One of my fave threads on CL and really enjoyed it, didn't win but the knowledge ITT is unreal. Thanks all
I had Rathvinden who looked good but Tiger Roll is just too good. Combi tricast Tiger Roll, Rock the Kasbah & Anibale Fly. not the worst effort. Quality race RIP to Up for review
So, can any of you regular punters help me with this one.
As I said on Saturday, I had Tiger Roll on an e/w bet with William Hill. Put £2.50 down. Picked up my winnings this afternoon & got back £8.50. Certain the lady paying me said the odds were 4/1. Obviously all done electronically now so no reason to think she can't count or I've been diddled / had one over them. Just can't fathom how I got back £8.50 on a £2.50 stake if the odds were 4/1.
Just to clarify......I also had £2.50 each way (so a £5 total bet) on Livelovelaugh which I don't think counted on my betting slip as it wasn't top 5.
So, can any of you regular punters help me with this one.
As I said on Saturday, I had Tiger Roll on an e/w bet with William Hill. Put £2.50 down. Picked up my winnings this afternoon & got back £8.50. Certain the lady paying me said the odds were 4/1. Obviously all done electronically now so no reason to think she can't count or I've been diddled / had one over them. Just can't fathom how I got back £8.50 on a £2.50 stake if the odds were 4/1.
Just to clarify......I also had £2.50 each way (so a £5 total bet) on Livelovelaugh which I don't think counted on my betting slip as it wasn't top 5.
Any thoughts ??
Golfie- you say you put £2.50 down but was that £1.25 each way or £2.50 each way ?
Soapbox Sam before the National Crisp was classed as a 2m-2m4f horse........
KK, my point is, if Crisp was consided a 2m-2m.4f horse, why did he carry 12st and go off joint Favourite. He must've run a longer race somewhere before that National ?
So, can any of you regular punters help me with this one.
As I said on Saturday, I had Tiger Roll on an e/w bet with William Hill. Put £2.50 down. Picked up my winnings this afternoon & got back £8.50. Certain the lady paying me said the odds were 4/1. Obviously all done electronically now so no reason to think she can't count or I've been diddled / had one over them. Just can't fathom how I got back £8.50 on a £2.50 stake if the odds were 4/1.
Just to clarify......I also had £2.50 each way (so a £5 total bet) on Livelovelaugh which I don't think counted on my betting slip as it wasn't top 5.
Any thoughts ??
Golfie- you say you put £2.50 down but was that £1.25 each way or £2.50 each way ?
It had £1.25 on the betting slip as far as I can recall
So, can any of you regular punters help me with this one.
As I said on Saturday, I had Tiger Roll on an e/w bet with William Hill. Put £2.50 down. Picked up my winnings this afternoon & got back £8.50. Certain the lady paying me said the odds were 4/1. Obviously all done electronically now so no reason to think she can't count or I've been diddled / had one over them. Just can't fathom how I got back £8.50 on a £2.50 stake if the odds were 4/1.
Just to clarify......I also had £2.50 each way (so a £5 total bet) on Livelovelaugh which I don't think counted on my betting slip as it wasn't top 5.
Any thoughts ??
Golfie- you say you put £2.50 down but was that £1.25 each way or £2.50 each way ?
It had £1.25 on the betting slip as far as I can recall
See stakitsteve's response for the answer. Had you placed it with say Bet365 you would have got back 25p more as they went 1/4 the place.
Until you begin to get a full and detailed understanding of betting I wouldn't include this in the portfolio of investment options you offer any of your clients!
Before signing off on this year’s GN thread, in case you’re
interested, my views on how my new (Mark II) GN Model performed.
As always, it’s been a fun thread to participate in, regardless
of Saturday’s result. Obviously, the result represented a loss for me and sincerest
apologies for leading you astray.
It was, in my view, a case of the workman and not the tool that was to
blame. In fact, I reckon the new Model did a pretty decent job. The evidence?
The Model Mark II
On the going (GS or better), it identified 5 with Winning Potential. All had
ratings within the range of the top 90% of its back-tested ratings of 22 GN winners/near-missers
from 2005. 2 of them had ratings consistent with the top 25% of former
winners/near-missers.
