Determining trade size based on volatility

M

mpat89

Active member
Messages
134
Likes
1
Hi,
I am developing a trading system designed to catch trends in the FTSE 100 stocks.
The system will enter even in ranging periods but these are likely to give losses averaging less than 5% of share price. The idea (or hope) is that all the big moves will be caught and these can range from 1-2% upwards..in some trades I've seen over 30% being taken and in some stocks 10% wins appear common (recently anyway).

My problem is this...
I will be entering lots of trades due to the nature of this system. Right now I can see there would be over 30 open positions. I am wondering what the best way to handle money management would be on this and any advice in this respect would be greatly appreciated.
Should I set a risk per trade and use a pre-determined stop loss or should I divide my capital by 30 (with spreadbetting this means each trade's position size will start at a third of capital (meaning with maximum trades open I would be leveraged 10x).

My confusion comes because it is clear that some stocks trend in smaller percentages than others, and I want to somehow standardise this between stocks (thinking about beta values or something like that but unsure how I'd apply this). Just wondering if anyone has any input or opinions!
 
Just to give an idea..rio tinto..
long (stop at 9.2% )1st dec closed 10th dec -3.39%
short (stop at 8.6%) 10th dec closed 16th dec -3.3%
long (stop at 8.73%) 16th dec closed 12th jan 11.52%
short (stop at 6.29%) 12th jan closed 2nd feb 7.01%
long (stop at 7.93%) 2nd feb closed 5th feb -7.93%
net profit around 2-3% when compounded
anyway u can see from the dates that I am practically always taking a position. my testing has shown me I will make lots of small losses like the 3%'s above, now and again I will be stopped out if the market turns against me too quickly but generally stops are not hit and losses are closed for much smaller % losses than where the stop is placed.
i feel at any one point there are always enough 'moving' stocks to outweigh the losses generated by the active nature of the system but drawdown will occur.
I am just wondering whether anyone has opinions on whether stops should be fixed % of my account or if I would do as mentioned already and open positions each worth 33% of my account (if stop loss is 10% then I am risking 3.3% of my account per trade. This is high but stops are wide and not hit often.
 
mpat89 said:
Hi,
I am developing a trading system designed to catch trends in the FTSE 100 stocks.
The system will enter even in ranging periods but these are likely to give losses averaging less than 5% of share price. The idea (or hope) is that all the big moves will be caught and these can range from 1-2% upwards..in some trades I've seen over 30% being taken and in some stocks 10% wins appear common (recently anyway).

My problem is this...
I will be entering lots of trades due to the nature of this system. Right now I can see there would be over 30 open positions. I am wondering what the best way to handle money management would be on this and any advice in this respect would be greatly appreciated.
Should I set a risk per trade and use a pre-determined stop loss or should I divide my capital by 30 (with spreadbetting this means each trade's position size will start at a third of capital (meaning with maximum trades open I would be leveraged 10x).

My confusion comes because it is clear that some stocks trend in smaller percentages than others, and I want to somehow standardise this between stocks (thinking about beta values or something like that but unsure how I'd apply this). Just wondering if anyone has any input or opinions!
Click to expand...
I won't comment on your system, but merely on your question about stops. It is a good idea to standardise risk across your portfolio. Before talking about stops you should consider doing this when you open a position by not basing postion size on a fixed percentage of overall capital. $100 of microsoft stock does not have the same risk as $100 of BNI stock for example. In consequence it would be more appropriate to fix the risk across your portfolio rather than the position size. Thus if stock A is twice as risky as stock B then stock A should have half the position size as stock B. So the question is how do you measure risk. One easy approximation is to look at how far a stock moves on average in any period and this is measured by its ATR. So you could divide up your capital in the ratio of the ATRs of each stock. The same kind of thinking can be applied to the stops. If, on average, you are likely to hold open each trade for say 15 mins, then look at the movement in cents over that period on average. Some stocks will naturally move more over a defined period of time that others. Again ATR can be used as an approximation. So determine the ATR over the period in question. A commonly used formula is 1-2 ATRs measured for the holding period.

