1. This site uses cookies. By continuing to use this site, you are agreeing to our use of cookies. Learn More.
  2. There seems to be a problem with some alerts not being emailed to members. I have told the hosts and they are investigating.
    Dismiss Notice
Welcome to the UK Betting Forum. Please consider registering, it is free!

Early days

Discussion in 'Ratings & Utilities' started by davejb, Jul 1, 2017.

  1. davejb

    davejb Filly

    Joined:
    Apr 4, 2017
    Messages:
    287
    Likes Received:
    300
    upload_2017-9-27_10-55-6.png

    There we go, seems the 9f standard time has dropped 1.0s and the 7.5f was 0.2s different...
    Dave

    Very odd - I logged back in and couldn't see my earlier paste job of the revised times, so I edited to add the table and now there are 2 there.... let's hope this is the final tidy up needed to get this posted and coherent!!!


    @TheBluesBrother There it is, I redid it. I must admit I was quite surprised to get so close to your figures Mike, as I wrote above I had slightly different standard times at 7.5 and 9f to those used on the RP results page, after updating my standard times file I reran the results and got the above.... which pulled the allowances even closer (I must admit to a wry grin when I saw that happen, I was quite happy with the original until I spotted our comptimes were slightly different for the round course).
    I do admit to being a little bit wary of these figures, whilst it's blatantly obvious that the first couple over the straight 5f were way faster than the rest of the card I find it hard to credit that the difference would be so huge - away from the straight it must have been like glue to slow them down so much extra.
    Dave
     

    Attached Files:

    TheBluesBrother likes this.
  2. davejb

    davejb Filly

    Joined:
    Apr 4, 2017
    Messages:
    287
    Likes Received:
    300
    Tomorrow's cards attached.

    Yesterday - gloomy again, 2/20 handicap winners at 7.2 and 6.42, 1/13 non handicap at 5.4, 0/6 all told over jumps. Using the filters that becomes 1/12 in handicaps at 6.42, and straight zeroes for non h/c flat (0/3), h/c jumps (0/3) and non h/c jumps (no selections).

    Dave
     

    Attached Files:

  3. davejb

    davejb Filly

    Joined:
    Apr 4, 2017
    Messages:
    287
    Likes Received:
    300
    Daily cards update etc
    Yesterday Flat 2/11 handicaps at 2.68 and 5.43
    2/12 non handicaps 1.82 and 2.23
    1/5 jumps handicaps at 1.22
    3/6 non handicaps 2.12 2.27 and 2.14
    Filters 2/4 handicaps on flat 2.68 and 5.43
    1/4 handicap jumps at 1.22
    no selections for non handicaps flat or jumps
    So, pre filter 34 out 19.91 back, after filtering 8 out 9.33 back
    Dave
     

    Attached Files:

  4. davejb

    davejb Filly

    Joined:
    Apr 4, 2017
    Messages:
    287
    Likes Received:
    300
    Daily stuff...
    the main card spreadsheet is a bit big, as there are 7 meetings on the go. The selections.txt has a lot of handicap picks on it, I still haven't found a way to reduce the numbers effectively whilst keeping the winners, all I'll say is that some/most days the win rate is about 1 in 6 on average over the past 5 weeks or so, during which there'd have been 363 qualifiers - so if you plan on using the ratings to pick winners then (a) using the filters - ie sticking to the ones that appear in 'selections.txt' each day - is eminently sensible as you'll miss a good few losers but few winners, and (b) have some means of filtering the selections further that can also reduce the number of bets.
    The non handicap flat selections have returned 49 winners from the past 112, which is actually a lower rate than was evident in the mid summer period.... at one point the hit rate was up near 70%.

    Yesterday - flat handicaps 2/12 at 3.3 and 2.63, filter version 1/5 at 2.63
    non handicap 2/9 at 23.41 (yes, nice odds) and 5.3 - no filtered selections
    jumps handicaps zero/7, filter 0/5
    non hcap 2/7 at 3.03 and 9.6, filtered 1/1 at 3.03

    Still watching the last few results coming in, for a change the handicaps have been doing quite well - so far there's 3 winners from t 11 on the flat at decent odds with one still to go, and there were 3/6 over the jumps. Non handicap flat is up a little with 1 winner from 2, 1 from 2 over jumps is a loser due to the winner being odds on....about 50% up on the day, makes a nice change lately!
    Dave
     

    Attached Files:

  5. davejb

    davejb Filly

    Joined:
    Apr 4, 2017
    Messages:
    287
    Likes Received:
    300
    Tomorrow's cards as usual....
    A small change to the selections.txt file, a small improvement to handicap win rates seems to be possible using the fairly simple idea that the horses higher in the handicap tend to win more often than those at the bottom. At the moment I have set a second filter to show the handicap selections that are in the top half of the field for each race - it may be that this would be better set at a different fraction, it may benefit from a bit of a tighten up, but for now I'm satisfied that the result of this second filter is to improve results whilst it seems to work happily alongside the market filter (it can be that you identify two factors such as this, which individually improve results, but they turn out to clash - one selects what the other throws away, that sort of thing, so it's important that any additional filters work in conjunction with existing ones).

    So, FOR HANDICAPS there is an extra symbol on the selections listing, an asterisk * indicates that the selected runner is in the top half of the handicap for the race in question. It is possible to have the ! symbol indicating 'less than 80% of runners rated' as well as the 'top half of handicap' * symbol, or just one, or neither. On tomorrow's listing for Musselburgh for example Raselasad has the * symbol, so he's top rated, in the top half of the betting and in the top half of the race weights.
    The same for See the Sun and Cricklewood Green at Epsom. Invermere and Jonny Delta have neither symbol, so they are top rated and in the top half of the betting, but they are in the lower half of the race weights (and considered, therefore, less likely to win).
    At Tipperary Timi Roli has both a * and a ! by their name, showing this horse is top rated, in the top half of the betting, but less than 80% of the runners in that race have a current rating.
    To remind those looking at this stuff - ! means 'less than 80% in race have been rated' - this does not mean the rating is considered dodgy, or the top rated isn't likely to win, it does mean that if you want to bet on it you need to look carefully at the opposition and satisfy yourself that there isn't some unexposed type in there that is going to wipe the floor with their exposed opponents.
    In the ideal world a runner with the * symbol should indicate a slightly improved chance over those without - in backtesting the difference has amounted to about a 5% improvement in win rate, it also just about halves the number of selections which tends to make life easier also.

    Yesterday - 4/16 flat handicaps at 4.27 4.5 5.26 and 8.8 filtered to 3/12 losing the 4.5 winner
    non handicaps 3/14 at 3.7 7.6 and 2.65 filtered to 1/2 at 2.65
    jumps handicaps 3/7 at 3.7 4.34 and 3.5 filtered to 3/6 (same 3 of course)
    non handicaps 2/7 at 1.41 and 2.39 filtered to 2/3 (same winners)

    So pre filtering it's 44 out and 43.32 back, after filtering it's 23 out and 36.05 back, nearly 60% up is fine in my view....the filters aren't spectacular but they are consistently improving returns significantly.