So, those 5 were rated by the model (and came home on
Saturday) as follows:
Joint #1: Tiger Roll (1st) and Ramses de Teillee (PU)
Getting 1st and 5th from 5 candidates
is a decent result, particularly as 2 of the other 3 Fell (VC out of the
running at the first). But there’s a further observation worth making:
The Model also identified 4 with Strong Place Potential, “next best” in stat-rating, consistent
with profiles of bottom 10% of former GN winners/near-missers and a number of others
making the frame (rating = c.15L behind TR/RDT):
Joe Farrell (PU), Rathvinden (3rd – 5L),
Ballyoptic (F), Walk In The Mill (4th
– 16L)
So, its top 9 selections produced: 1st, 3rd,
4th and 5th (i.e. from 22.5% of the field, it found 67%
of the 1~6 paying places and 80% of the first 5 home) – 3 of the other 5 Fell.
As it happens, that’s 1 more of the first 5 (or 6) places than
were filled by the Virtual Grand National’s top 9 selections (represented by its
predictions of the first 9 home).
So why the loss?
The model’s designed to make my GN bets as objective as
possible - ordinarily
I aim to back all the Winning candidates (scaled stakes) each-way. But, of course, the collective odds of that group dictate the risk-reward profile of that approach and moving parts (final weights, final preps, odds, bookies’
terms and going) make the process of translating the model's output to a betting slip more art than science - some years more than others. And that’s where things
have gone awry for me 2 years running.
I won’t bore you with the whys and wherefores - the simple
fact is that, with 3 of the 5 Winning candidates at/near the head of the market, my normal approach simply wasn't economically viable. In glorious retrospect, I should have had a stake-saving win bet on the Tiger (as did the far more astute @PaddyP17) but instead I
had the bright idea of giving him (once again!!) the swerve, in favour of
a 50/1 selection from the next best 4.
The outcome? Of the top 9 identified by the model:
· 5 Backed = 1 place, 4 fail
· 4 Not Backed = 1 win, 2 places, 1 fail
CONCLUSION: Tool = OK ........ Workman = tool.
To be continued ........... "To tweak or not to tweak?"
Ballydine and Chic Name were on my radar. Waiting for the day and extra places though. Looks like ground could be very quick too so sitting on sidelines currently!
Fantastic stuff as ever, Peanuts, and I've been radio silent on this thread last few days, but only so I can formulate a few questions on the model, because it certainly seems to do a good job.
Human fallibility is human fallibility, so your interpretation of the model is what it is, and regardless of whether I win or lose, I find placing my shillings as a result of your thoughts (and my own responses to them) very enjoyable.
That said, a couple of questions do abound:
- Have bookmakers' odds ever come into the model's factoring?
I realise SPs are probably the only thing you can go off, unless (for instance) you're able to recall heavily punted fancies, like Baie Des Iles coming in from 66s to 14s last year.
But, I would possibly guess that - certainly with this year and last - your choice of scaled each-way bets means you're poking at generally higher odds, and thus those you leave unbacked at shorter odds "show you up" (absolutely not a criticism btw) a little?
Also - the Mighty Tiger winning at 12s and then 4s(!) also helping establish the "strength of the bookies might be worth considering.
I realise you said in the immediate aftermath that to be successful you need an identifier of "value", so the above question might be moot, but even so.
- How do you weight the significance of each individual stat/profiling measurable?
Not to use the same horses over and over, but my clearest memories here are of this year and last. So, with Baie Des Iles and Ramses De Teillee both running below expectations as seven year olds, have you thought of adjusting the model accordingly?
That said, I realise that very few 7 y/os have fit a winning profile anyway, but those that have (or that were SPP/MPP) might not run to form because of their age?
- More a follow-up to my previous question, but I presume certain things are more important? For instance, facing the GN fences is a more important variable than, say, age?
-------------------
There are probably a few more questions that'll come to mind, but either way, I'd be intrigued to know some of your thoughts. It's an immense undertaking, all for one race (albeit one of the greatest races), and I'd love to know more!
Cheers @PaddyP17 All good questions. If you don't mind, I'll get back to you after I've posted a reply to Chunes and described the tweaks that are appropriate for the model. Sets a bit of context to answering the questions you put. Please bear with .....
