Tuesday, March 30, 2010

Position sizing

One of the most important subjects in trading is position sizing, yet it is very difficult to find more information on it and to have quick and magic formula. A position sizing should tell you how much of a position to take based on your account size or, if you look it from our side of trading ALSI contract, how much one must have per contract traded.

My opinion is that for traders who made it to the next level of trading and who are constantly making money, the most dangerous is to trade with too big positions, as that could cause them to go belly up.

On the other hand, risking too little will prevent growth at speed desired. One should try to find very thin line between destroying his account when system gets into drawdown on one side and trading large enough in order for progress to materialize and for his account to grow.

So, what is the solution for this? How can we determine how much should we trade?

Some books say that one should not risk more than 1% of the equity on any trade. This rule was made for share trading/investing with no leverage added. My opinion is that this rule must be modified to 5% per trade for valid, sound, well-tested systems. So, what does that mean?

Let’s say we trade ALSI contract and average loser is 300 points. If you don’t want to risk more than 5% per trade, then you need to have at least 6000 points in account. So, if you trade one contract and average loser is R3.000, then I would recommend having at least R60.000 in your account. This will give you approximate chance to comfortably survive even 2500 points drawdown and have money to continue trading.

Another approach to measure position sizing is based on Kelly formula which was used by famous Turtle traders:

The Kelly's formula: Kelly % = W - (1-W)/R


Kelly % = percentage of capital to be put into a single trade.

W = Historical winning percentage of a trading system.

R = Historical Average Win/Loss ratio.

Third approach is Optimal F formula which is small modification of Kelly’s work. I’ve found out about this formula from http://www.verticalsolutions.com/

Optimal F is maximum number of contracts one can trade for given account size. One should never trade more than optimal F suggests, because risk of destroying account is growing, but one can and probably should trade fewer contracts than optimal F suggests.

Optimal F formula: F = (Largest losing Trade)/((((1+win/loss ratio)*% wins)-1)/win/loss ratio)

One of the drawbacks of the Optimal F formula is that Largest losing trade is not fixed number and it could be changed in any time. Unfortunately, in trading there are few if any determined values, so one must live with it if one wants to use this formula. For most of our systems we are using this formula and for system published on twitter formula says that one must have R25.000 on top of the margin to trade one contract or R42.000 per contract with margin of R17.000.

As a side note for folks who like Optimal F – If you have tested your system over long period of time, but you have discovered later that your system was tested with data when Average True Range (or any other adequate volatility measure) was, let’s say, 2-4 times smaller than present volatility level, I would strongly suggest for exposure to be cut 2-4 times, as big loser (which will change Optimal F value) is just around the corner.

Also, from my experience, having in mind an assumption that your systems are robust, well tested and executed, I would suggest trading with 30–100% extra (in account per contract traded) over Optimal F value, subject to black swan events; robustness of your system and faith you have in your own research, and market conditions.

Last approach that I know of is Van Tharp’s approach from one of his books where he says that in account one must always have double the amount of known maximum drawdown (for the system) on top of the margin. For our average systems drawdown is between 2000 and 2500 points. So, that comes again to 5000 points plus margin for each contract traded.

Obviously, there is no magic formula that will tell you how much one must risk in every trade. There are four different calculations in this post and decision must be made according to risk profile of each trader. One must be very careful with how he handles money, because when money is gone the game is over, but again, one should not be too conservative because then there is no need to trade highly geared instruments like futures.

This subject is never ending story and the more we trade the more we will find what our optimal position sizing is. In some instances, when markets perform better or volatility is low, one can increase size or do the opposite when market conditions do not favor one’s systems. But again, this can and will come only with experience and for the start one should just stick to what is explained above.

Trade with trend!

Monday, March 29, 2010

Top40 - end of the day recap for week ending on March 28th

Last week on Top40 was one to forget. Top40 lost 4 (FOUR!) points for a week with volatility dropping to extremely low levels (please check previous post). Top40 traded only during 4 days (Monday was public holiday) and it is important to mark that Thursday and Friday were inside days (price never broke above or below Wednesday’s daily range).

Indicators are still at high levels, with stochastic dropping just a bit and currently is around 70.

There is still potential for double top and Top40 has not closed above 25922 level.

Next two weeks will bring us four-day trading weeks and I’m not sure if volatility is going to increase soon. My opinion is that we are to wait for full 5-day working week to see some volatility.

