I have just read peter's "Lay the Draw" blog and it is very interesting (as normal) the problem it poses for me is
How can you be sure with any strategy that it is profitable in the long run or just deviation from the norm?
The blog shows even a sample of 15,000 may not be enough and if you say have a strategy around say Soft ground horse races over chase fences on a Sunday you could be lucky to have 50 occurrences.
"Lay the draw" blog
Hi Buyshirts
The problem with backtesting a system is that is tells you about the market's historical inefficiencies. However, the challenge you're up against is that many people - including Betfair themselves - are looking for an edge, meaning an edge based on fundamentals is prone to disappear once the cat is out of the bag and everyone rushes to exploit it.
IMHO, one thing that doesn't change in the market is irrational human behaviour, so if you can find an edge that exploits market irrationality, you'll have a long-term edge. Also, if you're good at gauging the true price of a horse, greyhound, football team, etc winning, then you will be able to obtain value that way too.
An edge that once existed and may still exist is outlined in an academic paper (http://ihome.ust.hk/~dchoi/Betfair_ChoiHui.pdf):
'We find that whether bettors over- or underreact to the first goal of a match depends on how “surprising” the goal is. For a goal that is “expected” (i.e., the “favorite” team scores first) or only moderately surprising, bettors tend to underreact. In contrast, bettors tend to overreact when the goal is very surprising.'
Personally, I won't be trying to exploit that possible edge, as it may well disappear in the future (hence why I'm happy to post it here!).
Another fundamental based system is to back horses run by trainers who have a good ROI and strike rate at the course in question. A few years ago, someone who used that system sent me a spreadsheet with highly impressive results, but I've no idea if it still works...
Jeff
PS This is the link to the blog post: http://www.betangel.com/blog_wp/2012/04 ... revisited/
The problem with backtesting a system is that is tells you about the market's historical inefficiencies. However, the challenge you're up against is that many people - including Betfair themselves - are looking for an edge, meaning an edge based on fundamentals is prone to disappear once the cat is out of the bag and everyone rushes to exploit it.
IMHO, one thing that doesn't change in the market is irrational human behaviour, so if you can find an edge that exploits market irrationality, you'll have a long-term edge. Also, if you're good at gauging the true price of a horse, greyhound, football team, etc winning, then you will be able to obtain value that way too.
An edge that once existed and may still exist is outlined in an academic paper (http://ihome.ust.hk/~dchoi/Betfair_ChoiHui.pdf):
'We find that whether bettors over- or underreact to the first goal of a match depends on how “surprising” the goal is. For a goal that is “expected” (i.e., the “favorite” team scores first) or only moderately surprising, bettors tend to underreact. In contrast, bettors tend to overreact when the goal is very surprising.'
Personally, I won't be trying to exploit that possible edge, as it may well disappear in the future (hence why I'm happy to post it here!).
Another fundamental based system is to back horses run by trainers who have a good ROI and strike rate at the course in question. A few years ago, someone who used that system sent me a spreadsheet with highly impressive results, but I've no idea if it still works...
Jeff
PS This is the link to the blog post: http://www.betangel.com/blog_wp/2012/04 ... revisited/
Life gets quite interesting once the cat is out of the bag.Ferru123 wrote:Hi Buyshirts
The problem with backtesting a system is that is tells you about the market's historical inefficiencies. However, the challenge you're up against is that many people - including Betfair themselves - are looking for an edge, meaning an edge based on fundamentals is prone to disappear once the cat is out of the bag and everyone rushes to exploit it.
If 'the cat' starts to attract a little money,the market inefficiency remains but on a reduced scale. Although profit can still be made,it is reduced.
If 'the cat' attracts yet more money,the market inefficiency disappears,becomes completely efficient and results in a break-even situation.
If 'the cat' attracts even more money,the market inefficiency appears again,but in the opposite direction.In this case,laying the draw creates a loss and backing the draw becomes profitable.
But,I suspect that it isn't even that simple.
Once the cat is out of the bag,people start laying the draw.The market becomes less inefficient and layers make less profit than hoped for.Some will be happy with the smaller than expected profit and continue to lay the draw.Others will be unhappy and will stop.
Those that stop will cause some inefficiency to reappear and so more people will join the fray.As a result,the market becomes less inefficient and layers make less profit than hoped for.Some will be happy with the smaller than expected profit and continue to lay the draw.Others will be unhappy and will stop.And so the cycle continues.
A classic case of damned if you do and damned if you don't.
In the lay the draw situation,it looks to me like a case of a market inefficiency opening up due to randomness and then closing down for similar reasons.
74.5
The way to get over large drawdowns is to increase the margin you seek when you are looking for value. The bigger the margin the less you need to worry about drawdowns. When you are looking at a trading strategy then the draw down and reversion should be tighter as your strike rate will be higher.
Part of seeking value is finding enough to over come these bad runs against you.
Data mining can also mislead people into thinking they have found an edge. The only way to really learn quickly is to put it to use in the market. The odd thing I have found is that by participating in the market you end up influence the outcome, so you need to take that into account as well. Value seeking can be a tough business.
Part of seeking value is finding enough to over come these bad runs against you.
Data mining can also mislead people into thinking they have found an edge. The only way to really learn quickly is to put it to use in the market. The odd thing I have found is that by participating in the market you end up influence the outcome, so you need to take that into account as well. Value seeking can be a tough business.
I understand the greater the margin of profit the protections that gives you in assessing your system.
What rate of return on top of commission do we think is enough to say a system is A) profitable B) Have enough inbuilt margin to give us time to evaluate the system?
What rate of return on top of commission do we think is enough to say a system is A) profitable B) Have enough inbuilt margin to give us time to evaluate the system?