Account draw down/growth cycles

[jon1971]

Are these cycles caused by changes in underlying Market dynamics (e.g. trend to consolidation to trend etc.) or by changes in the traders psychological perspective or a combination of the two?

What are your thoughts and ideas for combating any potential negative effects on account equity that these cycles may inflict?

I'll start the ball rolling with: reduce lot size when you think you may be entering a draw-down phase...

 

[tomorton]

Are these cycles caused by changes in underlying Market dynamics (e.g. trend to consolidation to trend etc.) or by changes in the traders psychological perspective or a combination of the two?

What are your thoughts and ideas for combating any potential negative effects on account equity that these cycles may inflict?

I'll start the ball rolling with: reduce lot size when you think you may be entering a draw-down phase...

Challenge accepted.
I would prefer it if it was the market's fluctuations, rather than my own. The market can be gauged objectively and its cycles quantified (even if after the event): this allows for design of a trading system that automatically 'expects' draw-downs but makes profit over the long term.

The implication for me then is, stick to the plan, size positions according to the strategy's objective inputs only (i.e. distance to stop-loss) - don't reduce position size on additional subjective criteria.

A further analysis of the above is that I am aspiring to be a great system designer and don't particularly rate / desire the skills traditionally seen in traders. I wannabe a sniper, not a gunslinger.

 

[meanreversion]

Are these cycles caused by changes in underlying Market dynamics (e.g. trend to consolidation to trend etc.) or by changes in the traders psychological perspective or a combination of the two?

What are your thoughts and ideas for combating any potential negative effects on account equity that these cycles may inflict?

I'll start the ball rolling with: reduce lot size when you think you may be entering a draw-down phase...

When I've backtested reducing stake size (and I mean in percentage terms, not in absolute amounts) during drawdown, the results are not good - it simply takes longer to attain a new peak in equity.

This is the first ever thread I've seen on this website with "drawdown" in the title (other than one I posted). The preoccupation seems to be with "500 pips a day", "20 pct a week", "100 pct success rate" etc etc - drawdown is ignored, but this is what everyone encounters and how you deal with it is perhaps the most important aspect of trading, in my opinion.

I'm running a trend model and May has been fantastic so far, truly awesome. HOWEVER I keep thinking back to the Seykota quote "the trend is your friend until the bend at the end". I know that "bend at the end" is going to happen at some point, I just don't know when. I would like to think I'm ready for it, but drawdown is never pleasant.

 

[jon1971]

So that's two of you in favour of maintaining position size...

I thought maybe reduce risk until you've figured out what the problem may be, but "stick to the plan" is certainly excellent advice, it also adheres to a disciplined approach and of course there's that old chestnut of once you reduce your lot size that's invariably when you get a good entry in a nice trend, hmmm.....

 

[jon1971]

When I've backtested reducing stake size (and I mean in percentage terms, not in absolute amounts) during drawdown, the results are not good - it simply takes longer to attain a new peak in equity.

This is the first ever thread I've seen on this website with "drawdown" in the title (other than one I posted). The preoccupation seems to be with "500 pips a day", "20 pct a week", "100 pct success rate" etc etc - drawdown is ignored, but this is what everyone encounters and how you deal with it is perhaps the most important aspect of trading, in my opinion.

I'm running a trend model and May has been fantastic so far, truly awesome. HOWEVER I keep thinking back to the Seykota quote "the trend is your friend until the bend at the end". I know that "bend at the end" is going to happen at some point, I just don't know when. I would like to think I'm ready for it, but drawdown is never pleasant.

Ha ha, well I try and be realistic.

Re. bend at the end, maybe look at using a tighter trailing stop system as the trend becomes over extended?

 

[montmorencyt2w]

<Devil's advocate mode>But doesn't tightening the stop almost guarantee the probability of it being hit increasing?

That appears to be my experience. I'm not even sure if this is a bad thing, but it sure makes following trends for any length of time more difficult.