@PeanutsMolloy Thanks for this thread and your picks. I broke even and had a good time doing it!
Some years ago I asked you a boring question (that you kindly said was not boring). It was basically whether your model could ever be over-modified... Say if you adjust it one year based on the results, but the results were because of anomalies, and so your model becomes less accurate. You said that it was not likely since you tried not to include things that were clear anomalies.
As a follow-up, I was just wondering whether you ever run the stats post-race based on older versions of your model, to see how they would have fared? And when you adjust your model do you do it on an exclusionary basis (excluding horses that won't win) or inclusionary, or both?
Forgive the boring questions again. Just find it interesting.
I'll post shortly about whether/why/how I’ll tweak
the new model based on Saturday's result ….. if only to serve as another sleep-aid
To answer your 2 questions, both not boring
(well not to me and you) and highly pertinent:
1.
I’ve not normally re-tested races with a former
version of the model and, for a couple of reasons, I don’t particularly want to
this time.
·
The old model was for another era, referring to a
database for which 23 of the 28 GNs were over appreciably stiffer fences (and a
longer trip, as it happens), many of which also had significant numbers running
from out-of-the-handicap. I don’t think referring back to that system is going
to be all that helpful for the future, to be honest.
Not that the old model was broken by last
year's result (it had 1st, 3rd and 5th, though I naturally only had 5th on my
slip ) but, as Pleasant Company was first “outlier” to go close to winning
since I created it in 2006-07, it felt time for a revamp based on a more
relevant database.
·
Leaving my betting faux-pas aside, the new model really
did a pretty decent job.
Expanding on my earlier
post today, there was only 1 “outlier” in the first 8 home - even One For Arthur
(25L 6th), Regal Encore (28L 7th) and Singlefarmpayment (28L
8th) all ran broadly to expectation (all in a group of 10 runners
rated c. 20L behind TR/RDT).
Admittedly, it was a
significant outlier - the 2L runner-up - but I know (from sad, indelible
familiarity) that no version of the old model would have predicted Magic Of
Light’s run (she was expected c. 30L behind TR).
2.
If and when I tweak the new model this time, I’ll
do what I always do and back-test it against the whole field (inclusionary) and
all previous GNs (winners and finishing <30L) in the database to ensure it
"fits" with them at least as well as the untweaked version had. I get
particularly excited (sad isn't it?) if I find that a tweak to reconcile a
runner this time also helps reconcile a small anomaly that I’d let ride
previously – it does happen and, naturally, gives me confidence that it’s a “good
tweak”.
But, to emphasise, a tweak to reconcile an outperformance
is only likely to be a minor adjustment to a threshold, bestowing/removing a +/-.
As we talked about previously, I only ever look to reconcile out-performances
that marginally, but crucially, missed expectations (e.g. I’ll be looking
closely at Rathvinden’s stats to see if there’s a borderline miss of a + for
which it might be appropriate to make a tweak)
Of course, there are underperformers too but,
unless there’s a clear logic as to why that may have happened (possibly there is
a common one this time for RDT and General Principle), we have to accept that racehorses
may have off days or a problem we may never hear about (e.g. a bleed, lose a
shoe).
Just to expand on the evolution of the new model. As
said, it uses a more recent database but there’s a good deal in common between the
reference factors used by the old and new models.
It was in reconciling Pleasant Company (he’s from
the family of Comply Or Die) that I decided to incorporate pedigree (for the
first time) to create Mark II. This is the main departure and I think is a very
interesting area (a few specific bloodlines are represented surprisingly often in
winners/near-missers of GNs, pre- and post-2012); one I’m intending to research
more.
Otherwise, the new
model’s built on the same principles and uses largely the same factors as the
old one. The formula’s slightly different, reflecting
slightly different “tests” within the factors and slightly different +/-
weightings.
Rattling off the factors (in no particular order):
Age
Experience
of the GN course
Absolute
weight to be carried, compared to chase record
GN handicap
mark, compared to 3m+ career-best RPRs
Chase
experience
Frame-making
strike rate
3m+
chase Win or near-miss strike rate
Chase
form at 3.25m+ and 3.5m+
Jumping
record, particularly at stiff-jumping tracks
And how many horses go through that process and do you only start with first declarations?
