Fixed stop loss
A friend of mine, he trades Gaps at https://daxgaptrading.wordpress.com/ and the DAX futures overnight session at https://daxovernighttrading.wordpress.com/, pointed out the following idea when trading a fixed stop loss.
This year I suffered several “stopped and back” trades ie: the market hit my stop loss and then reverts to breakeven or even to a profitable night. There are times when an ATR based stop loss formula can be too tight when market volatility increases. Generally speaking, stops hurt long term performance.
Examining this years SPY performance, you will note a stronger equity curve than that produced through the use of a stop when trading ES futures. SPY is refererred to as the ‘cash’ market and does not use post market stop loss orders. This year SPY has benefited from not using stops and has consequently recovered quicker from drawdowns than the ES futures.
But with futures and low margins trading without stop loss orders can be very risky.
The core idea with fixed stop loss orders, is to reach a better compromise between no stops and the need to minimize losses. This allows for the sufficient ‘room’ to let the trade work out but not too large to become a “catastrophic stop loss”.
Looking back as far as futures data are available, from January 1998 to July 2015 (time of backtesting) and considering only nights with ES closing price above its 200 SMA, results follow:
Approximately 1400 trades fit the criteria and only 8 trades suffered a loss greater than 2% (close to open basis) whilst a loss greater than 1% covered about 6% of all the trades.
I included the following plot:
The optimal drawdown (%) level is where the average distance between one dot to the next in the plot starts rising (the tail of the distribution).
From the chart a 2% stop appears to be the best stop loss level. But with a 5% future contract margin this is a huge loss, about 40% of the invested capital. Traders can decide to stop here and to keep more money in cash. Instead of 50% or 60% of capital invested, traders can keep 80% or 90% of capital in cash.
I prefer expoiting leverage more and optmize the stop loss better.
I included the following table:
I stopped at 1% because I don’t want to place the stop loss “in the middle of the jungle”: the dots are too close below 1% and trades stopped out becomes too frequent, at 1% there are 6,3% of trades stopped out. I would like to cut down on the number of stops triggered to at least below 3%.
If I placed the orders at 1.5% from entry I would have been stopped 8+10=18 times since January 1998. About 1.29% of trades stopped out of 1400.
The evidenc is clear as shown on the plot chart how a big gap stands out.
With a 5% futures account margin, we managed to cut the potential loss to 30% of the invested capital. It is still high. With the minimimun money management approach of 50% of capital invested the total loss is equal to 15%.
I can see the last clear gap on the chart plot at 1.3%. That level triggered just once. So I would have been stopped out 8+10+(10-1)=28. That’s equal to only 2% of the trades.
With the futures margin considered above, loss is 26%. With the minimun money management approach it’s down at 13% of the total equity.
Just one last concern, historically the market daily range increases with time over the years as the total index value increases. This is the only data mining potential risk in this analysis.
If that consistently happens, I will come back to 1.5% and rise % of cash to keep the triggers frequency below 3%. It may or may not happen, it depends as usual on the market.
How Nightly Patterns would have performed with this stop loss will follow on another article.
Marco Simioni
NIGHTLY PATTERNS - Overnight Trading 