    I've been watching the racing today, I enjoyed Clemmie's race, I wasn't too wild about the rest to be truthful - I appreciate the history behind major races like the Cambridgeshire but find it exasperating when my wife asks me what I fancy - as if anyone has a hope of picking the winner in what resembles an outtake from the 'Charge of the Light Brigade'.... all they need is David Hemmings galloping up and down the running rail while the charge heads for the post. (Apologies if that's too ancient a reference for some).
    Tomorrow I am somewhat torn by several of my favourites running against each other, Enable isn't that high in my ratings (she's tended to pulverise her opponents at less than blistering pace, but you can hardly mark her down for winning by miles slowly, can you?) I have Ulysses above her, but I think he'll find himself running low on gas towards the end, I always liked Winter but I don't think the O'Brien runners are going to find Enable anything other than bulletproof. I'm thinking of a win bet on Enable and an each way on Order of St George, in a small way.

    As Marsha is out for a historic double, she's a club horse (my wife's in Elite) and destined to be sold - this should be her last race sadly - so I'd love to see her go out on a blistering high. She needs holding up though, and she'd be much better on faster ground, so I'm worried she might not get the run she needs and might find the going hindering her acceleration in the final furlong. I'm a bit bemused by the comments on TV about Battaash - the inference seems to b e that if he'd not got excited before the Nunthorpe he'd have beaten Marsha... Marsha beat a very much in form Lady Aurelia, there was nothing flukey about it, I think both fillies had Bataash's number regardless. (I also think it's not a given that he'll behave any better tomorrow).
    So I'll be cheering Marsha on too - but if I have a bet on her it'll be a small one!

    Still waiting for the rest of today's results, I thought I'd mention how that * indicator did on handicaps...it's only one day's racing so statistically insignificant, but as I managed to program it in for tomorrow's cards I went back over today's races manually checking what we'd have seen if I'd managed to get it done a day sooner. So this is working through the selections.txt I posted last night, but with the added info about which selections would have been asterisked...

    4 x NH handicaps would have turned into 2 selections - all 4 lost, so no extra back but 2 less losing bets is okay.
    Flat handicaps - 16 selections, 1 NR so 15 left. 1 is still to run but wasn't asterisked. Out of the 15 selections left, a total of 8 were asterisked. There were a total of 3 winners, at 5/2, 5/1 and 7/2 out of those 8 selections. So far NONE of the non-asterisked selections has won (in fact only one has placed)... and there's one non asterisked runner still to go in the 9.20 at Chelmsford.

    I would be overjoyed if this kind of thing continued, I suspect that like a winning system it may stop working as soon as I decide I'm really getting somewhere, but them's the breaks.... today has worked better than expected by some way, I expect a small improvement, not almost a doubling of the win rate as happened today, but anything that reduces the number of bets without damaging the returns is good in my book.

    That's plenty of waffle for now - I shall now go and make myself an Enable/Marsha party hat for the celebrations.... (no, not really).
    Dave
     

    Attached Files:

    TheBluesBrother and pete like this.
  6. davejb

    davejb Filly

    Joined:
    Apr 4, 2017
    Messages:
    287
    Likes Received:
    300
    Tomorrow...
    After yesterday's epic this will be rather shorter. I had the small win bet on Enable, sadly my small win on Marsha promptly used the winnings up.... still, at least I got evens and a thoroughly enjoyable afternoon glued to the TV. Hard luck Marsha, it's a shame she couldn't beat Battaash - who I have to admit won very impressively - but she's been a top runner for the club and I hope she goes on to have some equally speedy offspring in the future. Nice to see Rhododendron win also, I thought.

    Yesterday - flat handicaps 5/21 at 4.0 7.66 6.86 5.38 and 4.0 again. Filtered (market) to 3/15 at 4.0 6.86 and 5.38 using both filters see last night's post.
    non handicap 2/13 at 2.02 and 1.79 filtered 1/2 at 1.79
    Jumps handicaps 2/7 at 6.8 and 22.15 (not bad odds there :) ) filtered to 0/4 oops
    non handicaps 1/6 at 3.4, 0/1 filtered

    Dave
     

    Attached Files:

    TheBluesBrother likes this.
  7. davejb

    davejb Filly

    Joined:
    Apr 4, 2017
    Messages:
    287
    Likes Received:
    300
    Tomorrow attached.
    Yesterday's efforts -
    Flat hcap 2/9 at 24.29 and 8.71 filtered to 1/4 - remarkably it was the 24.29 winner. Asterisks turned that into 0/2....
    non hcaps 2/7 at 3.68 and 1.43 filtered to 1/2 at 1.43
    jumps hcaps 0/1 filtered 0/1 asterisk 0/1
    non hcap 1/4 at 2.26 filtered to 1/1 at 2.26
    pre filtering 21 out 40.37 back, so almost 100% up for a nice change. After filtering (market) 8 out 27.98 back which is almost 250% profit, and using the asterisks we get 3 out and zero back from the handicaps.

    Please note, the asterisks are only used for handicap races, so the full rnage of bets for the asterisk version would include the non handicap filter results - ie add 3 extra pts out and 3.69 extra back.
    We'll see how the handicap returns compare between using just the market filter and using the market filter and the asterisk (top half of weights filter).

    Slight addition to both the full ratings and the daily full card spreadsheet, I've used a bit of spare room up to add a bit more trainer info to both - adding data for trainers past 14 days records, showing how many winners/runners they've had and displaying the win% also... this info is at the end of each runner's 'full ratings' line, with the percentage coloured to highlight the better percentages, and it's before the 'trainer/jockey/going' etc columns on the full racecards. It's pretty obvious when you look at the spreadsheets. I decided to add this as the trainer 0/1/2 column is okay for spotting quickly that a runner's trainer is doing okay lately, but if there's a zero there it could mean the trainer is either not running horses or doing anything from 'not desperately well' (eg 1 winner in the last 12 runners say, as it takes a 10% win rate to post a '1' in that column) to abysmal - a trainer who has had 60 consecutive losers in the past 10 days would still have a zero.... with the extra column of trainer info you can now see who's a zero due to not having many runners and who's a zero because they're running carthorses.

    Oh by the way Mike @TheBluesBrother did you get two allowances for Haydock on Saturday? I can't find the scrap I made the note on (damnit) but from memory I had the round course significantly slower than the straight...