Effectively I run the slide rule over the top 80 when the weights are published but that's not as bad as it sounds. I've got an initial screen, based on a couple of key stats that, if they fail both, I know they're not going to figure so I stop there. Not strictly an elimination but effectively so. That quickly weeded out about 30 this year, so I went through the whole process for about 50. 14 runners on Saturday had initially been screened out - nearest finisher was Outlander (9th - 31L). Worked OK. Cheers.
My takeaway from last few years, and to an extent this year too, is that this race is now largely just a classy staying handicap.
Last few winners been 8 or 9 year olds, 4 of the first 5 were 8 or 9 this year. Think it's going towards the traditional look for the upwardly mobile younger horse as opposed to the more experienced stayer that was typically well suited to the test of the race previously. Certainly don't think it'll be long before a 7yo wins it (hopefully next year, *cough* Mister Malarky *cough*)
Still maintain Tiger Roll at 4s is one of the worst bets you'll have, regardless of the result.... did nearly come down a few times and some of his main rivals fell by the wayside. Will certainly be taking him on again next year....no room for sentimentality in this game!
My takeaway from last few years, and to an extent this year too, is that this race is now largely just a classy staying handicap.
Last few winners been 8 or 9 year olds, 4 of the first 5 were 8 or 9 this year. Think it's going towards the traditional look for the upwardly mobile younger horse as opposed to the more experienced stayer that was typically well suited to the test of the race previously. Certainly don't think it'll be long before a 7yo wins it (hopefully next year, *cough* Mister Malarky *cough*)
Still maintain Tiger Roll at 4s is one of the worst bets you'll have, regardless of the result.... did nearly come down a few times and some of his main rivals fell by the wayside. Will certainly be taking him on again next year....no room for sentimentality in this game!
Re-viewing the race, it's remarkable how inconsequential even gross mistakes at fences have become. To see Magic Of Light absolutely plough through the Chair like it wasn't
there (like Moses parting the Sea) and go on to be 2L runner-up (after ploughing through the last as well) suggests that the fences appear to have become since 2013 (literally) softer still by their construction.
Amazing though he is, I seriously doubt that Tiger Roll could
have won a GN over pre-2013 fences. While the handicapper’s 172
rating of Saturday's win is relevant (in making it certain he’ll carry
11.10 if he returns), for the Racing Post to dub him therefore “the best GN winner in at
least 46 years” is meaningless crap IMO.
Peanuts - a friend posits that you're approaching the fence issue backwards, and that it was a real achievement for Magic of Light to finish second after ploughing those two fences.
I'm on your side, and think the fences have got more give in them, but is there any merit in his point?
Peanuts - a friend posits that you're approaching the fence issue backwards, and that it was a real achievement for Magic of Light to finish second after ploughing those two fences.
I'm on your side, and think the fences have got more give in them, but is there any merit in his point?
I completely agree but I'm not looking at it backwards (I don't think). She did fantastically well but the point still stands about the changing nature of the race that, in former times (I would contend even since 2013), that mistake at the Chair would (99% of the time) have ended any horse's race, one way or another.
It's better seen head on but this is the best I can do with a still from the normal view. She's actually in the middle of the fence (yellow sleeves & noseband of course), having gone into it no more than half way up. There appears to be no core to it whatsoever. Bizarre. We're talking the most feared obstacle on the course bar Bechers. All credit to her for a fantastic run and she clearly has bucket loads of guts and stamina - but is this really "The Grand National"?
Thanks again peanuts for making me interested in something I would otherwise let pass me by
Sadly I think that's it for me and hunt racing, when I saw the grey go down and thrash about I stopped caring about the result and felt hugely sad for the horse and whilst horses being euthanized and hurting themselves isnt a new concept to me I felt bad about this one
Thanks again peanuts for making me interested in something I would otherwise let pass me by
Sadly I think that's it for me and hunt racing, when I saw the grey go down and thrash about I stopped caring about the result and felt hugely sad for the horse and whilst horses being euthanized and hurting themselves isnt a new concept to me I felt bad about this one
I totally get it @Carter The grey is actually fine and running again this Saturday but it was the horse he brought down that died. Very, very sad. Cheers mate.