In suitable software such as Tradestation all these calculations can be set up as programmable values that will automatically calculate position size and stop size and to warn if the total portfolio size is nearing its limit.

By the way I have just realised you are looking at FTSE100 stocks, so just convert all my dollar examples to pounds and pence !!

Charlton
 
Thank you for taking the time to share your thoughts. ATR seems like an excellent way to standardise risk and also help get out of a losing trade early as stops will generally be more narrow (with obvious trade-off). Now I am facing a problem. The system will profit from the volatility of stocks but by using the ATR I am reducing risk for these volatile stocks whilst possibly increasing risk for the less volatile stocks.

A question I am asking myself as a result is what is the comparison of trend strength, size and quality between the less volatile and more volatile stocks. I will spend time trying to figure out the answer to this one.
Another thought is whether I should spend more time studying charts of a much wider range of stocks and handpicking the ones which I will follow. For now I have decided to trade the DJIA 30 stocks as this gives me only 30 stocks to worry about as opposed to 100.
 
You have to be a bit careful with using ATR. When the ATR is low, you end up with smaller stops, and larger position sizes and that seams reasonable as it normalises risk, and thats good. The problem is that some types of method tend not to do too well under conditions of low volatility, so you can inadvertently end up with your largest positions, at precisely the time that your methodology is performing at its worst.

It might not apply to you, but it is definately something to watch out for.
 
I don't see this as too much of a problem as if risk is normalised then the position size is not important unless there is a nasty gap in the price. Regardless of position size I should only end up losing a fixed percentage of my account assuming my stop loss gets filled at the correct price.
 
I have not yet decided but I am going to consider using the risk %age of account method rather than position size being %age of my account. I will use whichever stop is closer, ATR or my own stop.
 
mpat89 said:
I don't see this as too much of a problem as if risk is normalised then the position size is not important unless there is a nasty gap in the price. Regardless of position size I should only end up losing a fixed percentage of my account assuming my stop loss gets filled at the correct price.
Click to expand...

Agreed 100%, if your using stops, and theyre honoured you'll be fine. However, it can be a problem if your trading methods where the max risk isnt identifed in advance, for example stop and reverse type systems, or systems where trades are closed on TA type signals etc.
 
You might find Implied Volatility more useful than ATR, simply because it is looking ahead rather than behind.

At the most basic level I have always used an old British Army saying about what life is like in the Army.
i.e. "Hurry up and Wait".
In other words, if you get a significant fast move (hurry up) then it is almost always followed by a consolidation (wait). Each imply that the other is coming next, in all timeframes
Also markets only trend (hurry up) for about 15-20% of the time.

Glenn
 
You might find Implied Volatility more useful than ATR, simply because it is looking ahead rather than behind.
Click to expand...

Glenn, do you know how this is calculated ?


Paul
 
There are all sorts of methodologies you can use to do this. In an institutional context, the most commonly known one is VAR (or one of its many flavors). VAR is not a very difficult concept, neither is it difficult to implement. The main idea with VAR is that to know the risk of a portfolio it's not enough to know the risk (volatilities) of each of the individual trades. You also need the correlations. Most commonly people use historical data to obtain both the volatilities and the correlations (you don't want to use mkt implied vols, since these are often systematically wrong). To address the issue with historical risk being inherently backward-looking, people also rely on secenario analysis/stress tests.
 
You must log in or register to reply here.

Similar threads

P
position sizing and determining leverage with kelly formula
Replies
1
Views
3K
HowardCohodas
HowardCohodas
L
How is options implied volatility for a stock determined?
Replies
5
Views
5K
Windlesham1
Windlesham1
R
Volatility Based Intraday Trading
Replies
0
Views
4K
radha55
R
G
trading size
Replies
9
Views
3K
alexander
A
L
Who determines the spread?
Replies
9
Views
3K
CityTrader
CityTrader
Top