    Here we go, redid it....

    upload_2017-10-2_21-32-17.png

    Dave
     

    Attached Files:

    Last edited: Oct 2, 2017
    jammy54 and markfinn like this.
  8. TheBluesBrother

    TheBluesBrother Gelding

    Joined:
    Feb 9, 2013
    Messages:
    1,457
    Likes Received:
    3,140
    Hi Dave, Haydock can be difficult to assess even when the going is on the good side.
    I only had one going allowance, this was mainly due to the speed figure Century Dream achieved in the 2nd race.

    Haydock.png
    CentuaryDream.png

    In the last race Royal line was an impressive winner, the time was incredibly slow, so much so, Dave Edwards even adding the 3yo (9lb) WFA allowance
    gave the horse a speed figure of 24, mine was 17 (No WFA).

    Mike.
     

    Attached Files:

    • WFA.pdf
      File size:
      29.9 KB
      Views:
      3
    Last edited: Oct 3, 2017
    davejb likes this.
  9. davejb

    davejb Filly

    Joined:
    Apr 4, 2017
    Messages:
    287
    Likes Received:
    300
    @TheBluesBrother
    Cheers Mike,
    good to have a second opinion on it. Watching the meeting on TV it looked absolutely dreadful there - I do wish somebody would invent an anti watering device and give us some decent ground for them to run on! Still, good for the fish I suppose ;-)

    Dave
     
  10. davejb

    davejb Filly

    Joined:
    Apr 4, 2017
    Messages:
    287
    Likes Received:
    300
    Tomorrow:
    Yesterday: Flat handicaps 1/5 at odds of 9.0, same winner out of 2 passing the market filter, sadly it was in the lower half of the card so the second weight filter knocked it out to give 0/1.... so that's 1 out of 5, 1 out of 2, zero out of 1 depending on which filter(s) you do/don't use***.
    non hcap 1/6 at 8.0, filtered to 0/1 (market filter only for non handicaps)
    jumps handicaps 2/6 at 4.61 and 5.16, 1/4 filtered at 4.61 (market) , 0/2 filtered (weight)
    non handicaps 1/4 at 16.89 0/2 filtered

    So, not exactly an unheard of occurrence, there are days when the filters don't help, and even less frequently it happens on q day when the unfiltered results have big odds winners in... so here we had the raw ratings throwing up a day with 21 pts out and 43.66 back. Using just the market filter we get 9 pts out and 13.61 back

    ***Using the weight filter and market filter on handicaps, and market filter on non handicaps, we get 6 pts out and sweet miss F Adams back. In the long run, or actually in the fairly short term, I would expect this combination to be the most profitable however - the raw ratings without filtering, as those following the thread will know, don't win anything like this usually, whilst the market filter has been very good at boosting returns to date. What I'm actually doing personally is as follows:

    Using the selections.txt as a shortlist, I look at the different sections and pick out the runners that don't have the ! mark against them - I DO quite often look at the card for these runners to see if in fact they look pretty good compared to the rest, for example if I saw a runner rated 85 and 20 pts clear of the second rated then I would consider putting it into the shortlist... but if it's got a ! and isn't looking pretty spectacular I'll bypass it.

    Once I've gone through the list, pretty much erasing the ! runners barring the aforementioned hotshots, I look at the * indicator in handicaps - if it's a handicap and the * isn't present then I'll remove it from my shortlist.
    I'll then look at the trainer info, and I annotate each runner with a tick if there's a positive win rate showing, if the trainer has a zero win rate then I look to make sure it's because of having a small number or zero runners showing, if i see the trainer has had a bunch of runners and no winners that's enough to score them off the list.

    After that I might just indulge in a bit of card reading, divination using chicken bones, or even reading the tips in the racing Post. (Only kidding, I mean chicken bones okay, but tips in a paper?!)
    I get a pretty short short list, after that it's personal form checking and what have you, ie I've reached the end of just letting the computer do it all - I figure it has given me a handful to consider, and as a quick way to reduce the size of the problem it suits me.

    Dave
     

    Attached Files:

    jammy54 and pete like this.
  11. TheBluesBrother

    TheBluesBrother Gelding

    Joined:
    Feb 9, 2013
    Messages:
    1,457
    Likes Received:
    3,140
    @davejb

    The correct distance for the 5:10 Sedgefield NHF race was 2m145yds.
     
    Last edited: Oct 4, 2017
    davejb likes this.
  12. davejb

    davejb Filly

    Joined:
    Apr 4, 2017
    Messages:
    287
    Likes Received:
    300
    @TheBluesBrother
    Hi Mike,
    Yep, I get the BHA distance corrections as part of my data download, I have that one as being 2m 178y with a correction of -33yds - ie 2m 145y. (You had me guessing a bit though, Catterick is Sedgefield.... :) )
    Dave
     
  13. TheBluesBrother

    TheBluesBrother Gelding

    Joined:
    Feb 9, 2013
    Messages:
    1,457
    Likes Received:
    3,140
    My mistake, yes it was Sedgefield.

    **********************************************

    I sent Sedgefield an email this morning regarding the NHF race distance.

    Morning Michael,

    Many thanks for your email. The race distances were initially posted as a 6 race card. With race 1 dividing this meant the order changed, unfortunately I didn’t amend the running order to a 7 race card.

    I can confirm that the NHF flat race was run over a distance approx. 2m 145yrds.

    Hope the information supplied is helpful, if you require any further info please let me know

    Kind regards

    Apul
     
    davejb likes this.
  14. davejb

    davejb Filly

    Joined:
    Apr 4, 2017
    Messages:
    287
    Likes Received:
    300
    @TheBluesBrother
    I use the HRB data for cards and results each day, and the results file includes the distance corrections - as far as I have checked them they always agree with the information that you can see posted on the BHA website regarding this. Unfortunately for my purposes HRB helpfully changes the distance info to reflect this , which means my program looks for the 'wrong' distance in the standard times file - I have to edit the correction out of the download so my program can find the correct 'nominal' standard time, which my program then calculates the appropriate time correction for. This has, as you can imagine, given me quite a lot of practise with the standard times and corrections for rail movements..... it's almost a minor hobby!

