Backing Drifters

[ian2005]

Hi everyone,

Just found this article, few months old now -

Published: 04/03/2005. David Ashforth, Racing Post

Backing horses who drift in the market is more profitable than backing those whose odds shorten, according to Betfair. The betting exchange`s claim represents the latest ingredient in Betfair`s attempt to counteract allegations that exchanges pose a threat to racing`s integrity, with non-triers allegedly being laid on the exchanges, where their price drifts before the horse involved loses.

On the basis of an analysis of six months` data, from the second half of 2004, Betfair said: "The evidence of nearly 2,000 races and 20,000 horses shows that more money would have been made backing drifters than backing horses that shortened."

Defining a `drifter` as a horse whose price moved out by more than 20 per cent during the final five minutes before the official start time of each race, and a `shortener` as a horse for whom the reverse was true, Betfair claimed: "The analysis from every runner in 1,864 races over the summer and autumn of 2004 shows categorically that punters who bet on horses that drift in the betting in the final five minutes do better than those who bet on horses that shorten in the market."

A £1 level-stake bet immediately before the start of each race produced a loss of £345 on shorteners but a profit of £570 from backing drifters.

A Betfair statement added: "These results completely contradict the stated view of those who use anecdotal evidence to project the idea that the biggest drifters in the betting market are `prepared to lose`, and thus represent bad value bets for punters.

"Detractors of exchanges are keen to highlight horses that drift and lose but conveniently forget those that drift and win. This analysis serves as a timely reminder that there are many reasons why horses drift in the market.

"It proves that the view that a drift in price indicates some form of skulduggery is a lazy and incorrect one.

"Drifting horses do not drift on the basis of either nefarious activity or inside information; they drift because market forces dictate the price and, as financial markets show, market forces can get it wrong."



------------

I just wondered how the profits of this compare with the systems on here?

It says (for 6 months only I realise) that at £1 level stakes would have returned £570 profit. Am I correct in thinking I could just multiply that by 10 to get say £5700 to £10 stakes and compare it to the JTS at £2955?

I could build this system easily enough and wondered whether it would be worth it. What's the catch or what have I missed!!

Many thanks,

Ian

[bigcumba]

Interesting idea Ian... I think this was mentioned on ATR on Saturday or Sunday as well... I'd want to monitor it for a while to be certain, and to build up a picture of how much a drifter has to be moving before it becomes a bet.

[Profit Seeker]

There's got to be software that can inform you of when a horse has gone out 20% in such time. But you'd be backing every outsider with massive odds, I find it hard to believe those claims. But who knows....probably Keith.

[bigcumba]

Actually I think the RP commentary box now has a 'market movers' feature..

[presto]

i personally wouldn't try it.

Defining a `drifter` as a horse whose price moved out by more than 20 per cent during the final five minutes before the official start time of each race

to identify these drifters 'in real time' before a race would be an enormous task. especially the longer odd's.

placing the bet's would also be very difficult. as odd's often drift 20% then come back in. in the last 5 mins. also 'the last 5 mins' can also be very difficult to define. as races never go off on at the defined time. you could place a bet (after 5mins of drifting) only for the race to go off 2mins early and, to shorten again.

there will also be many qualifiers, quite possibly 1/3 of the race.

if you did decide to look into it a bit more then why not click the 'bot' link at the top and get a quote on a piece of software to help out.

[ian2005]

Hi guys,

Thanks for the advice. Yeah I'm going to run a 'dummy test' over x amount of time to see what happens, nothing to loose.

I can write my own bots, certainly not thinking of doing this by hand!!

One thing I guess nobody would really know is the size of the loosing streaks on this thing, could imagine they would be loong.

I agree presto, the odds drifting and coming back in could be a problem. Simply don't know if betfair counted these. Think I'm going to take a snapshot of the odds 5 mins before official start time then 30 secs before and then back anything that has drifted 20%+.

