Trailing stop loss algorithm

[meanreversion]

I'm trying to develop a systematic approach for a trailing stop loss. Does anyone use the Chandelier method, and if so, what experiences good or bad have you encountered? My system isn't trend following per se, but I am keen to lock in profits fairly quickly whilst leaving enough room for the move to continue and not getting stopped on noise.

 

[amnonco]

but I am keen to lock in profits fairly quickly whilst leaving enough room for the move to continue and not getting stopped on noise.

IMHO this is quite contradictory.
the problem with chandelier stops is the multitude of parameters. you have to backtest the factor, the range etc...

For me the best solution is simply accept the noise, and not "force" locking in profits. if it trends - I win.
but you can use a simpler ATR stop from the last X bars high, or even simpler Y% off your P&L...

good luck

 

[meanreversion]

Since my original post, I've done extensive backtesting on either raising stop to breakeven or using a trailing stop, and my conclusion is... don't. If you enter a trade with a stop of (e.g.) 50 pips and a take profit of 150 pips, then leave these orders, don't tinker. Raising the stop sometimes leads to being taken out of what would have become a profitable trade.

However, it seems the "right" thing to do.... raise the stop to breakeven, use trailing stop etc etc. but my analysis shows that it's not. Maybe that's why trading is so hard; so much of it is counter-intuitive..

 

[meanreversion]

Reminds me of when I learnt to ski, the instructor said "Right, lean your weight forward" when all I wanted to do was lean back and get away from the scary slope.

 

[Charlton]

Since my original post, I've done extensive backtesting on either raising stop to breakeven or using a trailing stop, and my conclusion is... don't. If you enter a trade with a stop of (e.g.) 50 pips and a take profit of 150 pips, then leave these orders, don't tinker. Raising the stop sometimes leads to being taken out of what would have become a profitable trade.

However, it seems the "right" thing to do.... raise the stop to breakeven, use trailing stop etc etc. but my analysis shows that it's not. Maybe that's why trading is so hard; so much of it is counter-intuitive..

Have you also tried backtesting on scaling out - say 50% or 75% - at some level which is defined by technical analysis rather than simply pips, followed by a breakeven stoploss on the remainder, together with a final profit target again based on technical analysis.

In my experience you can often make the same profit out of the remaining 25% left to run as you did out of the original 75% closed out, but with the added advantage that you have already taken profit and the remaining trade is risk-free because of the break-even stoploss.

Charlton

 

[meanreversion]

I'll look into that, but it's probably beyond the limts of my programming ability to backtest that ...

 

[amnonco]

couple of things (once again IMO:
1. don't use pip/tick targets. you never know what the market will give you. your exit strategy should always be price/time based.
2. scaling in and out is good, but only if the entry/exit signal is not price based. i mean - if your exiting at the low of X bars, why scale? but if you exit after X bars, then you should definitely scale out.

you should try ATR breakouts (of a long range EMA) as an exit, they move towards the price in a quiet market and away from price in a volatile market - and therefore, you get what you want - not stopped out on noise.