    Tomorrow's cards etc attached:

    Yesterday: Flat handicaps 4/28 at 5.11 10.19 17.11 3.59 (total 36.0) after market filter 2/11 at 17.11 and 3.59 (20.70) after additional 'weight' filter 2/5 (same pair as before, of course, returning 20.7)
    non handicaps 2/8 at 11.5 and 6.4 (17.9) filtered 0/2 (only market filter applies to non handicaps)
    Jumps handicaps 1/6 at 3.13 market filter 1/5 at 3.13 weight+market 1/4 at 3.13
    non handicaps 1/7 at 3.4 filter 1/1 at 3.4

    So, raw ratings out 39 returning 60.43 = +55%, using market filter only 19 out 27.23 back = +43% using both filters 12 out 27.23 back = +127%

    One thing I forgot to mention yesterday was that I don't usually bet odds on - I don't mind something I consider a good thing (like Enable the last couple of times I've backed her) but I avoid odds on normally, and I have a mental stop at 4/6 even if it's Enable, Frankie Dettori up, strapped to a Bugatti Veyron. Consequently, for those who've looked at the selections from yesterday, should be able to see that my possibles for today were Helvetian, Indian Chief, Ayrad, Mezmaar on the flat and - having checked the ! runners I had Ply and Cuirassier DEmpire over jumps - with Ply not being a bet due to the price. Of the 5 left 2 won, at 4/1 and 2/1 (SP) returning +60%

    Dave
     

    Attached Files:

    TheBluesBrother likes this.
  15. davejb

    davejb Filly

    Joined:
    Apr 4, 2017
    Messages:
    287
    Likes Received:
    300
    usual daily post and attachments:

    Yesterday flat handicaps 2/14 at 5.71 and 1.66 same 2 winners from 11 using market filter, same 2 from 4 selections using both filters
    non handicaps 2/15 at 4.64 and 2.9. filtered to 1/4 at 2.9
    jumps handicaps 1/4 at 3.25 market filtered to 1/3, both filters made it 1 from 1
    non handicaps zero from one runner, filtered to zero from zero
    Prefiltering then 34 out 18.48 back, market filtered 18 out 13.88 back, both filters 9 out 13.88 back.

    I'd say that the filters work pretty well, it only really goes to pot when the ratings don't produce a few winners - if you look back through the results you should see this... there are very occasional days when the unfiltered results are good and do better than the filtered version but it's so rare as to not be worth trying - the filtered results are better, on winning days the percentage return tends to be better and there are fewer bets, which makes things much handier all round. On days where the ratings just don't throw up a few winners then results are a bit more random - the filters don't turn a badly losing day into a winning one, they do seem to turn a moderate day into a good one quite often. On a day when the ratings miss endlessly (some days seem to be full of runners coming second or third whilst being allergic to the top spot) then the filters can't do anything to turn a profit - they do, at best, help reduce the loss. As with everything there are very good days, and very bad days.

    Dave
     

    Attached Files:

    Donny and pete like this.
  16. TheBluesBrother

    TheBluesBrother Gelding

    Joined:
    Feb 9, 2013
    Messages:
    1,457
    Likes Received:
    3,140
    @davejb

    Hi Dave,

    The worst racecourse to compile speed figures for, has always been Tramore (IRE), when Mark Nelson used to be the "Time Test" man for the Racing and Football Outlook, he refused to even to look at Tramore.
    When you look at the Racing Post's standard times for Tramore you have similar standards for both the hurdles and chases, my rule of thumb says there should be at least 12.0s difference.
    The major difference between Dave Edwards and myself, If I know that something is wrong with the speed figures I am producing, I will stop and analysis the problem, and not continue to produce crap speed figures
    year after year.

    Tramore.png

    Using my new standard times for Tramore, the figures I had for yesterday's meeting looked spot on to me.

    Tramore2.png

    Mike.
     
    Last edited: Oct 6, 2017
    davejb likes this.
  17. davejb

    davejb Filly

    Joined:
    Apr 4, 2017
    Messages:
    287
    Likes Received:
    300
    @TheBluesBrother
    Hi Mike,
    upload_2017-10-6_12-56-17.png

    We're using the same standard times, as you can see my figures all come out a bit higher based on a different allowance - I'm a bit surprised to have this big a variation from the same starting point given how close we seem to be these days in general. It's probably down to you applying your past experience to this meeting, while I'm a bit hampered probably by the lack of OR's ... only 3/7 winners having an OR to base the calcs on was/is a worry.

    Let's see how I get on substituting 0.00

    upload_2017-10-6_13-6-25.png

    Well, no surprise there - within a lb all the way through. I think in this case I'll stick to your allowance figure, as mine is based on limited info from the ORs, and they're hardly stellar values to use as the 'best' for the meeting.... I'm happy to have different values to yourself when I've good data to work off, but usually in that case we end up with very similar results - I'd need to be happier about the data before I'll ignore your version.
    Thanks for the heads up.
    I wouldn't want Dave Edward's job at any price, and I'm wary of criticising his work when I know so little about how he works or any constraints he operates under - I like reading his 'Weekender' column each week, but I find myself doing so in a very critical manner at times, especially if I think I've detected a bit of dodgy reasoning. This week the last half of his piece covers the 2 races I was most interested in last Sunday - Enables Arc and the Abbaye (due my Marsha interest). He says that Battaash 'earned the highest topspeed rating of the whole season with an explosive wide-margin success...' (etc) giving a rating of 123. He does this in a time that is 'just 0.09s slower than the 5f par'. Enable on the other hand 'just bettered standard time when routing her rivals...' and earns 116 'Even so John Gosden's golden girl hasn't quite clocked a time that does justice to her undoubted brilliance'. I'd agree with him that Enable hasn't really clocked up any high figures for speed, it's not uncommon for top class races to set fairly moderate figures, but you'd expect 5f fireballs to do the business, but it seems a bit off to suggest that Battaash was a total rocket compared to Enable when it was the latter who ran a faster than standard time. Yes Battaash carried more weight, but the appropriate weights for both horses for group 1 races... )
    Dave
     
    TheBluesBrother likes this.
  18. TheBluesBrother

    TheBluesBrother Gelding

    Joined:
    Feb 9, 2013
    Messages:
    1,457
    Likes Received:
    3,140
    I have to agree with him on that one.

    Chantilly.png

    With regards to the going allowances, I will go through each race looking at several horses in each race to see what they have achieved in the past, and build up a mental
    picture, then I might decide that I have been too generous with the going allowance and alter it, as I have said before it is all about opinions.

    Mike.
     