Moving averages, which help us to determine long, medium and short-term trends:

  • long-term (200sma) - UP
  • intermediate (89sma) - UP
  • short-term (10ema) - UP
Trade with trend!


Wednesday, March 24, 2010

Average True Range, again

I wanted to take one more look at Average True Range for Top40.

Presented chart (end of the day) starts somewhere in the middle of 2003. We were going through nice bull run, then through massive and fast bear run, with new bull run starting in March 2009.

Top part of the chart represents 5-day Average True Range, expressed in percentage terms. This is done to compare apples to apples, because it is not the same when ATR is at 1% and Top40 is at 10.000 points or ATR is at 1% and Top40 is at 30.000 points.

What we see on this chart is that for the most of the time, over last six to seven years, 5-day ATR (%) spent time in the range between 1% and 2.5% (two horizontal black lines). There were some massive peaks when market sold off and especially during bear market in 2008 and 2009.

Currently 5-day ATR (%) is around bottom of that range and during last few months it was there only in late December and early January, when volumes were low and traders were away from their screens.

Unfortunately, our swing systems are not producing good results with volatility being as low as it is now. However, if there is something to look forward, ATR should go up from here and if that is to happen our system should start to perform better. I am not sure for how long this period is going to last, but be ready to pick up nice points and be emotionally ready for fast swings.

Trade with trend!

Monday, March 22, 2010

Top40 - end of the day recap for week ending on March 21st

Top40 added 194 points last week or some 0.76%. Week started and ended with red days and we had three green days in the middle. Thursday was futures close out and that brought us nice up day, with new year-to-date high - only 2 points above the previous one, but that counts too.
Volatility is still on the low side. Average ATR for last 5 days is only around 1.2%, so in terms of volatility we are re-visiting conditions of end of December or very early January, when most of the traders were not participating in market.
Indicators are in overbought position and there are no visible divergences.
Stochastic is very close to 100 and on Thursday’s close it recorded 98.262. RSI smacked against 70 line, as that line very seldom gets breached. MACD is solidly above 0 level.
There is potential double top on the chart, but it is still far away from confirmation. Level around 26000 is very strong support for now and final battle between bulls and bears is going to be fought around that level. Many long term investors consider 25922 on end of the day bases, as level where bears will surrender their defence. If Top40 closes above 25922 it will signal to them that the end of secular bear market is in place and beginning of new secular bull market is underway. By itself that means that many stop losses will be triggered around that level. We should just wait and see what is going to happen.
Last week brought strong Rand which broke below 10.00 for EURO and by many global strategists that strength will continue.
Next week will bring us two-day Reserve Bank’s Monetary Policy Committee meeting that will result with interest rate decision on Thursday 25th of March around 15:00.
Moving averages, which help us to determine long, medium and short-term trends: 
  • long-term (200sma) - UP
  • intermediate (89sma) - UP  
  • short-term (10ema) - UP

Trade with trend!


Friday, March 19, 2010

EURZAR and USDZAR daily charts


Today we are going to check two charts: USDZAR and EURZAR, both daily time frame.

Yesterday EURZAR has broken bellow psychological level of 10.00 and is currently trading around 9.95, on very long support that goes back all the way to 2007. In addition, EURZAR is locked in clearly defined channel that started in November 2009 and is still intact. Top of that channel is currently around 10.19.

If EURZAR manages to break below 9.95 next supports is in the region of 9.33 where triple bottom was formed in April, July and October 2007, which eventually caused EURZAR to go all the way up to 15.26.

Second chart is daily USDZAR chart and we can see similar picture there. From early August 2009 USDZAR is locked in trading range from 7.22 to 7.85. Currently is trading around bottom of that specific range. If it manages to break down we can see 6.60 where target for break of the range is or even 6.40 where support made in November 2007 is.

So, why ZAR is getting stronger?

I’m not forex or global picture expert, so I would use comment made by Jimmm Slater, one of the regular posters on page88.co.za forum, who works for “MEGABANK” (as he likes to say) and whom I trust big time when it comes to global pictures and macro views.

He said: “FX flows on macro not the micro. The ZAR is just part of the risk on risk off trade and risk on looks like its back driving the market at the moment. ZAR is a deep, liquid market that big funds can play around in. We (they) don’t care about what is actually happening in the country, we just use the ZAR as a way of playing the global macro trend. Facts, even economic facts, have very little to do with the performance of risk assets.”

