said:

    No, that's not the one I was thinking of ...... my one was definitely since I came back to racing after a dozen years 'break' and that was in 2008, so it's since then

    The one I'm thinking of might have won group races or at least run in group and listed races

    Material World...... ran in 2008 World Hurdle @ Chelt

    Racing Chat- Wednesday 2nd December

    in At The Races - Racing Forum

    Posted

    2 hours ago, Villa Chris said:

    Obviously the more horses in a race the less % your selection has of winning, that’s if you rate up all the horses in the race. Wouldn’t it be more accurate to get rid of the horses that you think have little to no chance of winning.  

    Some races are competitive where all have their merits, but there are races where some horses are literally just making the numbers up. 

    All horses have some chance of winning  however slight it may be,by omitting some horses you would create a false market
    The no hoper could be the faller that brings the fav down.

    Quote

    Would that give you more accurate odds for the horses that are left in the race?

    In a word ...NO

    Racing Chat- Wednesday 2nd December

    in At The Races - Racing Forum

    Posted

    43 minutes ago, MCLARKE said:

    I think your definition of basic stats and others definition might be a bit different !

    You are right though, the more work you pit in the better you will get. I've been doing this 40 years but still learning and improving.

    I still can't price a race though.

     

    I think your confusing basic maths with basic stats totally different

    Racing Chat- Wednesday 2nd December

    in At The Races - Racing Forum

    Posted · Edited by Valiant Thor

    4 hours ago, Villa Chris said:

    I had her down as 4/1 in terms of what her odds should have been. Whether I should have taken lesser odds is another story.  Hopefully in the long run it pays off, but I’m new to this waiting around for a horse to get to certain odds before backing it. Usually I’d just back my selection but I’m trying to find value in my bets along with trying to do ratings. 

    I wouldnt worry about it too much if I were you, I know people who cant even price up a £5 note, never mind a horse-race despite 20+ years of trying.
    I had it priced @ 9/4 and even using a cheap and nasty method using RPR's (below) it was a  5/2 shot, so I thought you were a bit OTT looking for 4/1 but practice makes perfect. ?

    image.png.510fbe7212a321b54e3975dff4a74a72.png

    (anyone with basic stats should be able to work the above out and how I came about it) :ok

    4 hours ago, Villa Chris said:

    Bit annoyed with myself but not been in great form lately so maybe it was a confidence thing too.  Put the selection in the nap of the day, so at least I’m off the mark for the month in that .

    If you want to do what you say you want to do dont get disheartened if things dont work out straight away, & dont do what 99% do and take the easy way out and throw the towel in after a few week and go back to knee jerk selections.
    The more work you put in the better you'll get its not an overnight project
    This is regarded as the BIBLE OF FORECASTING I bought it many moons ago and is an excellent book quite expensive now but if you can get hold of one cheaply its highly recommended , plus bin the pen and pad crap , sort yourself out and get excel its soon to be 2021 you know ?
    ATB
    VT :ok

    VT's NFL Selections

    in NFL Predictions

    Posted

    WEEK 11
    Think Ive found the problem, The date and time for the games downloaded from 1 site were causing an error when looking for the odds on another site due to time/date of game being changed or in a different format. (times and dates for the games have been published in advance and any changes are beyond my control)
    Anyway see how it goes .
    **Weeks 9 & 10 will be missing from the Season to Date results as no selections were made due to no spreads being recorded.**
    image.png.67d9859e76c60e2689cda440a7c13437.png
    Just the NAP bet this week

    GENERAL RACING CHIT CHAT

    in At The Races - Racing Forum

    Posted · Edited by Valiant Thor

    3 hours ago, Villa Chris said:

    To anyone that does their own ratings. Do you take into account course form? If so I wondered how that works out. For example  you have an 8 horse race. 3 of them have ran at the course before. The other 5 haven’t.  Out of the ones that have only two of them have ran at the course more than once. Doesn’t this scenario leave blind spots, especially if your top rated hasn’t run at course before and second rated has . 

    Its down to what weighting the cd winner has or has not got over non cd winners

    Lets look at some imaginary data from 295 10 horse races (2950 runners no dead heats so 295 winners)
    From the 2950 runners we have 550 previous cd winners of which go on to win 65 of these races (11.8% sr)
    That leaves 2400 non previous cd winners that win 230 of these races (9.6%)
    image.png.7d7de86a6e3985af02f140244b78d096.png
    So to find a weighting for a previous cd winner you divide the races cd winners 65 by races total winners 295 =.22%
    then divide total previous cd winner runners 550 by the total runners 2950 = 19%
    Finally  divide cd winners total 22% by runners total 19% = 1.18 which is your weighting (1 being the mean)
    So by being a previous cd winner you are 18% more likely to win than a non cd winner (1.18-1=18%)
    so therefore when rating a race and 1 runner is prev cd winner then you multiply your base rating by the cd weighting 1.18
    All being equal a weighting of 1 is the mean so >1 positive rating <1 negative rating
    So if your base rating was 100 then a prev cd winner would have an improved rating of  100*1.18= 118
    Obviously the rating would be the inverse if it was a negative weighting of say 0.84 then it would be 100*.84= 84
    this can be done for multiple factors , jocks ,trainers,beaten favs, etc etc once you find the important ones then your rating start to take a shape which can the be used to produce a viable tissue .
    hope this helped a bit.
    ATB
    VT :ok

     

    Paddy Power Gold Cup

    in At The Races - Racing Forum

    Posted · Edited by Valiant Thor

    4 hours ago, richard-westwood said:

    The paddy power has always been a strong trends race .....