    Last edited: Oct 6, 2017
  19. davejb

    davejb Filly

    Joined:
    Apr 4, 2017
    Messages:
    287
    Likes Received:
    300
    Yes,
    I said on Monday that his win was very impressive, but it was the way he describes 0.09 slow as blisteringly fast whilst being faster than standard is a bit pedestrian.... I think I'd have said the first bit, but I'd have considered Enable's time very decent under the circumstances.
    Looking at my own ratings for UK racing I have Battaash and Harry Angel on 109, Caravaggio and The Tin Man on 108 so far.
    Dave
     
  20. davejb

    davejb Filly

    Joined:
    Apr 4, 2017
    Messages:
    287
    Likes Received:
    300
    Daily upload:
    Yesterday: Flat handicaps 0/11 yikes, lousy day. Obviously the filtered (9 market, 4 both) didn't score
    non handicaps 1/6 at 1.47 filtered this became 1/1 at that price.
    1/10 jumps handicaps at 6.38 0/5 filtered (market) 0/4 (both)
    2/8 non handicaps 2.74 8.24 became 1/3 filtered at 2.74
    raw ratings 4/35 for 35 out and 18.82 back, filtered (market) 18 out 4.21 back both filters 12 out 4.21 back

    Well, if I were advertising this I'd be typing 'recent winners advised at 11/2, 7/4 etc' but it's a good illustration of what I was saying yesterday - unless the raw ratings pick a few winners the filters can't do anything much to improve returns. On the other hand I'd rather lose 12 pts than 35. Let's see how the next few days go, there've been poor patches before often enough, the question is how long this one will last.
    Dave
     

    Attached Files:

    mlmrob and dave58 like this.
  21. davejb

    davejb Filly

    Joined:
    Apr 4, 2017
    Messages:
    287
    Likes Received:
    300
    The usual:

    Yesterday: Flat handicaps 3/11 at 6.68 4.68 5.66 Market filter (this'll be filter1 from now on) 3/8 (same 3 of course) filter2 (market filter+weight filter) 2/5 6.68 and 4.68
    non handicaps 0/7 filter1 0/0
    jumps handicaps 1/9 at 18.55 filter1 0/6 filter2 0/0
    non handicaps 1/2 at 2.37 filter1 1/1 2.37

    raw results therefore 29 out 37.94 back, filter1 15 out 19.39 back, filter2 6 out 13.73 back.

    Dave
     

    Attached Files:

    ArkRoyal and pete like this.
  22. davejb

    davejb Filly

    Joined:
    Apr 4, 2017
    Messages:
    287
    Likes Received:
    300
    As usual then:
    Yesterday flat handicaps 2/17 at 12.5 and 4.9 filter1 1/10 at 12.5 filter2 0/5
    non handicaps 2/20 2.01 and 1.51 filter1 0/6
    jumps handicaps 2/6 5.71 9.2 filter1 1/2 5.71 filter2 0/1
    non handicap 0/6 filter1 0/1

    So raw results out 49 back 35.83 using filter1 it's 19 out and 18.21 back filter2 13 out 0 back
    Well, that was a bad day then. The last 4 or 5 days have swung backwards and forwards with little to show for it, last Tuesday was the last time things really gelled. I'm waiting for things to improve, which might occur anytime of course - but with the back end of the season upon us and the going changing so much I really don't intend to make any predictions.
    The selections.txt carries a blurb about the filters now - essentially trying to identify those runners that I consider to be genuine ratings based selections - ie the ones I would consider actually betting on. This file has a list, sometimes quite a large one, which shows top rated horses that pass the market filter - further ! and * symbols indicate those considered weaker or better candidates. Based on this there would have been 3 selections today (Alphabet, Keep Believing (NR), and Awkright) - ending up sadly as zero from 2 runners, and tomorrow there'd be 4 (Black Bess, Muntadab, Breden and Sweet Pursuit). Although I did bet on some of these selections earlier in the past week, the loss of form of the ratings means I am now sitting on my hands waiting to see some signs of them working as intended.

    Apologies for the later post - I'd been reading the RP online, and having watched a video clip I saw there was another clip on YouTube showing the match race between Seabiscuit and War Admiral, so I watched that, and then youube kindly informed me that the film shot about 15 yrs back was available so I watched that..... I'm a sucker for watching old racing films.... I really must see if there's a clip of Bruni winning the St Leger, maybe I'll spot myself and friends in the crowd. (Looking upset, we didn't back him).

    Dave
     

    Attached Files:

  23. TheBluesBrother

    TheBluesBrother Gelding

    Joined:
    Feb 9, 2013
    Messages:
    1,457
    Likes Received:
    3,140
    @davejb

    Hi Dave, unlike you I have no programming skills to make up daily lists, so I have to rely on my tracker, which at this moment in time only holds 3 horses,
    mind you two ran yesterday winning at odds of 4/6 & 67/10.

    Tracked.png

    As I am the only person I know that compiles speed figures for Auteuil, when Master Dino won at Auteuil on the 1st Sept he earned a speed figure of 117, and he is only a 3yo!
    My par for winning the Triumph Hurdle is 125, and yesterday RPR gave him a rating of 144.

    The best speed figure I had for a 3yo before that in France, was Bonito Du Belais (RPR 150) who earned a speed figure of 123 as a 3yo, so it was no surprise to me when
    Master Dino won a Grade 2 hurdle yesterday at odds of 67/10.

    When Beyond the Clouds won at Worcester on the 3rd Sept in a NHF he earned a speed figure of 100, which to me makes him above average, a lot of Willie Mullins
    bumper horses don't earn speed figures like that.

    Mike.
     
    Last edited: Oct 9, 2017
    pinemarten, mlmrob, davejb and 2 others like this.
  24. davejb

    davejb Filly

    Joined:
    Apr 4, 2017
    Messages:
    287
    Likes Received:
    300
    @TheBluesBrother
    Hi Mike,
    the difference is that you undoubtedly work harder at it, and as a result do better! Having looked at the level of commitment that I think would be essential to do it your way I decided I'd better learn to program it all up to reduce the load - I think it's also important to have a mental back catalogue of races watched, and to have immersed yourself for some time in racing - you need to know the horses, trainers, jockeys and courses so that you can make good use of that background knowledge when forming a judgement.... having only got back into it 9 months or so back I'm only now about to go into the second season of racing, NH, where I'm not starting totally from scratch. Your results are definitely better than mine right now - the last 5 days has been abysmal ;-)

    Notwithstanding the air of 'the triumph of hope over reason' I again upload tomorrow's cards - although if the current run of form continues much longer I shall rename this thread 'Dave's Lay bet Extravaganza'. I think it's important to avoid going quiet when it all goes south, but unless using the selections as lay bets I'd just watch and see what happens - that's what I'm doing right now.

    Yesterday - only 2 handicap flat runners, they both lost. One of them passed the market filter, it still lost. No runners passed the weight or 80% checks.
    Non handicaps were zero from 5, filtered zero from 2, or zero from 1 if you followed the advice re backing runners with a ! next to their name. (This is something I've programmed into my daily checks, as I want to make sure it's a sound idea).
    Jumping: 12 handicap runners, 1 won at odds of 11.47 - so actually almost breaking even! (You would, however, have a heart attack watching all the losers going in). zero from 9 using the market filter, 0/4 using the weight filter (*) and 0/0 using ! also.
    non handicaps 3 from 11 at 5.06 11.0 and 3.86 - none of which passed the market filter which rejected all 11, returning 0 from 0 therefore.