What will happen to ZAR in near future? Me view is that it will continue to get stronger, at least until appetite for risk is there. Maybe this ZAR strength will influence Top40 shares, so THEY (from Jimmm’s comment above) are going to increase purchase of shares and bonds here in SA and that will push our markets up.

Trade with trend!

Tuesday, March 9, 2010

Top40 - end of the day recap for week ending on March 7th

 It is time to update Top40 daily chart in order for us to see what the odds are for this new week.

Last week we had 5 green days and all together seven in the row so far. I really don’t have to point out who is winning the battle currently. Top40 is very close to year to date high, which is around 26000 and the odds are very good that we are going to test that level. At the same time, 26000 is 61.8% Fibonacci level for bear market from May to November 2008.

Stochastic is very high, but it can stay high in the prolonged trends. RSI is high as well, just short of 70. MACD is above 0 line and is pointing to more upside. There are no major negative divergences and for these to happen Top40 should reach new high.

If I remember correctly, Top40 made 9 green days in a row few years back, and as well 9 red days in February 2009. So, there is a possibility for more upside to come.

Moving averages, which help us to determine long, medium and short-term trends:

  • long-term (200sma) - UP 
  • intermediate (89sma) - UP 
  • short-term (10ema) - UP 

Bulls are controlling Top40 so far. Don’t try to short this market based on end of the day chart. 

Trade with trend!

Wednesday, March 3, 2010

When to be afraid of next trade?

Last few weeks all trend following systems (which are on my screen) on shorter time frames of 15 minutes and 30 minutes behaved dismally, losing quite a big chunk of points. Current drawdown for most of followed systems is between 1500 and 2500 points. That correlates to about 15-40% of the equity for the period that is recorded. Number of losing trades outstripped winners by big margin when compared to previous longer performance periods. Simply, we had streaks of five or six losing trades in the row.

Of course, this is not what happens often, but it is something that will happen for sure once a year or maybe once every two years. Times like these are epitome of plans to stop using systems; tweak them and to look for holy grail of trading one more time. Many novice traders, who started with trading recently will place question mark and will start thinking about trading: How this is possible? Should I stop trading? Should I skip trade or two?

Many will stop with trading once the pain threshold exceeds level one is comfortable with. They will, in most cases, stop trading just before big winner comes in to wipe good chunk of painful losers. Many will be very afraid to place next order, knowing that it will/can be a loser, as they can’t take it any more.

Should you be afraid of placing next order?

Answer to this is definite YES. You should be afraid of placing new order every time when you are OVEREXPOSED to the market. This is the only case that you should be afraid of placing trades. Simply, if you have profitable system and you have strong belief, you should not be afraid to place new orders.

One should always look at this game as at game of averages. If your average trade for your system is, as an example, 50 points, then you should put yourself in state of mind that whenever you place the order you actually make 50 points. That way you will remove yourself from big strings of losers and for that matter you will mentally cut strings of winners, as these can cause overconfidence.

However, when you are overexposed it is different ball game. One should never be overexposed and should always be on the safe side. First rule in trading is to survive; second rule is to make money by consistently adapting and adjusting to new market conditions. So, in order to adhere to rule number one, one should never be in position to lose his own capital. Overexposure (due to overconfidence or other reasons) is probably the biggest killer of all traders - freshmen, seasoned, experienced, etc.

There are many formulas on the Net and in books where proper sizing for one’s trading account is to be found, but more about that next time.

Trade with trend!

Monday, March 1, 2010

Top40 - end of the day recap for week ending on February 26th

It is time to update Top40 daily chart in order to see what the odds are for this new week.

Last week, after nice up day on Monday, market drifted lower. Close on Friday was at 23994, just a tad bellow 24000 mark. Market tested again 23.8% Fibonacci level for the current up move from March 2009 to early January 2010.

Stochastic (was overbought last week) is around 50 now. RSI dropped a bit as well and it is around 50 too. MACD that broke bellow 0 (with correction which started in January 2010) is still bellow 0, but above signal line.

Moving averages, which help us to determine long, medium and short-term trends:
  • long-term (200sma) - UP 
  • intermediate (89sma) - FLAT
  • short-term (10ema) - FLAT 
Market is in no man’s land with no determination to up or down side for short and medium-term.

Trade with trend!