Reducing lot size seems like a good defensive move to me, and if survival is everything, well...live to fight another day, etc.


Disclaimer: I'm not sure I really believe in trading "systems" as such, but I strive to keep an open mind.

 

[jon1971]

<Devil's advocate mode>But doesn't tightening the stop almost guarantee the probability of it being hit increasing?

That appears to be my experience. I'm not even sure if this is a bad thing, but it sure makes following trends for any length of time more difficult.


Reducing lot size seems like a good defensive move to me, and if survival is everything, well...live to fight another day, etc.


Disclaimer: I'm not sure I really believe in trading "systems" as such, but I strive to keep an open mind.

I was only suggesting tightening stops if the market is giving indications that the trend is over extended and may possibly be breaking down. I agree with you that you have to leave it room to breathe so as to avoid getting shaken out too easily, like most things in trading and life generally, it's about finding the right balance.

BTW, that's one of my particularly annoying bug bears, I really wish I could find the optimal way of trailing stops (even if one exists).

 

[meanreversion]

The only way to know about trailing stops is to program them and then backtest. I use Amibroker and it's an excellent piece of software. I don't think there's any way around this.. you need to code stuff and then see how changing parameters affects the performance.

For example, I hear a lot of comments along the lines of "move your stop to breakeven as soon as possible". This doesn't hold up in backtesting, it deteriorates the performance markedly in cases.

A popular type of trailing stop is the Chandelier, where the stop (e.g.) for a long position is x * ATR below the highest high for the last 20 days. 'x' is usually set to 3, but you can of course vary it. It's a pretty effective method and can keep you in trends for a long time, but of course you will always give back some equity at the end when the trend reverses.

There is no way to avoid this "bend at the end". The trailing stop is the thing that keeps you in the move for a long time, but it will always be at the expense of handing some money back.

It's impossible to pick turning points, so don't even try. But think about this.. if a FX rate has a large move of 15 pct and you capture 8 pct of it through a trailing stop method, isn't that pretty good anyway? The guy who tells you he got all 15 pct is a liar anyway.

 

[meanreversion]

Furthermore, I've found that taking a conservative route usually worsens performance as well. For example, tightening in stops sounds like the right thing to do, but it can sometimes kill the strategy. Your stop needs to be outside the "noise" (Elder has a clever stop method in one of his books, check it out).

Also, let's say you trade 10 securities, a conservative approach might be to limit yourself to a max of 4 positions, sounds right doesn't it? This will usually worsen performance as well. The best results will be taking signals for all 10 assets, if margin allows it.

 

[tomorton]

I go along with meanreversion's words above. Don't try to second guess the market - trade what you see. Your good system's standard stops will prevent excessive losses even if price reverses 1 minute after you enter, and will very soon give you a new entry signal in the new direction. If you use technicals to generate signals, then a signal is a signal and you should take them all.

 

[meanreversion]

Ah yup, agree with the last bit there many times over. Second guessing my system cost me a lot of money back in Feb, so now I just endeavour to do exactly what it says (in other words, I'm a crap discretionary trader).

 

[jm99]

Are these cycles caused by changes in underlying Market dynamics (e.g. trend to consolidation to trend etc.) or by changes in the traders psychological perspective or a combination of the two?

What are your thoughts and ideas for combating any potential negative effects on account equity that these cycles may inflict?

I'll start the ball rolling with: reduce lot size when you think you may be entering a draw-down phase...

its a combination of the 2 and depends on which sort of trader u are. if a long term systematic trendfollower (clearly some are here) then if will be the change in market conditions will be a bigger factor than individual performance.

At the opposite end of the scale, if your a disc trader, scalping every few mins performance is more important imo.

 

[montmorencyt2w]

Since I am almost completely discretionary, I can't relate directly to the comments on the more mechanical approaches already mentioned, but I will throw in a few thoughts if I may.