    Okay, maybe not that odd I suppose, the flat is winding down and although not spectacular the NH side of the ratings has been markedly outperforming the flat ratings for the past week. Unfortunately the raw ratings are delivering, rather than the filtered version, so there's a lot of selections involved - over the past 5 days NH handicaps have returned 6 winners from 41, which isn't a great strike rate, but returned a total of 54.46 for the outlay of 41. NH non handicaps have also had 6 winners, this time from 28, returning 33.27, so the NH has been profitable provided you can stand having a low strike rate and lots of selections. The flat on the other hand has been dire, handicaps are at 7 wins from 55, and a return of about 41 (quick total in my head glancing at the notes, it'll be close to that) and 5 winners from 53 in non handicaps returning about 12.50

    Dave
     

    Attached Files:

  25. davejb

    davejb Filly

    Joined:
    Apr 4, 2017
    Messages:
    287
    Likes Received:
    300
    This is doing no good at all to my season's stats.... oh well, tomorrow's ratings attached - no sign of the depths being fully plumbed as yet though...
    Yesterday there was no NH to rate, and on the flat 1 winner from 23 romped in at odds of 3.4, the market filter turned that into 0/4 and adding In the weight and 80% filters took this to 0/2
    non handicaps 15 runners, no winners - 3 were actually selected using the market filter, they still didn't win.
    On the plus side the filters did indicate that most of the rated runners weren't a good bet - the number passing the filters was very small.

    I don't think today was any better, I know that 1 of the two non handicap runners won, but that was at 4/5f, none of the 4 handicap selections (3 flat one jumps) won either.

    The beatings will continue until morale improves.
    Dave
     

    Attached Files:

    mlmrob likes this.
  26. TheBluesBrother

    TheBluesBrother Gelding

    Joined:
    Feb 9, 2013
    Messages:
    1,457
    Likes Received:
    3,140
    @davejb

    Just in case yesterday's meeting at Galway gives you any grief, here's my take on it.

    Galway.png

    Mike.
     
    davejb and mlmrob like this.
  27. davejb

    davejb Filly

    Joined:
    Apr 4, 2017
    Messages:
    287
    Likes Received:
    300
    @TheBluesBrother
    Hi Mike,
    thanks as ever for the help - I haven't processed yesterday yet, I usually get around to that at midday or so, I'll upload my version in a few minutes to compare. As you've got a single allowance across the whole meeting I'd expect to be fairly close...
    I'll be back in a little bit.
    Dave
     
  28. davejb

    davejb Filly

    Joined:
    Apr 4, 2017
    Messages:
    287
    Likes Received:
    300
    Hi again,
    I actually got an allowance there of -1.68 (Heavy) which surprised me, as I updated a couple of standard times to ensure my comparison times started out as per yours... I had no standard time for 2m 3f hurdles for example. Where we seem to diverge is on the OR's, I only had three of the winners holding an OR (as the RP results for the meeting) - are these OR's (for races 1,2, 5 and 7) which you've highlighted your estimates of what the OR should be, or did you use them in your calculations? I know you adjust for what you consider doable, for the three runners with an OR your ratings and mine are quite different - Persistent you make 70, I get 105 for example.
    I'll show you the initial calculation for Galway, then I'll tell my program to use -1.20 to see how that compares....

    Original:
    upload_2017-10-11_12-32-16.png

    Now using -1.20 instead:

    upload_2017-10-11_12-34-32.png

    Ok, much more the same - did you use the highlighted OR's in your calcs, or are they your prediction of the ORs for those 4 based on your ratings? Are they based on class pars for example?

    I know you alter ratings and allowances based on what you consider to be reasonable figures for the runners, your figures are well below the OR's but looking at past performances I can see why - only Persistent and Mahler have much in the way of ratings for this season, I have Mahler running a 33, then a 52, then 63 so it's unlikely he'd have got anywhere near that 118 OR this time out.... your 79 looks more likely than the original 111 using the heavy going figure. Persistent I have doing 34, then 94, then 96 - the 105 I originally calculated is certainly doable from there, for a horse rated at 115, I'm unsure to be honest whether your 70 or the 105 is more likely for yesterday's run. Genie's last run in April earned an 85 off me, doubtful that the original 130 was accurate, even if the OR is 148.... there's really no telling, is there?

    From my own point of view this exemplifies exactly how to introduce dodgy data into my ratings - with only three winners officially rated it rather glues the calculation of the allowance to those three races - there's no option to weed out a race where the pace was unduly slow or fast, and the result is to slew the results of every runner at the meeting. Combine that with a limited amount of form to compare figures with.... I expect this kept you busy!


    Dave
     
    TheBluesBrother and markfinn like this.
  29. TheBluesBrother

    TheBluesBrother Gelding

    Joined:
    Feb 9, 2013
    Messages:
    1,457
    Likes Received:
    3,140
    PM.png
    They are my RPR conversions to OR's.

    All I do is use the nearest standard time, and then treat the new distance as a rail movement, and then adjust the race times to suit.

    This file has every meeting I have worked on, updated daily.

    Speed figures2: Speed Figures.rar

    Mike.
     

    Attached Files:

    Last edited: Oct 11, 2017
    davejb likes this.
  30. davejb

    davejb Filly

    Joined:
    Apr 4, 2017
    Messages:
    287
    Likes Received:
    300
    @TheBluesBrother
    Actually that made me chuckle a bit, because that's what I usually do if I can't find a time - every so often I have to correct my daily download file to change things (as I said a few days back) because the compiler alters the 'official' distance by adding in railmoves - usually adding 0.5f as a sort of standard amount. As that's a bit imprecise I alter it back and use the actual railmove distance to correct the expected standard time. Sometimes, rather less often thankfully, courses seem to make up entirely new race distances - as happened here with a new hurdle distance slap bang between two existing ones. Had I not calculated the standard time you used by looking at your table I'd have added a 1s/15yds correction to the 2m2f to produce it.
    I think that's quite amusing, I copied you instead of doing my usual, and it turns out I was copying you doing what I do anyway....

    Thanks re the RPR->OR info, I would just have scratched my head over it if you hadn't explained it - I figured you were using past performances to determine likely ratings values, rather than basing it on the basic OR based calculation, which you can see in my original table (I imagine, initially, you probably saw similar).

    Well, onto the normal business for the day - upload for anyone finding the form printout useful, as I doubt the ratings are helping much currently... with no past seasons to call on I can't say if this is how it's always going to go at the end of the season, but the run of weather we've had along with the end of season run down (and as NH is only just thinking about starting properly) it doesn't surprise me, there are quite a lot of beaten favourites fading away in the final 2f and long shots romping home of late.

    Yesterday 20 flat handicaps, 2 NH handicaps, 5 NH non handicaps, all without a winner. 3 flat non handicaps were won by the top rated, at 1.88 4.31 and 3.11, from 12 runners. One passed the filter, the 1.88. using all the filters you got that one winner out of a total of 6 selections.

    Dave

    Oops, forgot to say thanks for the rpr conversion table, which I've added into my folder of lookup info.
     

