Statistical Validation on trading | Forex Factory

Options

Go to
Permalinks
Print Thread

Similar Threads

Statistical Probability Trading With Price Action 576 replies

Trading journal (trades based on statistical calculations) 42 replies

How do you gain, and measure, your trading system's statistical edge? 29 replies

Validation Date Code need... 5 replies

EurJpy Statistical Probability Trading With Price Action 46 replies

Statistical Validation on trading

  • #1
  • First Post: May 1, 2005 5:53pm May 1, 2005 5:53pm
I cant afford tradestation so I guess I'll have to do this manually,How many trades would it take to find out my true win\ loss ratios, downdraws, expectancy etc. of the system I am trading now?
  • #2
  • May 1, 2005 7:08pm May 1, 2005 7:08pm
I just keep track of every trade, then when I have 50 trades, I look at the win/loss ratio. You can see how many wins to how many losses, and what kind of pip profit/loss you have. I think any trading platform will give you a report. If you're asking for more in depth info, I don't know about that.
Because helping each other is a good thing.
  • #3
  • May 1, 2005 8:57pm May 1, 2005 8:57pm
Quoting howtotradeforex
I just keep track of every trade, then when I have 50 trades, I look at the win/loss ratio. You can see how many wins to how many losses, and what kind of pip profit/loss you have. I think any trading platform will give you a report. If you're asking for more in depth info, I don't know about that.
Hey howto,
After 50 trades do you see a consistency in the numbers like win\loss, etc? Thanks for you help.
  • #4
  • May 2, 2005 4:30am May 2, 2005 4:30am
Quoting bluemonkey
I cant afford tradestation so I guess I'll have to do this manually,How many trades would it take to find out my true win\ loss ratios, downdraws, expectancy etc. of the system I am trading now?
buddy, you cant afford NOT to have tradestation. its hard to play this game seriously with free software
Relax and be happy.
  • #5
  • May 2, 2005 10:06am May 2, 2005 10:06am
Quoting merlin
buddy, you cant afford NOT to have tradestation. its hard to play this game seriously with free software
Merlin, I do have to open an account with them in order to use the tradestation software right?
  • #6
  • May 2, 2005 11:11am May 2, 2005 11:11am
Quoting merlin
buddy, you cant afford NOT to have tradestation. its hard to play this game seriously with free software
Why is tradestation a must, Merlin? It is a recent phonomenon and people have been making money trading for a very long time. I can't believe that trading has changed so much that the only way to play 'this game seriously' is to use tradestation. That sounds very odd to me.
  • #7
  • May 2, 2005 11:58am May 2, 2005 11:58am
Quoting davide123
Why is tradestation a must, Merlin? It is a recent phonomenon and people have been making money trading for a very long time. I can't believe that trading has changed so much that the only way to play 'this game seriously' is to use tradestation. That sounds very odd to me.
David do you think it would be enough for me to test it by hand? is 50 trades enough or do I have to do hundreds of tests?
  • #8
  • May 2, 2005 1:20pm May 2, 2005 1:20pm
Blue Monkey,

Check out Ed Seykota's article here:

http://www.seykota.com/tribe/risk/index.htm

In that article he talks about stress testing. He says: "Stress Testing is a process of subjecting a model of the trading and risk management system to the historical data, and noticing the historical performance, with special attention to the draw downs."

Then he says: "The length of historical data sample for the test is likely adequate if shortening the length by a third or more has no appreciable effect on the results."

Ed is a pretty bright guy and a world class trader, so chances are this is a good rule of thumb. If you think your sample size is large enough to predict drawdowns, then make it 1/3 shorter. If the effects on the data are minimal, then you probably have a good enough sample size.

Word of caution - I'm not a competent system designer by any stretch of the imagination, so my opinion is probably worth exactly what you are paying for it.

James
  • #9
  • May 2, 2005 3:37pm May 2, 2005 3:37pm
Quoting Redleg
Blue Monkey,

Check out Ed Seykota's article here:

http://www.seykota.com/tribe/risk/index.htm

In that article he talks about stress testing. He says: "Stress Testing is a process of subjecting a model of the trading and risk management system to the historical data, and noticing the historical performance, with special attention to the draw downs."

Then he says: "The length of historical data sample for the test is likely adequate if shortening the length by a third or more has no appreciable effect on the results."

Ed is a pretty bright guy and a world class trader, so chances are this is a good rule of thumb. If you think your sample size is large enough to predict drawdowns, then make it 1/3 shorter. If the effects on the data are minimal, then you probably have a good enough sample size.

Word of caution - I'm not a competent system designer by any stretch of the imagination, so my opinion is probably worth exactly what you are paying for it.