Thanks again peanuts for making me interested in something I would otherwise let pass me by
Sadly I think that's it for me and hunt racing, when I saw the grey go down and thrash about I stopped caring about the result and felt hugely sad for the horse and whilst horses being euthanized and hurting themselves isnt a new concept to me I felt bad about this one
I totally get it @Carter The grey is actually fine and running again this Saturday but it was the horse he brought down that died. Very, very sad. Cheers mate.
Yeah I tend to agree with Carter, Up For Review twitching in pain on the floor really ruined the spectacle for me this year. That is ultimately my main memory from the race.
Thanks again peanuts for making me interested in something I would otherwise let pass me by
Sadly I think that's it for me and hunt racing, when I saw the grey go down and thrash about I stopped caring about the result and felt hugely sad for the horse and whilst horses being euthanized and hurting themselves isnt a new concept to me I felt bad about this one
I totally get it @Carter The grey is actually fine and running again this Saturday but it was the horse he brought down that died. Very, very sad. Cheers mate.
Yeah I tend to agree with Carter, Up For Review twitching in pain on the floor really ruined the spectacle for me this year. That is ultimately my main memory from the race.
I felt the same as you @ValleyGary even though my only bet on the race was the Grey, Vintage Clouds who fell at the first bringing down Up for Review. My personal disappointment disappeared over my £10 loss and just felt heartache for the stricken horse and stable lad or lass.
Comments
One of my fave threads on CL and really enjoyed it, didn't win but the knowledge ITT is unreal. Thanks all
I had Rathvinden who looked good but Tiger Roll is just too good. Combi tricast Tiger Roll, Rock the Kasbah & Anibale Fly. not the worst effort. Quality race RIP to Up for review
As I said on Saturday, I had Tiger Roll on an e/w bet with William Hill. Put £2.50 down. Picked up my winnings this afternoon & got back £8.50. Certain the lady paying me said the odds were 4/1. Obviously all done electronically now so no reason to think she can't count or I've been diddled / had one over them. Just can't fathom how I got back £8.50 on a £2.50 stake if the odds were 4/1.
Just to clarify......I also had £2.50 each way (so a £5 total bet) on Livelovelaugh which I don't think counted on my betting slip as it wasn't top 5.
Any thoughts ??
£1.25 to win @ 4/1 = £6.25 return
£1.25 to place @ 4/5 (fifth of the odds) = £2.25 return
He must've run a longer race somewhere before that National ?
Until you begin to get a full and detailed understanding of betting I wouldn't include this in the portfolio of investment options you offer any of your clients!
Morning all,
Before signing off on this year’s GN thread, in case you’re interested, my views on how my new (Mark II) GN Model performed.
As always, it’s been a fun thread to participate in, regardless of Saturday’s result. Obviously, the result represented a loss for me and sincerest apologies for leading you astray.
It was, in my view, a case of the workman and not the tool that was to blame. In fact, I reckon the new Model did a pretty decent job. The evidence?
The Model Mark II
On the going (GS or better), it identified 5 with Winning Potential. All had ratings within the range of the top 90% of its back-tested ratings of 22 GN winners/near-missers from 2005. 2 of them had ratings consistent with the top 25% of former winners/near-missers.
So, those 5 were rated by the model (and came home on Saturday) as follows:
Getting 1st and 5th from 5 candidates is a decent result, particularly as 2 of the other 3 Fell (VC out of the running at the first). But there’s a further observation worth making:
The Model also identified 4 with Strong Place Potential, “next best” in stat-rating, consistent with profiles of bottom 10% of former GN winners/near-missers and a number of others making the frame (rating = c.15L behind TR/RDT):
So, its top 9 selections produced: 1st, 3rd, 4th and 5th (i.e. from 22.5% of the field, it found 67% of the 1~6 paying places and 80% of the first 5 home) – 3 of the other 5 Fell.
As it happens, that’s 1 more of the first 5 (or 6) places than were filled by the Virtual Grand National’s top 9 selections (represented by its predictions of the first 9 home).
So why the loss?