    Attached Files:

    Last edited: Oct 11, 2017
    TheBluesBrother likes this.
  31. davejb

    davejb Filly

    Joined:
    Apr 4, 2017
    Messages:
    287
    Likes Received:
    300
    Cards as ever...
    Yesterday - flat, handicaps zero from 10. Apply the filters and you get zero from 6, 4, and 2 successively as each is added. As the idea is to use all three in handicaps that means the selections text had 2 runners on it in handicaps, both lost.
    non handicaps 2 won from 9, at 1.53 and 2.92 - the market filter turned that into 1 from 2 at 1.53, zero from 1 when you remove the one that didn't pass the 80% rated check.
    Jumps 0 from 5 handicaps (filtered to 0/1)
    non handicap 0 from 2, neither of the pair passed any filters.

    So far today the single non handicap flat runner didn't ... ie it was a non runner. 6 handicap selections were listed, 3 came second (9/1 10/1 and 100/30), one still has to run. Over the jumps, as has happened a few times recently, things went a little better - non handicaps there were no selections, but in the handicaps the two selections both won, granted Sissinghurst was 2/5 and therefore probably not a bet to get excited about, but Mercian King won at 6/1 which was much better. Apart from the odd non runner none of the other listed top rated runners came anywhere (wnb these had ! against them and are therefore not counted as selections... they don't pass the filters) . If this sort of performance continues Ia bit longer 'd be considering backing the NH selections, I wouldn't touch the flat ones with a barge pole however. I wouldn't have backed Sissinghurst by the way, as it's way too short a price.

    Dave

    I've updated the selections text file to add a section at the bottom showing the 'official, these guys have passed the filters' list, to avoid any confusion about which symbol means what. (Hopefully)
     

    Attached Files:

    Last edited: Oct 12, 2017
    pete likes this.
  32. TheBluesBrother

    TheBluesBrother Gelding

    Joined:
    Feb 9, 2013
    Messages:
    1,457
    Likes Received:
    3,140
    @davejb

    I know they had problems with the going at Worcester yesterday due to the quick going, the going allowance I ended up with of 1.02s/f after rail adjustments is the fastest I have ever seen.

    When I calculate the going allowance using the OR's (class), I like to see a variance of -28 for the jumps and -8 on the flat.

    Worcester.png

    So after the going allowance adjustment of 1.02s/f, Rene's Girl speed figure is 97.

    ReneGirl.png

    Mike.
     
    Last edited: Oct 13, 2017
    davejb and markfinn like this.
  33. davejb

    davejb Filly

    Joined:
    Apr 4, 2017
    Messages:
    287
    Likes Received:
    300
    @TheBluesBrother
    Hi Mike,
    Now that would explain some of the differences I've noted along the way - using -8 I got 0.77 for Worcester yesterday. Right, thanks for the info! Here's Worcester recalculated using -28:-
    (hang on, let me recheck some of those times...)

    upload_2017-10-13_21-39-46.png

    Right, I had a few fractions of a second or so different to yourself for most races, so I've double checked against the RP and the BHA results page, now we're agreed on times barring Shoal Bay (my times for these two are 3:32.7 and 3:40.2 using a standard of 3:34.0, railmoves +33 yds each.... as they're divisions of a bumper I suspect your standard time is maybe 1.0s different to mine and the RP's, which would explain it). apart from that bumper I'm generally 4lb lower than yourself, which is a margin I can live with. Thanks for the effort and info, it's good to be able to double check like this. Obviously I'm happy to send you anything I produce if it would help yourself, unfortunately I think you've long since finished doing your ratings etc by the time I start them but if tables like the above are any use to you just say so and I'll email them.

    Right, onto tomorrow, which will probably be another dire day for the flat ratings at least, but should be a good day to watch on TV at least! (Today was quite a nice opener, I even managed to pick the 12/1 winner of the last race of the ITV7 - sadly I only got one other, so no £50,000 for me this time).

    Yesterday - flat: 0/9 handicaps, 0/3 non handicaps. There were less selections when filtered, but they still lost.
    Jumps all 7 non handicaps lost from the raw ratings, only one of the 7 was a selection as such as only one of them passed the market filter. 3 out of 10 won in handicaps, at 8.88 8.2 and 1.49, which was okay, 2 of the 8 that passed the market filter won (the one at 8.88 and the 1.49) adding in the weight filter turned this into 2/4 (even better) and with all three filters applied it became 2/2, even bettererer! I'm not going to claim any great triumph here, but as I said before the filters only work if the basic ratings produce a few winners. As far as returns go the raw ratings gave 29 out and 18.57 back, all from the jumps of course (which was 17 out 18.57 back), with the filters applied as they're meant to be that turns into 7 out and 10.37 back, over the jumps alone that would be 2 out and 10.37 back.... it's easy to massage figures to look good, isn't it? In truth the ratings are not doing well at all now on the flat, they aren't doing too badly over the jumps, but that might go through a bit of a phase now as the jumps proper are about to kick off - there'll be lots of horses with no ratings coming up, as of course an awful lot of them haven't been running over the summer. Whilst I'd love to see a bit of a burst of winners, I suspect things are now going to be a bit slim for the next month or so.

    If you look at the selections text file appended you'll see the result of this season start, very few actual selections and lots of ! symbols indicating too many unrated runners in a race.

    Dave
     

    Attached Files:

  34. TheBluesBrother

    TheBluesBrother Gelding

    Joined:
    Feb 9, 2013
    Messages:
    1,457
    Likes Received:
    3,140
    You were right, I was out by 1.0s which I just corrected.

    The RP don't have this problem they just ignore the rail movements, on Thurs at Worcester the going allowance was +0.45s/f (good to firm), some of the speed figures were quick
    notably the 115 given to Rene's Girl.

    Rail movements added about 48yds to the race distance, the time was 8.1sec faster than the standard.

    48yds = 3.2s, so after the rail movement adjustment the winning time ends up 11.3s faster than standard.
    This causes a problem when assessing the going allowance, was it the fast going that help Rene's Girl achieve that fast time, or is it simply the case the horse has improved.

    Mike.

     
  35. davejb

    davejb Filly

    Joined:
    Apr 4, 2017
    Messages:
    287
    Likes Received:
    300
    Well, like yourself I think it's important to have as accurate a race distance as possible, and given how easy it is to adjust for rail movement (which has been up to half a furlong or so lately) I cannot see why anyone would ignore this factor. Obviously if, like myself, you have programmed this sort of thing up then you need to e able to adjust the code to reflect changes in your method, but if you can program the speed calcs then the railmove adjustment is a minor extra. I like the challenge of getting the program right, but I find staring at the spreadsheets a bit tedious, and if you find it tedious it's too easy to make mistakes, hence I try to turn as much as I can into code - computers don't get bored! (Neither do they make mistakes, it's the people feeding data in that do that). I do lose out on acquiring a proper feel for things though, if a runner jumps in rating from mid 60's to 110 you'll notice it a lot more readily than I do.

