Week of 2014 DE 05

IWM rose by 0.9% this week, while my account fell 0.5%.  The loss-floor has fallen to -12.2%.

End of week allocations: 97% equities, 26% silver, 10% cash

% gain size │ Max loss Sym Buy │ Fri Mon Tue Wed Thu Fri │ Beg End [out] │ 11 20 20 20 20 20 │ URTY ¹ OC28 │ +1.3 15 +0.6 15 +0.6 15 +0.6 15 +0.6 0 │ -0.5 -0.5 ATML NO24 │ +0.5 13 +0.3 14 +0.6 14 +1.1 14 +1.0 14 +1.0 14 │ -1.1 -0.6 YHOO NO26 │ +0.0 13 -0.4 13 -0.3 13 -0.4 13 -0.3 13 -0.2 13 │ -1.1 -1.1 UWM NO28 │ -0.0 14 -0.4 14 -0.1 14 +0.1 14 -0.0 14 +0.2 14 │ -0.8 -0.8 SRTY ¹ DE01 │ +0.0 11 -0.2 10 -0.2 10 -0.2 10 -0.2 0 │ -0.8 -0.8 SRTY ² DE02 │ -0.1 11 -0.2 10 -0.2 10 -0.2 10 │ -0.6 -0.6 AGQ DE04 │ -0.0 14 -0.3 13 │ -0.8 -0.8 URTY ² DE04 │ +0.0 14 +0.3 14 │ -1.5 -1.6 [cash] │ 34 14 3 3 -10 -0 │ IWM │   +1.3   -0.2   +1.0   +1.9   +1.4   +2.2 │ me │ -4.8 -6.5 -6.1 -5.6 -5.8 -5.3 │ floor │ -10.1 -9.8 -9.8 -9.1 -11.3 -11.2 │

[out]: Withdrew more money on Monday to put food on the table for my family.

URTY ¹: Sold at low-of-week.  :-(

ATML: Doing great!

[cash]: Negative balance on Thursday — I bought URTY with unsettled funds.

Stock-trading robot

Ticker Symbol		Buy date			Buy price			Sell date			Sell price			Acct Profit
		Model	Actual		M	A		Model	Actual		M	A		M	A
FRED		OC 20	(Skipped)		$14.31			DE 03			$14.27			-0.07%	+0.00%
URTY ¹		OC 28 13:00	12:00		$79.36	$79.03		DE 01 11:00			$83.10	$83.11		+0.70%	+0.64%
UWM		NO 21	NO 28 12:58		$85.85	$85.72		(Not yet)
ATML		NO 24 09:30	11:31		$7.41	$7.58		(Not yet)
YHOO		DE 04 09:30	NO 26 09:30		$50.19	$51.56		(Not yet)
TZA		DE 01 10:00	(Skipped)		$13.71			DE 03 11:00			$13.07			-0.49%	+0.00%
SRTY ¹		DE 01 12:00			$35.75			DE 02 10:00			$35.12	$35.10		-0.19%	-0.22%
SRTY ²		DE 02 13:00			$34.96			DE 03 10:00			$34.27	$34.29		-0.21%	-0.23%
AGQ		DE 04	09:45		$42.03			(Not yet)
URTY ²		DE 03 15:00	DE 04 11:00		$85.90	$86.52		(Not yet)

FRED: Finally got a ”sell” signal.  I saved a small amount of money by skipping this trade.

YHOO: Finally got a ”buy” signal.  Would have saved money if I had entered the trade properly and not bought so soon.

TZA: This trade failed to go through, apparently due to a momentary glitch at Schwab.  I tried manually buying it at the same price later that day, but the limit-order was not reached so I gave up on it.

AGQ: Bought at low of day!  Then it went lower the next day.

URTY ²: Not sure why this trade was late.  It seemed okay at the time, but now the model says I should have bought earlier.

New trading models

I finally finished a project that’s been on hold for months: I have refactored the trading models so the common code is pulled out into a separate file of macros.  This makes it much easier to just code up a new model when I think of something.  Since none of my current short-selling models seem to work very well, I’ve tried a whole bunch of new short models in the last two weeks — but none of them work well enough to use.  Here are two examples:

Momentum model

Short: Before 4pm, when TRIX(50,2) < 0 TRIX(14,6) < 0 hourly close < SMA(200)-0.8*STDDEV(200) This hour’s True Range > 7*ATR(15)

Cover: Before 4pm, when TRIX(14,6) > 0

Stop: Updated at 4pm, use the lesser of EMA(25) Low_since_purchase plus 4*ATR(15)

Basically, we short when slow-TRIX is favorable, price is below the lower Bollinger band, and has just dropped by a significant amount.  We cover when quick-TRIX is unfavorable.  There is a fixed stop at EMA and a trailing stop that’s based on the Average True Range rather than a fixed percentage.

Results using TWM (2× leveraged inverse ETF):

2002‥2007	+1.1%
2008	+1.8%
2009	-0.1%
2010	-1.1%
2011	+0.5%
2012	-0.4%
2013	-0.1%

This just isn’t very impressive.  In order to avoid losing money during 2002‥2007, I can make only 2% during 2008, when the market dropped 33% in a month!  It dropped 20% in 2011 but I get only 0.5%.

Staircase model

Short: Before 4pm, when Each of the last five hours has had a lower high than the previous one Price is below stop

Stop: Updated every hour, use high from two hours ago.

This model has the problem that Schwab doesn’t give me hourly highs, just high-of-day-so-far, so I would have to use barchart.com whose quotes are delayed by 20 minutes.  So actually implementing this thing would be difficult.

Results using TZA (3× leveraged inverse ETF):

2002‥2007	-7.8%
2008	+2.3%
2009	+0.1%
2010	-1.4%
2011	-0.8%
2012	-1.3%
2013	-0.1%

I studied the chart for 2008 while constructing this model — and it does okay for that year only!  It doesn’t manage to make anything in 2011 and its long-term average is a net loss.  Useless.