counter_violent
Legendary member
- Messages
- 12,670
- Likes
- 3,787
Are you asking on behalf of a friend ?
arabianights
Legendary member
- Messages
- 6,721
- Likes
- 1,380
um
you need to define that before testing the system, as it depends upon the system, fairly obviously.
If you have any sense you decide how much you are willing to lose and then tweak your position sizing so within a given confidence level you can be sure of the invalidity of the system if your drawdown is that much...
Next week: Karl Popper for dummies
you need to define that before testing the system, as it depends upon the system, fairly obviously.
If you have any sense you decide how much you are willing to lose and then tweak your position sizing so within a given confidence level you can be sure of the invalidity of the system if your drawdown is that much...
Next week: Karl Popper for dummies
Lol, no its not one of those my “friend has a nasty rash on his ......” It is a hypothetical question. With some questions behind it.
B
Black Swan
I'd say 15% but that would have to happen over a period of *sensible* trading time, with appropriate gaps to go on benders, get off your face in the Hacienda....oops...always warned there'd be flashbacks...
If you spunk even the 15% inside a week there's something wrong imho...but having said that IF I spunked 15% in a week or two I probably wouldn't panic these days...be fooking weird though...1% risk per trade, 1:3 risk reward ratio target. Flat for the weekend. You are happy with the back test.
arabianights
Legendary member
- Messages
- 6,721
- Likes
- 1,380
I just removed an extremely rude post towards mr swan that wasn't even remotely justified, luckily before I posted it.
Basically Jason you're looking at this from the wrong angle. You don't tell if a system works by the magnitude of any drawdown, you work on the probability of any particular drawdown and if the probability of that arising by chance in a winning system is less than x% (chosen by yourself, again - and the statistical methods for working this out again have to be chosen by yourself) then if that drawdown is reached then there is a 100-x% chance that your system hasn't the value you thought it had.
Capice?
Basically Jason you're looking at this from the wrong angle. You don't tell if a system works by the magnitude of any drawdown, you work on the probability of any particular drawdown and if the probability of that arising by chance in a winning system is less than x% (chosen by yourself, again - and the statistical methods for working this out again have to be chosen by yourself) then if that drawdown is reached then there is a 100-x% chance that your system hasn't the value you thought it had.
Capice?
B
Black Swan
aw 5hit, can you pm me it...please...?
arabianights
Legendary member
- Messages
- 6,721
- Likes
- 1,380
Arabian, I would agree with you. This is the sort of process I would go through, good to be confirmed by a profitable trader though. I do Capice and have Capiced on this.
I'd say 15% but that would have to happen over a period of *sensible* trading time, with appropriate gaps to go on benders, get off your face in the Hacienda....oops...always warned there'd be flashbacks...
What I am really driving at though is in Black Swan’s reply. I have seen several traders, who I am sure are profitsble and better traders than me, talk on BB about opening an account for a new system and risking 1% per trade (or so). Now if you have an uncle point on a draw down it is going to be no where near 100% of the account at 1% per trade on these sort of R:R ratios. So why do so many traders do this?
arabianights
Legendary member
- Messages
- 6,721
- Likes
- 1,380
I'm sure they do but I really think you're going to have to tailor the testing to the system. I doubt that each trade is independent, to start with... all systems are going to take advantage of certain market conditions and either you have a way of identifying that or you're going to have to adjust for it and at that point it's well beyond me mate.
You sound to me like you have your head screwed on, probably much better than me - and I'm sure know that - it's for the advantages of the lurkers I'm posting this...
My personal view would be to either simulate, as best as possible, the fund you want to take your system to, and whatever acceptable draw down on that would be, or alternatively be prepared to see it all disappear...
You sound to me like you have your head screwed on, probably much better than me
- and I'm sure know that - it's for the advantages of the lurkers I'm posting this... My personal view would be to either simulate, as best as possible, the fund you want to take your system to, and whatever acceptable draw down on that would be, or alternatively be prepared to see it all disappear...
B
Black Swan
I lost 3% of one account Wednesday last week and tbh (with hindsight) I deserved a good smackdown..well overdue. I had a good long think that evening and immediately binned the shorter TFs and went back to 3-4hr the next day. I could spot the problem, too much risk (sometimes in one direction) with up to 7 pairs all at the same time, I want to take each and every trade that matches the criteria, but having 8-10% of an account on the line at any one time was stoopid...