    Rene's Girl - probably both... all part of the puzzle of course, there are lots of factors affecting each and every race and runner, which is really part of both the frustration and the fun of the game.
    Enjoy the racing.
    Dave
     
    TheBluesBrother likes this.
  36. davejb

    davejb Filly

    Joined:
    Apr 4, 2017
    Messages:
    287
    Likes Received:
    300
    Well here we are, page 97 already and no winners since the last walkover - and that's only down to getting it each way ante post.....

    Okay, it's not quite that bad, but it has been a bit dull of late - today a few winners popped up, and as I watched the ITV racing it was apparent that the winners quite often featured in the top 2 or 3 of the ratings, away from the cameras there've been the odd winner, and also some that would have been good EW bets... not great, but better than total failure at least.

    Tomorrow's cards attached of course, yesterday's post mortem follows:

    The flat still very poor yesterday, 14 handicaps without a winner, as the various filters come into play this scores as 0/14 0/7 0/5 and finally 0/2 using all filters, as is supposed to be the case.
    non handicaps had 2/16 at 4.68 and 6.4 1/4 using the first filter (4.68) and final filter turne4d that into 0/2 - rats
    jumps 0/7 handicaps, 0/3 0/2 0/2 using the filters (ie 0/7 raw, the market filter turned this into 0/3, weight filter included turned it into 0/2, and the 80% rated runners filter made no changes so also gave 0/2)
    non handicaps 3/7 at 2.3 2.35 1.72 then 2/3 (filter1) 2.3 and 1.72, no runners passed both non handicap filters so 0/0 there.

    I know the filters take out the occasional winner, in fact they are cutting the number of selections right down - today only 3 ran (2 NRs) - the filters are pretty much suggesting to hang fire on betting, and as they aren't producing very many winners I think I'll take that hint.

    It'll be interesting to hear what's said about Expert Eye, did he just blow his chance a la Battaash at the start, or was something else amiss? At least the O'Brien machine did the business, and the winner put in a game run, I just find it hard to see that as having been as good a race as we were perhaps expecting to see. I had to admire the gamble on the Cesarewitch - I can't imagine paying 170,000 (I think that was the figure) for a horse at 3yo, gelding him, and planning to nab the Cesarewich as a 4yo.... the odds of even having him at the start ready to go on the day would have me quaking, never mind the small matter of beating 30 odd other runners. I can only say that the owner and trainer must have steel versions of what the horse no longer has.

    Dave
     

    Attached Files:

  37. ArkRoyal

    ArkRoyal Administrator

    Joined:
    Dec 28, 2012
    Messages:
    12,917
    Likes Received:
    13,852
    @davejb I believe that your filtered selections are rating qualifiers filtered by SP? Do you have any idea how those results would compare against just using the SP and ignoring your ratings :eek: The reason for asking is that this would, I believe :confused:, give a benchmark as to how well the ratings are performing rather than the effectiveness of the filter.
     
    Last edited: Oct 15, 2017 at 10:12 AM
  38. davejb

    davejb Filly

    Joined:
    Apr 4, 2017
    Messages:
    287
    Likes Received:
    300
    @ArkRoyal
    Hi Ark,

    I can check most things I can think of back to 3rd week of August fairly quickly - prior to that the filters weren't really in use (it's when I added filters that I redid the analysis routines to cope, and restarted the records for that analysis).
    I'm not entirely sure what you are suggesting, if I just use the SP I don't get a selection for the race?

    I can, for example, run a check to see how my top rated handicap runners won 105/793 races, with a return of 92% of level stakes compared to say forecast favourites which would have given 136 winners in 706 races, returning 87%. I have the actual odds returned (BFSP) but not the market position, as I'm picking the night before I keep the forecast odds and calculate the market position from that, which is then stored.

    Currently the non handicaps are filtered by:
    (a) Top rated, (b) forecast favourite, (c) 80%+ pf runners in race have a valid rating.
    Handicaps are:
    (a) Top rated, (b) in top half of the forecast (ie every horse in race has its odds compared, so rank all the forecast prices, find the one in the middle, check the runner is at that price or shorter), (c) in upper half of the race card (so upper 50% of runners ordered by weight), (d) 80%+ of runners in race have ratings

    If you can give me an example of the sort of query you mean I might be able to give a more accurate/complete answer. Currently I pretty much use level stake return as a benchmark, figuring a profit as the essential test - and I know that the raw ratings alone will not manage that. (Hence the filters, intended to reduce number of bets and improve profits). At the moment the ratings aren't performing as well as they did over the summer, but their win rate has always been quite low, compensated for by better prices.

    Dave
    ps - I really would appreciate an explanation of what you are suggesting, I'll take all the help I can get!
     
  39. davejb

    davejb Filly

    Joined:
    Apr 4, 2017
    Messages:
    287
    Likes Received:
    300
    Tomorrow:
    Yesterday - compared to the rest of the week an improvement, not great but still better....
    handicaps on flat 2/16 at 4.84 and 6.06, both passed the first filter giving 2/8, 2/5 by the 80% test, the weight filter zapped this to 0/1....
    non handicaps 2/15 at 7.40 and 6.45, narket filtered both out to 0/2 (0/2 still with 80% check)
    jumps handicaps 1/9 at 12.64 1/5 passing market filter (same one of course) - nothing passed last two filters
    non handicaps 1/6 at 2.45 1/1 on market, 0/0 on 80% check.

    Well, back when I started trying to identify filters the first to come up was the market one - and it's worked okay this past week, if you look back over the posts you'll see that. Unfortunately this usually leaves quite a lot of selections, usually about 15-20 a day, and at that point the 80% check cuts the numbers right down and tends to remove most of the bets and all the winners. (The logic here would therefore be to ONLY bet on those that had a ! by their names.....) All I can say is that the market filter is worth applying, the results of applying that filter to the ratings is the full list given on the selections sheet each day. Currently the other filters (80% ! check) and the weight check in handicaps don't seem to be helping.

    Dave
     

    Attached Files:

  40. ArkRoyal

    ArkRoyal Administrator

    Joined:
    Dec 28, 2012
    Messages:
    12,917
    Likes Received:
    13,852
    Hi Dave,

    I wasn't sure how you were using the F/C prices for your filters.

    So what I am saying is, what are the returns for non-handicaps just using (b) and (c) and how does that compare when using (a), (b) and (c)

    Handicaps are a bit trickier in that you will have more than one qualifier in a number of races but it would be the same principle ie compare returns for all qualifiers from (b), (c) and (d) against (a), (b), (c) and (d).
     
    davejb likes this.

Share This Page