Early days

[davejb]

Yes,
I can see how even with practise that would take you some time to work through each day!
My latest method, which is very early in development currently, has been looking at calculating a more accurate ga for the race - perhaps a bit simplistic of me but I figure if I get the allowance right I'll be pretty much there. I've been basing it on the idea that a 100 rated runner will manage standard time on good going carrying 9-0, and trying to use the OR, weight, distance of a runner to calculate a correction factor to reflect my 100/9-0 runner. So far I'm working this out for the first and second in each race, to allow for races where the second performs to a higher rating than the winner, and looking at the best race result each meeting to decide what to hang my hat on - figuring that most races will not reflect horses running to their OR but the best figures for the meeting are probably pretty close.

I'm still working on it, I hope to have something to show you in the not too distant future....
It's very kind of you to go to this trouble to explain things, I do appreciate your efforts.
Dave

 

[TheBluesBrother]

I figure if I get the allowance right I'll be pretty much there.

The going allowance would be the most important factor when compiling speed figures, you could have the the best set of standard times available,
without the knowledge on how the going on the day, was slowing or speeding horses up, you would be up the creek without a paddle.

Some days you will see speed figures that just don't make sense, this morning I asked Dave Edwards this by email.

Morning Dave,

With regards to the official race distances in Ireland, here is something to ponder over, was the 1m4f distance at the Curragh correct?

How on earth did the rider of the Amateur race Miles to Memphis carrying 16lb more than Enable, get within 2.73s of a impressive classic winner.

The speed figure of 49 I have for Enable is so bad, I am having a hard time trying to work out what went wrong here.

I had the going allowance as firm…

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

Morning Mike

Strange goings on maybe they went slow early in the Classic

Definitely firm ground though.

Dave.

 

[davejb]

I'm still beavering away, and was just about to take a break when I noticed that your figures for comptime at Hamilton are fairly different to mine - somewhat unusually HRB, RP and Timeform all give the same times for Hamilton yesterday, and I'm using the same standard times as yourself (unless you've altered them in the past week) - from your race distances (most noticeably the last race, which was 8.5f and you have it as 9.3) it looks like you've factored railmoves in or something? According to the BHA site race distances were 'as advertised' with no rail movement to account for. I get comptimes of 3,3, 3.04, 3,41, 6.12, 5.81, 4.0, and 5.26 for Hamilton. Races 1,5,6,7 are all a fair bit different, and as races 1 and 2 were over the same distance I don't see how we could agree on race 2 yet not race 1.... any ideas?

Dave

 

[davejb]

Three for tomorrow,
Ayr, Windsor, and the AW at Wolverhampton
Dave

 

[TheBluesBrother]

I noticed that your figures for comptime at Hamilton are fairly different to mine

The comparison per furlongs times I use are taken off the Racing Post website, remember I update the standard times list most days.



I see Dave Edwards has used 2 going allowances for Hamilton?

Standard times: https://drive.google.com/open?id=0B3rVp0WNvwn8anUySlNvSjN1bXM

 

[davejb]

Ah,
okay - I'm still using slightly different times to the RP for 8.3f and 11.1f.
Yes, and that is a tweak (change of going) that I'll have problems with - there's a sneaky but more labour intensive option for dealing with that and I'm going to give it a go.The problem I have currently is that I have more ideas than I have time to actually code up!
Dave

 

[davejb]

Here's 3 cards for tomorrow,
anyone following the exchange between myself and Mike ( TheBluesBrother) will hopefully understand that I post these for interest, I believe some of the filtering my program produces would be of use to anyone looking to analyse form, and that this is still very much a work in progress. I offer these cards each day as an aid to those who are interested, I hope one day they might be regarded as bloody useful! (If they do become so I hope everyone will appreciate the large part Mike has played here in getting me to think the right way).

Dave

 

[dicko14]

davejb
Thanks for putting these up for our perusal.
How you guys do this is amazing and way over my head.
Saw an interesting segment whilst watching ATR today where they showed 2 divisions of the same race where in the first race they went hell for leather from the off and in the other they ran to even splits which was the faster race. Think this maybe your next project?
Dave

 

[Kevinc]

Interesting thread ive always liked speed ratings but never had the commitment. With the Hamilton standard times , does the rp having a separate standard time for when they race in the loop

 

[TheBluesBrother]

does the rp having a separate standard time for when they race in the loop

There are only one set of standard times for Hamilton.




The racecourse has just been remeasured.

Mike.

 

[TheBluesBrother]

davejb

One thing that I have failed to mentioned is the max and minimum weights I use for both the Flat and jumps.

Flat - max 10-0. min 8-0
For example, if a horse on the flat is carrying 11-6 in an amateur race, I only adjust to 10-0.

NH - max 12-0 Min 10-0



Example from yesterday's meeting from Ayr, as you can see Royal Shaheen rating was only adjusted to 10-0.

The importance of looking at the rails movements certainly applied to Ayr yesterday.

AYR/MON 17 JUL

GOING/TRACK

Good (GoingStick: 7.8 on Monday at 06:30)
Rails: Inner rail out 6yds/Top Bend out 8 yds

Race:3 Distance as advertised
Races: 1,2,4,5 & 6 +24yds (subtract 1.6s)
Race: 7 +42yds (subtract 2.8s)

PLEASE NOTE: the current set of RP standard times for Ffos Las cannot be used on yesterday
course configuration, there is a major discrepancy, I sent an email to Dave Edwards.