James
Thanks any help is appreciated by me.
  • #10
  • May 2, 2005 4:22pm May 2, 2005 4:22pm
Quoting bluemonkey
David do you think it would be enough for me to test it by hand? is 50 trades enough or do I have to do hundreds of tests?
There is nothing in principle wrong with 'doing it by hand'. I am not a big fan of backtesting and think that watching trades in real time or demo trading them is superior even though your data base is smaller. This is because you watch the trades as they unfold and study the behaviour of the price movement in real time and get a very good feel for what is happening. If I observed a pattern repeating itslef over 50 trades, I would certainly trade that pattern. Of course that is highly unlikely, since some of those 50 trades must have been losers. For that reason 50 is probabaly a bit too few. However, I would be much more interseted in the logic of the system that is being observed rather than the frequency of wins (assuming there are enough of them to give you a positive expectancy). This point is very important because the outcome of any test depends on the conditions of the market at the time.

You want to see that the system that results from your tests and observations is robust enough to handle both trending and sideway markets, for example. I hope this helps.
  • #11
  • May 2, 2005 4:55pm May 2, 2005 4:55pm
Quoting bluemonkey
Hey howto,
After 50 trades do you see a consistency in the numbers like win\loss, etc? Thanks for you help.
I'm not consistent yet, but an improvement that I've noticed is less consistent losses.
Because helping each other is a good thing.
  • #12
  • May 2, 2005 6:49pm May 2, 2005 6:49pm
Quoting davide123
There is nothing in principle wrong with 'doing it by hand'. I am not a big fan of backtesting and think that watching trades in real time or demo trading them is superior even though your data base is smaller. This is because you watch the trades as they unfold and study the behaviour of the price movement in real time and get a very good feel for what is happening. If I observed a pattern repeating itslef over 50 trades, I would certainly trade that pattern. Of course that is highly unlikely, since some of those 50 trades must have been losers. For that reason 50 is probabaly a bit too few. However, I would be much more interseted in the logic of the system that is being observed rather than the frequency of wins (assuming there are enough of them to give you a positive expectancy). This point is very important because the outcome of any test depends on the conditions of the market at the time.

You want to see that the system that results from your tests and observations is robust enough to handle both trending and sideway markets, for example. I hope this helps.
Yes it does very much so.
  • #13
  • May 2, 2005 6:50pm May 2, 2005 6:50pm
Quoting howtotradeforex
I'm not consistent yet, but an improvement that I've noticed is less consistent losses.
I can relate to that howto. I'm in the same boat.
  • #14
  • May 9, 2005 5:43pm May 9, 2005 5:43pm
Quoting bluemonkey
Merlin, I do have to open an account with them in order to use the tradestation software right?
nope. but if you have an account it costs $100 a month instead of $200.
Relax and be happy.
  • #15
  • May 9, 2005 5:54pm May 9, 2005 5:54pm
Quoting davide123
Why is tradestation a must, Merlin? It is a recent phonomenon and people have been making money trading for a very long time. I can't believe that trading has changed so much that the only way to play 'this game seriously' is to use tradestation. That sounds very odd to me.
i didnt say "'to play this game seriously' is to use tradestation".

i said its hard to play seriously with free software.

trading has not changed overall, but the amount of analysis that goes into building a successful mechanical trading system sure has. back in the day all you had to do was buy breakouts, but with the advent of computers the market became more efficient to most of the patterns that were once winners. this position is shared by most of the market wizards.

sure, you can spend days (or weeks) analyzing your trades, but with good software you can do the same task in seconds. literally. to not utilize the computer to its full capability is to be at a time disadvantage, and why i say that its hard to play this game with free software.
Relax and be happy.
  • #16
  • May 10, 2005 5:52am May 10, 2005 5:52am
Quoting merlin
i didnt say "'to play this game seriously' is to use tradestation".

i said its hard to play seriously with free software.

trading has not changed overall, but the amount of analysis that goes into building a successful mechanical trading system sure has. back in the day all you had to do was buy breakouts, but with the advent of computers the market became more efficient to most of the patterns that were once winners. this position is shared by most of the market wizards.

sure, you can spend days (or weeks) analyzing your trades, but with good software you can do the same task in seconds. literally. to not utilize the computer to its full capability is to be at a time disadvantage, and why i say that its hard to play this game with free software.
How much does Tradestation cost?
  • #17
  • May 10, 2005 10:32am May 10, 2005 10:32am
Quoting davide123
How much does Tradestation cost?
I pay £105.

F
  • #18
  • May 12, 2005 1:48pm May 12, 2005 1:48pm
Quoting davide123
How much does Tradestation cost?
are you admitting i won that debate?? YES!!! thats my first win in a while LOL (fiji is always schooling me )

...$200 a month.
Relax and be happy.
  • #19
  • May 12, 2005 3:27pm May 12, 2005 3:27pm
Quoting merlin
are you admitting i won that debate?? YES!!! thats my first win in a while LOL (fiji is always schooling me )

...$200 a month.
Merlin is it 200.00 per month does that include the data feed?
  • #20
  • May 14, 2005 4:44pm May 14, 2005 4:44pm
Quoting bluemonkey
Merlin is it 200.00 per month does that include the data feed?
unfortunately no. i spend about $200 per month for the platform and data. i get the platform for $100 because i have an account with them, and my data comes to about $100. forex data is free for the moment so you wont have to buy that.
Relax and be happy.
0 traders viewing now