Although my discretionary approach was quite successful for a while (months, rather than weeks, but not years ), I did make three or four serious errors of judgement - at least one was about the market itself; at least one was about trade and money-management - which led to some noticeable, not to say painful, dents in the account.

Probably because of feeling bruised, I've been taking a much more defensive approach just lately. Hopefully it will help me survive for longer, but I can see it has already impacted progress in the sense of account growth.

For example, although I have argued against the practice here in the past, I have been experimenting with moving stops to break-even (or break-even + spread, or variations, locking in various amounts of profit). Perhaps because this practice is a bit alien to me, and I am still not very good at it, it does not seem to have been very successful in terms of account growth. Maybe it has had some less obvious benefits, but anyway, this is still early days in my particular experiment. However, so far, invariably the break-even (or better) stop has been hit with a 100% hit rate. So rather than even bothering to move to break-even or better, I might as well just close the trade there and then and settle for whatever the market has given me, which is sometimes nothing, sometimes a relatively small win, but not really what I might have been hoping when [thinking of] moving the stop.

I could probably ramble on about this for longer, but will leave it there for now.

 

[meanreversion]

I've backtested the moving-stop-to-breakeven idea, and it almost always deteriorates performance. In fact I don't think I found a single instance where it improved anything.

 

[tomorton]

I don't usually move stops to break-even either but do occasionally when I have a position large enough to split in half, taking profits on half and running the other half. Howver, the added gain is usually marginal.

The way I see it, TA that signals an entry point must also indicate your exit points with equal quality. So if my TA suggests I can expect a 50pts gain (and that my stop-loss will be no more than 50pts the other direction), there is no TA justification to expect a 100pt gain from this set-up just because I moved my stop-loss. More comfortable and, to me, rational, to take the 50pts as soon as its hit and get on to the next trade.

 

[meanreversion]

From the testing I've done (on various timeframes), it seems that the more uncomfortable a system, the more profitable.

Moving stops to breakeven might make you feel better but it will reduce your return.

If you trade 4 currency pairs and get signals in all 4, take all 4. Restricting yourself to 2 or 3 will also reduce profitability over time, even though it might seem "sensible".

Furthermore, mechanical systems seem to work better on longer timeframes, like dailies. This means keeping positions open overnight more often than not.

Quite a few of these things are anathema to traders, and I'm certainly not saying it's not possible to be successful doing none of the above. But statistically, it doesn't seem to hold up that well.

 

[Adamus]

mean, your backtesting is always on programmed strategies. Who knows what sort of trading system the disc trader has, and therefore who can prove whether it would benefit from moving the stops to b/e or not?

 

[meanreversion]

You're exactly right, that's a good point. I'm not able to test chart patterns (support, resistance, Fibonacci etc) or (of course) discretionary. Thus I don't know if these general ideas apply to other, non-mechanical styles of trading. However, I would be surprised if they didn't apply to at least a small extent.

 

[Adamus]

However, I would be surprised if they didn't apply to at least a small extent.

Actually I'd be surprised too

 

[Vaco]

Interesting thread,

imho, moving stops to b.e is used often in intraday trading to proect capital from drawdowns and limit losing trades to those that go against you from the off. What your really looking for with this style of trading is the one or two trades a day that zoom off in your favour and never look back.

I did some tests a while back on my own results and found that closing he position when i would normally have moved the stop to b.e. would have resulted in larger gains, even though i would have missed out on all my home runs.

On trend following systems how quickly you move your stop all depends on how you got in on the trend in the first place, was it a break out or a pullback entry.

On break out trades drawdowns are suffered when the market ranges and moving the stop or reducing size is not advisable as you risk geting stopped out of the payday trade or been in i with smaller size.

Try testing reducing size after each winning trade and increasing after each losing trade, this may leave you more in sync with the market as a ranging market will break sooner or later and a trending market will stop trending and get stuck in a range again.