Mike.

 

[davejb]

Rail movement,
yes, it makes a sizeable difference quite frequently, and it's not that long ago I recalculated everything to include them as you probably recall. Your idea abut max weight bears thinking about - I'll add it to the list! I haven't done yesterday's figures yet - it doesn't take very long and I prefer to keep working at the update, as I'm not betting until I get this all sorted out it's a higher priority at the moment to see if I can improve things.
Thanks as ever for the information.
Dave

 

[davejb]

Okay,
TheBluesBrother
I said earlier that I'd post something for you to look at if I managed to get anything to work on my second method of calculating the allowance, and here are a couple of days I've been using for testing purposes with allowances calculated by my updated program.
I was using these two as I have your speedfigures up to around then - I don't want to just copy your stuff, I'd just like to be somewhere on the same sheet with my efforts!
From a fairly quick glance I've got some allowances that match yours spot on, I am usually within 0.1 or less of your figure.... I'm not sure I'm going to get any closer as I suspect the remaining differences are the result of your own expertise being used to tweak the figures until they match your expectations.
I'm going to trial this version's speed figures for a while now to see if they perform better than my original set - so the next hour or so will be spent recompiling a second database of figures for the past 18 months.
Thanks again for all the help, I'm going to trial things for a while now before making any significant changes to anything - I need some thinking time to decide if I can improve things further.

dicko14 sorry, I didn't spot your post initially. I'm currently on summer holiday - which will lead seamlessly into my retirement, so this is a project to keep me and my noggin busy doing something I have been interested in for about 45 years. If I ever get to the point of wanting to do something else I'll probably have a crack at another interest from my misspent youth and try handicapping greyhounds. (I used to be better at that).

I'm not sure that split times for racing over here is all that practical - it's hard enough getting reliable timing data for the races as you probably realised from some of the chat on this thread. I know Timeform produce the figures, in fact decades ago I bought the first few of their annual timefigure booklets.. mainly because I couldn't afford the racehorses annual!
Dave

 

[TheBluesBrother]

I am usually within 0.1 or less of your figure

Make the figures your own, my method of using the 10lb constant, to calculate the going allowance is fairly recent, I had that figure in my head for a long time, but it works.

After a while you come across figures that are just not right, like Rowlestonerendezvu at Ffos las yesterday, the 1m4f distance is now 1m3f209y, I took 2.0s of the standard time to correct it
or the 48 rated winner would have had an exaggerated high speed figure.



Whenever I change a standard time, the lbs per length figure is also updated.

Mike.

 

[TheBluesBrother]

davejb

Here are my figures for the mixed meeting at Killarney yesterday, note the 2 different going allowances for both the flat and jumps,
in the majority of cases, you have to take the official Irish race distances with a large pinch of salt.



Killarney standard times updated.

Mike.

 

[davejb]

Make the figures your own, my method of using the 10lb constant, to calculate the going allowance is fairly recent, I had that figure in my head for a long time, but it works.

Well, I don't really have any option not to - as you said a little while back I would find it extremely difficult to program in the different tweaks you do, so the ratings will evolve as I alter things, I did want to be somewhere in the same ballpark as yourself just to ensure than I wasn't producing imaginary Unicorn racing ratings

Handling multiple allowances for a single meeting would be difficult I think to program in, but it's not off the radar - it just isn't likely to be a mod I can put in any time soon.
Cheers again, gotta go outside.....eek!
Dave

 

[dicko14]

davejb
Congrats on getting out of the rat race. I see your retiring at the same age I did which was 2 years ago now, how time flies!
Enjoy the holiday and your retirement.
Dave

 

[davejb]

Daily cards for tomorrow then, these cards are all using the new version of the ratings, having recompiled them all for the past 18 months or so.
I've left the mixed turf/AW card at Lingfield out of it - it's no harder to do than any other but I would rather watch the others on turf.

TheBluesBrother Tomorrow's results from Lingfield, being mixed, will be a bit of a test.... my program SHOULD treat it as two separate meetings (obviously the database records whether the race is on turf or AW, and there's coding inside that ensures the correct set of standard times is applied) but I wouldn't be surprised to see something needing a quick fix when I actually come to process it. This might be a possible route for handling cases of the going changing during a meeting, process each part of a set of results as a separate meeting.

dicko14 Well, the last few years at work were pretty hard going, so I'm not sorry to be out of it at last - the trick is to accept there is more to life than work, and I have a number of ideas for how to make retirement go the way I'd like, so we'll see what comes up.
Dave

 

[mlmrob]

What was Love Oasis real figure for last Saturday as I think 141 must be an error.

 

[davejb]

You're quite right, I'll have to step through Chester for that day and see if there's some reason it's going allowance was calculated to be rather larger than expected - a bit of a 'back of the fag packet' manual rating looks to put Love Oasis at 58 or 59 for that run. Sorry, this stuff is still developing (as the thread illustrates on a daily basis) but after I find the reason for the miscalculation (which usually just takes time and a degree of bloody mindedness) I'll hopefully be able to check for it automatically in future.
Dave