The market certainly has changed and we have to look at new methods to make money. The old swing trading stops us out too early, and as soon as we are out, the market moves up again.
I have found that day trading and very short trading gives me the best profits at the moment.
I am putting together a revised plan for this type of market. Any ideas would be welcome.
I have had limited success day trading index CFD's, I start by identifying a trend using hourly charts, then using a 3 min chart for an overbought/sold situation, then a 1 min chart for entry. I started looking at a system last week, I'm only trading stocks now in extreme oversold situations. Would be interested in your comments or system Eric. porkpie
Ok guys! Just ran a back test on this system with SPI day session data for this year.
This is my first every back test so, pick away. Parameters as follow,
From your sample were you able to identify at which times you were better off staying out of the market or waiting until a direction was established enough to trade to your advantage?
Ok guys! Just ran a back test on this system with SPI day session data for this year.
This is my first every back test so, pick away. Parameters as follow,
Running a test of the 7/15 cross on a 2 min chart based on the morning session( up to 11:30 am as per the first post in the thread) on a lower vol environment(July 06- Jun07), and allowing for $5 a side brokerage + 1 point slippage on entry and exit(for a grand total of $30 cost entry + exit per trade), based on 24 hour data on a single contract traded gives the following results-
683 trades
227 winners(33.24%)
456 losers(66.76%)
av win $285.81
av loss $176.67
payoff ratio 1.62
profit factor 0.81
max drawdown $-20075
Sharpe -6.95
Pretty dodgy results. But I think most of us knew that before it was even tested
Just to add, this is only half the method tested as described earlier by prawn- I have no idea how to code in the change over to a higher timeframe half way through the session. Running the same test based on 5 minute data, and trading the entire session instead of just the morning period actually manages to turn a profit....just. But with a profit factor of 1.17 it leaves the possibility of that profit quickly turning into a loss if slippage turns out to be greater in real trading.
Scratch both of the posts I made above- I didn't have the backtester set up properly- it was coded to buy the open of the bar where the cross took place, not the bar after, so those results should be considered quite flattering to the method, and real trading would result in much, much worse results
Ran another test on BHP entries from 10:02 to 12:00 excluding the lunch period when the Averages are going to merge and give you false signals. For this year. long and short. 100 shares per trade. no brokerage or slippage.
Well, guys, the point I always made is that it is NOT a purely mechanical system. It creates a framework, you still have to use your brain to select the right stocks to enter, as well as when to exit. That is why mechanical backtesting will never work.
The fact is, that, especially in this market, day trading according to this method still gives me my best income.