The model’s designed to make my GN bets as objective as possible - ordinarily I aim to back all the Winning candidates (scaled stakes) each-way. But, of course, the collective odds of that group dictate the risk-reward profile of that approach and moving parts (final weights, final preps, odds, bookies’ terms and going) make the process of translating the model's output to a betting slip more art than science - some years more than others. And that’s where things have gone awry for me 2 years running.
I won’t bore you with the whys and wherefores - the simple fact is that, with 3 of the 5 Winning candidates at/near the head of the market, my normal approach simply wasn't economically viable. In glorious retrospect, I should have had a stake-saving win bet on the Tiger (as did the far more astute @PaddyP17) but instead I had the bright idea of giving him (once again!!) the swerve, in favour of a 50/1 selection from the next best 4.
The outcome? Of the top 9 identified by the model:
· 5 Backed = 1 place, 4 fail
· 4 Not Backed = 1 win, 2 places, 1 fail
CONCLUSION: Tool = OK ........ Workman = tool.
To be continued ........... "To tweak or not to tweak?"
Human fallibility is human fallibility, so your interpretation of the model is what it is, and regardless of whether I win or lose, I find placing my shillings as a result of your thoughts (and my own responses to them) very enjoyable.
That said, a couple of questions do abound:
- Have bookmakers' odds ever come into the model's factoring?
I realise SPs are probably the only thing you can go off, unless (for instance) you're able to recall heavily punted fancies, like Baie Des Iles coming in from 66s to 14s last year.
But, I would possibly guess that - certainly with this year and last - your choice of scaled each-way bets means you're poking at generally higher odds, and thus those you leave unbacked at shorter odds "show you up" (absolutely not a criticism btw) a little?
Also - the Mighty Tiger winning at 12s and then 4s(!) also helping establish the "strength of the bookies might be worth considering.
I realise you said in the immediate aftermath that to be successful you need an identifier of "value", so the above question might be moot, but even so.
- How do you weight the significance of each individual stat/profiling measurable?
Not to use the same horses over and over, but my clearest memories here are of this year and last. So, with Baie Des Iles and Ramses De Teillee both running below expectations as seven year olds, have you thought of adjusting the model accordingly?
That said, I realise that very few 7 y/os have fit a winning profile anyway, but those that have (or that were SPP/MPP) might not run to form because of their age?
- More a follow-up to my previous question, but I presume certain things are more important? For instance, facing the GN fences is a more important variable than, say, age?
-------------------
There are probably a few more questions that'll come to mind, but either way, I'd be intrigued to know some of your thoughts. It's an immense undertaking, all for one race (albeit one of the greatest races), and I'd love to know more!
All good questions. If you don't mind, I'll get back to you after I've posted a reply to Chunes and described the tweaks that are appropriate for the model. Sets a bit of context to answering the questions you put. Please bear with .....
Getting back to you @Chunes
I'll post shortly about whether/why/how I’ll tweak the new model based on Saturday's result ….. if only to serve as another sleep-aid
To answer your 2 questions, both not boring (well not to me and you) and highly pertinent:
1. I’ve not normally re-tested races with a former version of the model and, for a couple of reasons, I don’t particularly want to this time.
· The old model was for another era, referring to a database for which 23 of the 28 GNs were over appreciably stiffer fences (and a longer trip, as it happens), many of which also had significant numbers running from out-of-the-handicap. I don’t think referring back to that system is going to be all that helpful for the future, to be honest.
Not that the old model was broken by last year's result (it had 1st, 3rd and 5th, though I naturally only had 5th on my slip
) but, as Pleasant Company was first “outlier” to go close to winning
since I created it in 2006-07, it felt time for a revamp based on a more
relevant database.
· Leaving my betting faux-pas aside, the new model really did a pretty decent job.
Expanding on my earlier post today, there was only 1 “outlier” in the first 8 home - even One For Arthur (25L 6th), Regal Encore (28L 7th) and Singlefarmpayment (28L 8th) all ran broadly to expectation (all in a group of 10 runners rated c. 20L behind TR/RDT).
Admittedly, it was a significant outlier - the 2L runner-up - but I know (from sad, indelible familiarity) that no version of the old model would have predicted Magic Of Light’s run (she was expected c. 30L behind TR).
