Wednesday 25 September 2013

Week of 2013 SE 20

IWM rose by 2.2% this week, while my account rose 1.1%.  The loss-floor has risen to -5.6%.

US market news: On Wednesday, Federal Reserve chairman Bernanke announced that he was not ready to “taper” off the $85 billion of newly-printed money going into the stock market each month.  This is effectively an admission that the US economy is not doing well and needs continued life-support to prevent the lifestyles of the rich and famous from collapsing along with everyone else’s.  Markets jumped on the news, but there will probably be a major correction soon.

End of week allocations:
% gain
size
Max loss
Sym  Buy  Fri  Mon  Tue  Wed  Thu  Fri Beg   End
 
ATML AU29 +0.3 10 +0.4 10 +0.6 10 +0.6 10 +0.5 10 +0.4 10 +0.0 +0.4
URTY SE04 +0.9 11 +1.0 11 +1.3 12 +1.6 12 +1.6 12 +1.5 12 -0.4 +0.2
FNSR SE10 -0.1 9 -0.1 9 +0.0 9 +0.1 9 +0.0 9 +0.0 9 -1.1 -1.0
UWM SE13 +0.0 10 +0.0 10 +0.2 10 +0.4 11 +0.4 11 +0.3 11 -0.7 -0.3
IWM +27.0 +27.3 +28.5 +29.8 +29.6 +29.2
me -3.9 -3.7 -2.8 -2.1 -2.4 -2.8
floor -7.0 -6.7 -6.6 -5.9 -5.6 -5.6

Stock-trading robot

Ticker
Symbol
Buy dateBuy priceSell dateSell priceAcct Profit
ModelActualMAModelActual M  A  M  A 
ATML AU 29 09:30 $7.32 SE 04 09:30 SE 20 10:14 $7.32 $7.65 -0.03% +0.39%
URTY SE 04 11:00 $60.82 $60.80 (Not yet)
FNSR SE 10 09:30 $23.53 (Not yet)
UWM SE 13 09:30 $69.80 (Not yet)

ATML: Raised the stop again for Monday, and again for Thursday, and again for Friday — when it stopped out.  I tend to make my manual stops too tight; we shall see if this was a good move.  Anyway, I earned +0.4% on a trade that would have been roughly break-even if the robot system had worked properly.

ATML, YHOO: Both of these have been doing very well recently.  I held ATML due to a mistake, but didn't hold YHOO because the robot sale went through properly.  This shows how nasty the SMA(q) rule is: if price jumps and then hugs the SMA line on the way up, my robot will just sit on the sidelines.  I wish I didn’t need the SMA(q) rule, but otherwise the robot would have suffered extreme losses during some historical periods.

MACD-based trading model

I cannot get the double-TRIX model to produce decent results for SRTY, so I have switched it back to the double-MACD model.

Parameter Tuning
2008 2007
PEAKFREQ(n,size) 140,8 140,8
minfreq 8 8
MACD(qS,sS,nS) 15,66,9 15,71,9
MACD(qQ,sQ,nQ) 4,16,4 4,17,4
MACDchan 12 12
SMA 26 150
stddev 7 5
stddev-EMA 8 12
floor-init 1.027 1.027
floor-ticks 15 10
wait-ticks 2 3
     
Year Results (using 10% tranche)
-3×IWM 2008 2007
2003 -135% -3.8% -2.2%
2004 -51% -4.7% -4.4%
2005 -9% -1.5% -4.7%
2006 -51% -5.9% -4.4%
2007 +6% -5.8% -4.7%
2008 +114% +12.0% +14.7%
2009 -99% +3.2% +8.1%
2010 -71% -0.3% +3.1%
2011 +18% +3.9% +5.2%
2012 -43% -1.5% +0.4%
2013 H1 -45% -2.0% -1.9%

The "2008" tuning is the old one, based on 2008‥2011 with no indicator ramp-up period (because that was all the data I had then).  The new "2007" tuning is based on 2007‥2011 with the last two months of 2006 for ramp-up.

Unfortunately, the new tuning works well only for the years on which it was tuned.  Losses for earlier years are excessive.  This needs more work.

No comments:

Post a Comment