Trade System Design - Reducing Drawdown

[chesl73]

Hi,

I've been coding a few ideas of mine into a trading system using Amibroker for EOD data.

I've a few questions I'm hoping a few of you out there might be able over general advice.

1/
I'm not that creative so I find the whole process quite difficult but I find that I can come up with some reasonable entries but the system drawdowns are always awful (~50% or more). If I then just apply a trailing stop (say 3*ATR(14)) to take me out of the trades I then find that my profit drops dramatically and the drawdown still isn't that low.
So I was wondering how others who have been through a similar process generally go about trying to build a successful system? Should I be working more on my Sell criteria (which I find very difficult) or maybe the high drawdowns are a consequence of the type of system I'm developing?

2/
If you design a trend following system, would you still use it in a non trending market (say such as now for the ASX) or do you (personally - or your system) recognise this and step aside?

3/
One final thing, if my plan when trading is to look at the data during the evening and then issue the buys the following day, should I always use the Amibroker setting of: SetTradeDelays( 1, 1, 1, 1);
ie, does this then replicate my plan?
If I leave this as: SetTradeDelays( 0, 0, 0, 0);
then it seems to give me far far better results but then is this because I'm putting on the trade on the day before I've actually received the signal to buy?

Thanks

 

[Tradesurfer]

Do you find that you are exiting trades too early (taking profits too quick?)

I do like using the ATR to develop one of two exits. The ATR one is my initial line in the sand. Then my second one is when the trend indicator goes negative.

The other thing to think about is using the right position size. Since you are using ATR, do you also size your positions equally based on that volatility measurement?

 

[$20shoes]

Its not uncommon for longer term systems to suffer larger drawdowns. With shorter term systems the required win rate tends to make for a smoother equity curve.

When did you suffer the 50% drawdown? Was it during the GFC which in theory would have been tracking the indice falls at about the same percent. Is your drawdown worse than what your total universe experiences at the time ( XAO, XJO, XKO etc)? If it is a longer term system and you know it won't perform well in ranging or bear markets you could look to switch it off..then again you shouldn't be getting a hell of a lot of long entries for a long term system anyways.

Is margin involved here? Portfolio heat? Max positions? Fixed fractional risk?

 

[chesl73]

Looking into it further I think it's my exits (or lack of).
ie, this is probably quite common but I spend my time trying to get a good entry and then seem to forget about or not put as much effort into the exit other than just slapping on a trailing stop.
I was reading something yesterday that said something like:
1. Use stops purely for money management to protect your capital and minimise losses
2. Use a proper exit criteria to remove yourself from the trade when the trade conditions no longer apply.

I think this is a good way to look at it.

 

[OzAsh]

Look at the watchlist you are running the backtest on. I generally break my watchlists down into ATR levels and start from there. Also make sure you filter with volume so the manipulated garbage gets weeded out.
Hope that helps ...
Ash