Cheers,

Ian

[Win2Win]

All depends on the definition of 'drifter' and is open to different results for different people the same as the Pro-Losers.

[ian2005]

Well I ran the bot today on UK only races -

1 winner out of 53 selections found....

and that was Spunger @ 8.4

I guess if this system did actually work you'd need a huge bank to start with!

[Merlin]

Betfair, of course, dont make it clear what they monitored........precisely...

If they monitored the whole market , this could be said to be disengenous (in respect of counteracting claims of skulduggery) because of course, skulduggery is usually claimed when fancied horses drift.....

It would be more interesting I think to run a trial on say the first 6 - 8 in the betting - coz nobody complains when a 16/1 shot loses......

[icebreaker]

Ian, I'm following your thread with great interest.

May I ask two questions, please?
1) Across how many races were the 53 selections spread (I'm presuming there were some races which had more than one 20% drifter)
2) From what point in time before the race do you (or the bot) begin monitoring for the drift in price. In other words a drift in offered price could begin to be felt during the night, hours before racing begins. Or do you wait until a very strong market is formed (say 1/2 hour before the off).

Thanks in advance.

[ian2005]

53 seletions over 19 races today, yes some races had 3 or 4 and some had none.

The bot takes a snapshot of the odds of each race 5 mins before the official race start time and then again 15 seconds before the offical start time. Compares the 2 sets of odds and check whether any horses have drifted 20% or more.

Using only the final 5 mins before the race because that's whats quoted by betfair in the article.

Merlin - yeah I might just do that about the first 6-8.

[icebreaker]

Okay, Ian. That's great ........ thank you.

[Vic]

I'm with Merlin on this one - I've long believed that backing horses from the first 4-6 in the betting that drift from 1st show would be profitable. Invarably if a horse is drifting at bookie prices it is racing away on the exchanges as the layers try to lump on, and then pulls back in somewhere closer to the bookie odds just before the off. I've no hard facts, but from what I've observed I believe it to be true.Ladbrokes market makers are not stupid, if they have 3/1 for a runner at first show - and it drifts to 9/2, and while it's drifting you can get over 5.5/1 on the exchanges - there is a fair chance you have a value bet. It's also rare that you can't lay it off in running for a free bet of at least even money.

[Yoda]

I wouldn't recommend this to anyone.

I made quite a large bet on them to be X-mas number 1 in '78 only to be left very disappointed.

tricky business.

Jonah

[Win2Win]

Yoda!! How dare you try and change the tone of this serious thread, this is a respectable forum.......I'm off to play with Willy Wonka & the Chocolate Cream with the wife :D

[Profit Seeker]

First race, Villaricca was the biggest drifting horse before the off. don't know if it was as far as 20%, won 10/1sp

[markwales]

Just had an email from Betfair asking the dirfters question......

Anyone else received it?

[Iron Chris]

Yes.... :D

[wb]

does anyone think it would be possible that a name of a horse could affect its price. Im not talking about famous horses or anything.

Im not sure how many joe punters there are out there who back 'fun bets' by name, but my guess is that there are quite a few of them. (crazy people!)

say, just for example, in the last race, on a day that had been good for the bookie, where the punters have lost out a lot, and there is a horse called something like 'last chance' or whatever,

do you reckon' the amount of people that may go with the name will affect its price?

If so, are they distorting the market much?

[ian2005]

Yep did better today 4 out of 89 selections -

Villarrica Salisbury 15.00 W
Benwilt Gold Downpatrick 40.00 W
Tony Tie Newcastle 36.00 W
Summer Charm Lingfield 13.50 W

[Profit Seeker]

Another example, Cleaver in 3.40 Newcastle, the biggest drifter by far! I went through each horses graph a minute before the off, Cleaver had an almost vertical line hit about 30 before the off then went to 50 straight away in running. Wish I'd had a few quid now :(

[Win2Win]

You won't make much at those prices PS, you need something under 5.0