Olikea -
You are still testing with Close < SMA(10). What happens when you take this out?
First off, I am using http://www.candlesticker.com as a guide, and looking at the “high reliability” bullish patterns. They list a number of medium/low reliability patterns, but I can’t be bothered to test them, unless someone (ahem Stitts?) knows of a particular pattern that is likely to work!
The only one I haven’t so far tested is the “Concealing Baby Swallow”, (oh and the poisiness fish one 
The pattern is defined here:
http://www.candlesticker.com/Cs81.asp
I think the rules for Concealing Baby swallow should be as follows:
close(3) < open(3) black
hi(3) < 1.001open(3) no upper shadow
low(3) > 0.999close(3) no lower shadow
close(2) < open(2) black
hi(2) < 1.001open(2) no upper shadow
low(2) > 0.999close(2) no lower shadow
close(2) < close(3) closed down compared to yesterday
open(1) < close(2) gapped down
close(1) < open(1) black
hi(1) > close(2) trades into body of prior day
close(0) < open(0) black
hi(0) < 1.001open(0) no upper shadow
low(0) > 0.999close(0) no lower shadow
close(0) < low(1) engulfs prior day
open(0) > hi(1) engulfs prior day
I am not inffallible, please check for deliberate mistakes!
Anyway, the long and the short of that is that it doesn’t work:
“No stocks were bought in first 26 weeks. Your Buy rules may be too restrictive”
How amusing!
In reply to Stitts comment about a recent downtrend, well according to the defnitiions, these are bullish reversals, “characterised by a downtredning market”. But I have re-run all of the sims without the "price < SMA(10) " rule. Below I show the sims with a summary of the results in the form
“annualised return with rule” → “annualised return wihout”
“sim without rule”
Here are the results:
Thee Inside Up
31% → 29%
http://www.portfolio123.com/port_summary.jsp?portid=263520
Three White Soldiers
23% → 7%
http://www.portfolio123.com/port_summary.jsp?portid=263522
Three outside up
18% → 15%
http://www.portfolio123.com/port_summary.jsp?portid=263523
Morning star
12% → 3%
http://www.portfolio123.com/port_summary.jsp?portid=263525
Kicking - N/A
Abandoned Baby
-4% → -3%!
http://www.portfolio123.com/port_summary.jsp?portid=263526
Morning doji star
-6% → 0%
http://www.portfolio123.com/port_summary.jsp?portid=263527
So as you can see, in every case except for the negative ones, the loss of the “price < sma(10)” rule diminishes performance. This got me thinking, what is the effect of the rule by itself? Answer:
http://www.portfolio123.com/port_summary.jsp?portid=263530
Interestingly, its shows a very slightly positive graph, 10% annualised, implying short term weakness is anti-trending. But its not a strong effect.
Right, you ppl owe me a beer for doing this!
Right, you ppl owe me a beer for doing this!
Next time I make it to the U.K. I’ll buy you a pint! You can introduce me to the fine beers.
Steve
Ok, so I have been thinking hard about Stitt’s comments. I decided to use a different way to look for a downtrend,
-3 consecutive down days (short term down trend)
-black candlestick
-price above 50 SMA (overall trend is up)
-priceH/priceL > 2 (high volatility)
-market cap limited to less than $1bn
The sell rule is still nodays>3. I have had a bit of an experiment with other sell rules but wihout sucess so far.
The results are OK, 52% annualised:
http://www.portfolio123.com/port_summary.jsp?portid=263770
BUT with a profit per trade less than 1% average, its still a long way from being tradeable. 
I don’t see how without something radical such a system could work… Do we have any succesful swing traders willing to give a hint or two?