2. If and when I tweak the new model this time, I’ll do what I always do and back-test it against the whole field (inclusionary) and all previous GNs (winners and finishing <30L) in the database to ensure it "fits" with them at least as well as the untweaked version had. I get particularly excited (sad isn't it?) if I find that a tweak to reconcile a runner this time also helps reconcile a small anomaly that I’d let ride previously – it does happen and, naturally, gives me confidence that it’s a “good tweak”.
But, to emphasise, a tweak to reconcile an outperformance is only likely to be a minor adjustment to a threshold, bestowing/removing a +/-. As we talked about previously, I only ever look to reconcile out-performances that marginally, but crucially, missed expectations (e.g. I’ll be looking closely at Rathvinden’s stats to see if there’s a borderline miss of a + for which it might be appropriate to make a tweak)
Of course, there are underperformers too but, unless there’s a clear logic as to why that may have happened (possibly there is a common one this time for RDT and General Principle), we have to accept that racehorses may have off days or a problem we may never hear about (e.g. a bleed, lose a shoe).
Just to expand on the evolution of the new model. As said, it uses a more recent database but there’s a good deal in common between the reference factors used by the old and new models.
It was in reconciling Pleasant Company (he’s from the family of Comply Or Die) that I decided to incorporate pedigree (for the first time) to create Mark II. This is the main departure and I think is a very interesting area (a few specific bloodlines are represented surprisingly often in winners/near-missers of GNs, pre- and post-2012); one I’m intending to research more.
Otherwise, the new model’s built on the same principles and uses largely the same factors as the old one. The formula’s slightly different, reflecting slightly different “tests” within the factors and slightly different +/- weightings.
Rattling off the factors (in no particular order):
I hope that makes sense but happy to clarify.
Well done on breaking even
And how many horses go through that process and do you only start with first declarations?
That quickly weeded out about 30 this year, so I went through the whole process for about 50.
14 runners on Saturday had initially been screened out - nearest finisher was Outlander (9th - 31L). Worked OK.
Cheers.
Last few winners been 8 or 9 year olds, 4 of the first 5 were 8 or 9 this year. Think it's going towards the traditional look for the upwardly mobile younger horse as opposed to the more experienced stayer that was typically well suited to the test of the race previously. Certainly don't think it'll be long before a 7yo wins it (hopefully next year, *cough* Mister Malarky *cough*)
Still maintain Tiger Roll at 4s is one of the worst bets you'll have, regardless of the result.... did nearly come down a few times and some of his main rivals fell by the wayside. Will certainly be taking him on again next year....no room for sentimentality in this game!
Totally agree @PolzeathNick
Re-viewing the race, it's remarkable how inconsequential even gross mistakes at fences have become. To see Magic Of Light absolutely plough through the Chair like it wasn't there (like Moses parting the Sea) and go on to be 2L runner-up (after ploughing through the last as well) suggests that the fences appear to have become since 2013 (literally) softer still by their construction.
Amazing though he is, I seriously doubt that Tiger Roll could have won a GN over pre-2013 fences. While the handicapper’s 172 rating of Saturday's win is relevant (in making it certain he’ll carry 11.10 if he returns), for the Racing Post to dub him therefore “the best GN winner in at least 46 years” is meaningless crap IMO.
I'm on your side, and think the fences have got more give in them, but is there any merit in his point?
She did fantastically well but the point still stands about the changing nature of the race that, in former times (I would contend even since 2013), that mistake at the Chair would (99% of the time) have ended any horse's race, one way or another.
It's better seen head on but this is the best I can do with a still from the normal view.
She's actually in the middle of the fence (yellow sleeves & noseband of course), having gone into it no more than half way up. There appears to be no core to it whatsoever. Bizarre. We're talking the most feared obstacle on the course bar Bechers.
All credit to her for a fantastic run and she clearly has bucket loads of guts and stamina - but is this really "The Grand National"?
Sadly I think that's it for me and hunt racing, when I saw the grey go down and thrash about I stopped caring about the result and felt hugely sad for the horse and whilst horses being euthanized and hurting themselves isnt a new concept to me I felt bad about this one
The grey is actually fine and running again this Saturday but it was the horse he brought down that died. Very, very sad.
Cheers mate.