 |
logic_man
Registered: Oct 2010
Posts: 1489 |
08-17-12 07:47 PM
I was wondering if anyone knew about a website or any tools that can be used to generate random trades against which to test your own results, to see if they are significantly better than random.
I was reading an article a few days ago and the author used the profit factor of a random trade as his benchmark for testing the significance of his own profit factor. He used a PF of 1.2 for long trades and 0.8 for short trades, but he didn't talk about where he got those values or if they were from a specific time period only or if they were the PFs for any random long or short trade on the S&P 500 index for all time periods.
|
| |
|
Edit/Delete • Quote • Complain |
dtrader98
Registered: Nov 2006
Posts: 1818 |
08-17-12 08:12 PM
Excel.
=IF(NORMSINV(RAND())>=0,1,-1)
Run many times will give a sampling distribution of signals to compare.
It makes more sense to run against same period as system validation set.
|
| |
|
Edit/Delete • Quote • Complain |
dom993
Registered: Jul 2008
Posts: 540 |
08-18-12 03:59 AM
There is absolutely no need for doing that re. intraday trading ... taking random entries & exits will generate about minus 2-ticks (+ commissions) per trade in the long run.
You might want to test your trade management against random entries, or your entries against random exits - but what's the heck?
The thing about taking random entries, is for long-term trend-following strategies.
|
| |
|
Edit/Delete • Quote • Complain |
mgabriel01
Registered: Oct 2007
Posts: 975 |
08-18-12 06:01 AM
Quote from logic_man:
I was wondering if anyone knew about a website or any tools that can be used to generate random trades against which to test your own results, to see if they are significantly better than random.
I was reading an article a few days ago and the author used the profit factor of a random trade as his benchmark for testing the significance of his own profit factor. He used a PF of 1.2 for long trades and 0.8 for short trades, but he didn't talk about where he got those values or if they were from a specific time period only or if they were the PFs for any random long or short trade on the S&P 500 index for all time periods.
Ninjatrader Simulator
|
| |
|
Edit/Delete • Quote • Complain |
jcl
Registered: Jan 2012
Posts: 407 |
08-18-12 08:16 AM
Quote from dom993:
There is absolutely no need for doing that re. intraday trading ... taking random entries & exits will generate about minus 2-ticks (+ commissions) per trade in the long run.
You might want to test your trade management against random entries, or your entries against random exits - but what's the heck?
The thing about taking random entries, is for long-term trend-following strategies.
In his example above, with the PF 1.2 for random long trades, the price curve he used was probably strongly trending upwards. In such a case a comparison with random trades might make some sense, but it's a better method to just detrend the price curve for the test. Detrended price curves keep all their other characteristics.
|
| |
|
Edit/Delete • Quote • Complain |
logic_man
Registered: Oct 2010
Posts: 1489 |
08-20-12 04:49 AM
Quote from dom993:
There is absolutely no need for doing that re. intraday trading ... taking random entries & exits will generate about minus 2-ticks (+ commissions) per trade in the long run.
You might want to test your trade management against random entries, or your entries against random exits - but what's the heck?
The thing about taking random entries, is for long-term trend-following strategies.
Wouldn't the results for intraday trading depend on how long you held? Or are you saying that, on average, for any holding period intraday, price is 2 ticks away from any other time during that day?
In any case, thanks for the tip that this shouldn't be a concern for more short-term trading.
|
| |
|
Edit/Delete • Quote • Complain |
| Receive
an email whenever a new post is added to this thread by subscribing
to it. |
|
|
|
|