Overtrading, over leverage, over confidence...'kin nob 'ed..So back to the longer TFs and for once I'm more relaxed, no more constant tinkering/playing a few scalps on cable like its chavvy scratchcards ..perhaps the lesson was necessary for my psyche?
Important issue was immediate recognition that something was deeply wrong, no point in others telling me I had to discover it myself and the lesson wasn't too hard to spot and the remedy simple. I think on reflection it could have only happened once big (8-10%) before I'd have changed something and if I hadn't bailed all 4 trades I'd have lost 2 grand...not good...
I lost 3% of one account Wednesday last week and tbh (with hindsight) I deserved a good smackdown..well overdue. I had a good long think that evening and immediately binned the shorter TFs and went back to 3-4hr the next day. I could spot the problem, too much risk (sometimes in one direction) with up to 7 pairs all at the same time, I want to take each and every trade that matches the criteria, but having 8-10% of an account on the line at any one time was stoopid...
So back to the longer TFs and for once I'm more relaxed, no more constant tinkering/playing a few scalps on cable like its chavvy scratchcards ..perhaps the lesson was necessary for my psyche?
Important issue was immediate recognition that something was deeply wrong, no point in others telling me I had to discover it myself and the lesson wasn't too hard to spot and the remedy simple. I think on reflection it could have only happened once big (8-10%) before I'd have changed something and if I hadn't bailed all 4 trades I'd have lost 2 grand...not good...
Great Post! What can I say?
This is a great reminder for us all to treat trading with as much respect as you would for any other of your own businesses.
Thanks BS
Jason
* Alot of people claim "to risk no more than xx% on any one trade", but of course we are all customers with yachts.
* there are plenty of other options - Kelly, Optimal F etc...
* Have you tried treating your system results as a binomial series -> normal approximation? Then do hypothesis testing on your expectancy results?
* A stage further might be to test for autocorrelation (or more appropriately breakdowns in autocorrelation as an indication that your strategy does not operate most effectively in the current market conditions)
* The maths can't take into account that you gotta be able to sleep at night
* there are plenty of other options - Kelly, Optimal F etc...
* Have you tried treating your system results as a binomial series -> normal approximation? Then do hypothesis testing on your expectancy results?
* A stage further might be to test for autocorrelation (or more appropriately breakdowns in autocorrelation as an indication that your strategy does not operate most effectively in the current market conditions)
* The maths can't take into account that you gotta be able to sleep at night
Good suggestions MrGecko,
I am very keen on Kelly. I think of Optimal F as a little to much like wishful thinking for my point of view.
The application of binomial series -> normal approximation really fascinates me. As a kid I would spend hours tossing a coin in my bedroom (at least that’s what I told my mum I was tossing) with a calculator in the other hand. Looking back, probably a strange pastime for a 12 year old boy.
As for testing for autocorrelation, this goes a little over my head and a little to deep at the moment to be honest with you. Maybe this is something I can look at in the future.
Thanks again for your ideas MrGecko, I am inspired to take position sizing more seriously.
Jason
B
Black Swan
no more constant tinkering/playing a few scalps on cable like its chavvy scratchcards ..perhaps the lesson was necessary for my psyche?
Classic
Just being serious for a 'mo in some ways I envy those guys who play those cards and or who I see in the betting shops (through the windows
) scrunching up their slips...they seem to have far less emotional content/imput/hang ups than us lot when it comes to the simple game of *putting one on*...Perhaps I'll grab a few of 'em and make a chavtastic team out of them...be funny as fook if they all hit the ground running...
"Is that it like, when those three things line up you place the bet and get out when you've a few quid more or when those things cross and line up again...? 'kin 'ell, this is well better than the gee gees, it's only a two horse fookin race and you can change the jockey mid race, bail out the race early if it's goin bad, or double up if you're in the lead...!!"
Masquerade
Senior member
- Messages
- 2,543
- Likes
- 284
I'd say £0.
That's actually quite conservative - this is a leveraged game after all. For a lot of traders £0 might not be such a bad result
.
