Breakout system broken?

[Superboot]

Hi all,

I have been trading a breakout system the last few years reasonably successfully. However, the last 4 months have been out of character and it is not performing as I expect as my system has previously performed well when there has been a stong run in the all ords which we have seen over this period. I am trying to determine whether this is a small blip or whether I have a more serious issue to contend with.

To cut a long story short - I trade speculative stocks that have a large volume spike, a new high plus some other criteria.

Question 1 - has anyone else experienced some challenging times with speculative stocks over the last 4 months? If so, is there some logical reason why?

My analysis to help understand what is going on:

Failure Period (Dec 2012 to Mar 2013):
Initial capital 100000
Ending capital 86569.54
Net Profit -13430.46
Net Profit % -13.43%
Exposure % 47.09%
Net Risk Adjusted Return % -28.52%
Annual Return % -39.43%
Risk Adjusted Return % -83.72%

Winners 7 (21.88 %)
Avg. Profit % 20.82%
Losers 25 (78.13 %)
Avg. Loss % -14.10%


Extended Success Period (Jan 2000 to Nov 2012):
Initial capital 100000
Ending capital 6419739.18
Net Profit 6319739.18
Net Profit % 6319.74%
Exposure % 25.41%
Net Risk Adjusted Return % 24872.51%
Annual Return % 37.72%
Risk Adjusted Return % 148.47%

Winners 824 (45.20 %)
Avg. Profit % 24.08%
Losers 999 (54.80 %)
Avg. Loss % -13.00%

The key factor I am attributing to the failure period is the low % win to loss ratio ie. only 22% winners vs. expect 45%.

Some further analysis I conducted on the actual stocks to understand why there is a low win rate showed me the following -
Failure period: was comprised of 55% mining and energy stocks
Sample Success period: had 75% mining and energy stocks

Question 2 - I am thinking my breakout system is tailored for the mining and enegy stocks. A slowdown in these sectors may be the reason for the poor performance. Would others agree? Other theories?

Question 3 - Is there an index that tracks small speculative mining/energy stocks?

Apologies for the lengthy post.
All comments welcome.

 

[tech/a]

Yes a high correlation to mining Stocks will not help.

Much lower win rate also kills it.
If its outside your Montecarlo blueprint then
you should at least suspend trading and re test your method including this period.
The very small sample size isnt helping either.
in 12 mths time that (The sudden down turn) maybe in consequential.

Youve cherry picked a choppy period Im sure a similar period within your initial test
dates would return similar
Try and find one or three and test 3 mths as you have here.
Whats the longest string of losses in the initial test period?
If more or similar then its happened before!

You can always look at 1/2 position sizing until it turns
OR You could trade the equity curve when it turns up again
OR Look at your trading universe.

 

[pixel]

An article in the Easter Weekend West made a similar observation, commenting that many small-cap speccies had fallen out of favour and consequently out of bed over the past 6 months. A few charts underpinned the point, showing also a couple that bucked the trend - LKO being one of them.
My personal suspicion, based on frequent observation, is that 'bots appear to have moved down the chain, setting up "promising" patterns like breakout alerts, which in all likelihood will attract retail online buying. Any trading of small numbers - e.g. 7 shares at 4.6c every few minutes - is a red flag that makes me ultra-cautious. If they alternate with larger crossings at the next higher or lower level, suspicion usually turns to certainty. It's only a matter of time that those accumulated small numbers are used to turn the direction - as soon as enough credulous retailers have taken the bait and appear ready to get fleeced either by a sudden surge of supply or a buying spree.

Example:

 

[kid hustlr]

Great topic for a discussion. Very keen to hear what tech/a et.al have to say.

 

[skc]

My personal suspicion, based on frequent observation, is that 'bots appear to have moved down the chain, setting up "promising" patterns like breakout alerts, which in all likelihood will attract retail online buying.

Or may be falling commodity prices, fear of China slowdown, the big miners cutting back on exploration activities etc have more to do with it. Junior species rely on hope, hype, bluesky and optimism in order to attract speculative money and raise capital. A negative sentiment on the industry based on some real risk factors is enough to nip most breakouts in the bud.

Question 1 - has anyone else experienced some challenging times with speculative stocks over the last 4 months? If so, is there some logical reason why?

There were plenty to trade (on 1-3 day timeframe) before end of Feb but things changed noticably after that. Test your system with tighter stops to see if you notice any difference in your P&L.
The logical reason imo is exactly as per the hypothesis of your Question 2.

Question 2 - I am thinking my breakout system is tailored for the mining and enegy stocks. A slowdown in these sectors may be the reason for the poor performance. Would others agree? Other theories?

While not necessarily "tailored" it certainly seems to pick out plenty of mining stocks - as those are the ones susceptible to high volume breakouts.

Question 3 - Is there an index that tracks small speculative mining/energy stocks?

Not really but I'd create my own index if it is a concern. Simply take a random selection (say 60-80) of the stocks traded in the back test and do a price-weighted index. Assuming you are talking about nano- to micro- juniors (sub $100m) any capitalisation based index would be a waste of time.

Question 4 - What are some of your strategies to deal with the loss of profitability?

 

[Superboot]

Hi all,

Thanks for your comments so far

Some additional information - I backtested in 4 months blocks from 2000 to check for other 'poor' periods. I couldn't find a period as bad, however there were some that were close. The couple of examples I looked at were choppy markets and the trend was changing from bullish to bearish, in which case I expect poor performance. If these cases were bullish periods I would feel more comfortable that is it part of the system behaviour....

Period Net % Profit CAR # Trades # of winners % of Winners # of losers % of Losers

1/01/2002-1/05/2002 -7.12 -20.12 38 14 36.84 24 63.16
1/01/2005-1/05/2005 -4.15 -12.48 44 16 36.36 28 63.64
1/05/2005-1/09/2005 -4.35 -12.45 35 14 40 21 60
1/01/2001-1/05/2001 -4.05 -11.82 10 2 20 8 80
1/05/2004-1/09/2004 -3.24 -9.44 15 5 33.33 10 66.67

I haven't decided upon a strategy forward (hence my post), however as Tech suggested reducing my stake size is up there as my first option - this was my approach when I first started. Once the returns started coming in I increased my stake.

Some other things that I am considering which have been highlighted by both Tech and skc are:
1. tighter stops (possibly at the expense of return)
2. moving away from speculative stocks (possibly at the expense of return)
3. sitting on the sideline until the system moves back into profitablity
4. other suggestions?

Cheers

 

[Superboot]

Hi folks,

A quick update on how things have progressed.

I have changed my filter from the All Ords to a custom 'speculative' one. I randomly selected 40 or so stocks based upon an average daily traded value < $500k. The results are promising -

(4 months)
Failure period (with All Ords) = -15%
Failure period (with custom index) = -8%
This is more in line with my historical 'down' periods

(16 months)
1/1/2012 to now (with All Ords) = 0%
1/1/2012 to now (with custom index) = 27%
The other measures (e.g. MaxDD, W/L ratio ) are also in line with my system profile

The custom index was not strong when I back tested 4 yrs from 2000 - 2003, however I am thinking this is part of a system's 'evolution'.

Any thoughts on how often I should update the stocks in my custom index?