    My trends ...

    19/20   aged 6 to 9      coole passed 

    8/20   won at chelt previously   coole won over hurdles 

    Pref below 11st 2lb     coole was 10st 2lb 

    Only reason I didn't back was I just didn't think he was good enough based on recent form ..he had to improve  ....obviously the 10st 2lb helped him a lot on the ground receiving so much weight from the others....it happens....but he was definately on the shortlist after the trends chaff cutting ......I was gonna back him purely because he had passed trends and was top of time form ratings but was going to never did ......still I was happy with 3rd in such a hard race ?

    Oh ....I think he passed another trend   16/20 .....a previous winner at over 2m4f plus .....albeit hurdles (2m5f) so he was very much a trend horse ....like I said he had got to improve a lot .....but he did ....that's just life 

    19/20   aged 6 to 9 ..............80% of the field fell into that category so a blind man in a dark room could have seen that
    Pref below 11st 2lb...............So were 50% of the field
    16/20 .....a previous winner at over 2m4f plus.......As It was a valuable 2m5f race , I would at the very least expect the horse to be competitive at that distance or why else would they be entered , for the fun of it ?
    8/20   won at chelt previously......from 1079 pre race CD winners in the last 12 month 1743 won (13%) so sort of blows the 8/20 (40%) out of the water , when taking large data into the equation, Chelt or not.
    You forgot to mention only 2 greys have won it in the last 10yr , so disregard them and Coole Cody falls into that trend as well.:eyes

    See where Im coming from its a kind of horse-race pareidolia ( seeing patterns that aren't really there when you look at a bigger picture)
     

    10 hours ago, MCLARKE said:

    You could well be right, I don't normally look at trends for individual races, perhaps I was just lucky on this occasion.

    How do you define a similar race, is it class, going, distance, field size etc? What would you use that gives an adequate data size and would cover the Paddy Power ?

    1) I dont agree with luck good or bad..... Its a binomial event , one man's bad luck is another man's good luck so the result is always balanced, good/bad = 1 and vica versa.Its more of an excuse used by a loser rather than an admittance of a wrong decision.
    ie ...Boo Hoo Im having bad luck my 1/5 shot just got beat on the line (1/5 shots lose 17% of the time get over it and stop whining)
    2) As for data size , as much as I have of what you have stated ,plus Jock ,trainer,horse etc etc.
    You cant base a decision on 10 or 20 events as its just statistical noise, jock A might only have had 10 races at that distance, but all the jocks racing might have had over 1000 runs between them from which you can then make a more accurate average performance required put that up to the jocks riding in said race and it shows who is above or below the average required (just an example but get my drift).
    3) Mine is not a static database (just the results) its dynamic, meaning the result of one race effects the other races using weighted averages ( a sort of Butterfly effect in reverse, time series exponential smoothing to be precise ). these weighted averages are then pitched against each other ie jock A v jock B etc etc same with trainers , class , field size etc etc the product of which are then used for the final horse ratings.
    This creates a standard or normal distribution (Bell curve) from which z scores can be obtained and there in a reasonably accurate tissue can be made according to the final ratings value (IMO more accurate than the rating being divided by the sum of the ratings or guessing, hence the exp odds by my ratings are more important than the rating itself) The 100 for top rated is just an arbitrary number as I like all my top rated to be the same (its an OCD thing)
    Sounds a ball ache to do but once automated takes about 1 min per race so by the time Ive set it going and made a brew its done for the days racing
    ATB
    VT :ok

    Paddy Power Gold Cup

    in At The Races - Racing Forum

    Posted · Edited by Valiant Thor
    spelling

    44 minutes ago, MCLARKE said:

    A bit harsh Valiant !

    I do agree though that it is better to have a lot more data to draw meaningful conclusions. The trick is to have enough data whilst keeping the parameters of the races fairly narrow. 

    For instance should we expand the data to include races during November or other months, other distances, other classes etc.

    Currently my most profitable system is based on an analysis of all NH handicap races. This seems far too wide a parameter but it seems to work.

    It might be interesting to look at the top race each Saturday and see how successful the trends horses are compared to those selected by more traditional form study. 

    I dont think its harsh,
    The average horserace has around 9 runners (11%)
    A:4 wins from 10 runs = 40%
    B:72 wins from 400 runs = 18%
    A is trend data , B is race data from 400 similar 9 runner races ....which is a more accurate representation of a similar type race A or B?
    B is 10 times LESS likely to occur at RANDOM than A.
    Therefore is a more accurate representation of a likely outcome.
    Punters can do what they want as far as Im concerned its not my money, but if punters want to improve do your homework and dont fall for the crap .
    If they want trends just listen to the Blonde bimbo and her sidekick Jamie on ATR , Blah Blah Blah has a 33% sr at this track , you look they've had 3 races and won 1 whoopie f'kin doo lets pile on after all its a trend.
    If you want to check the trends go ahead , infact I think theres trend thread on here somewhere,but it stopped when they realised it was a pile of crap.
    Only trend Im interested in is how each horse could/should perform against each other in that particular race , Dobbins performance from 7 years ago has no influence whatsoever unless he/she is running in it.
    I bet the trend followers buy lotto systems as well :lol

    ATB
    VT :ok

×
×
  • Create New...