Week of 2013 OC 25

IWM rose by 0.5% this week; my account fell 1.1%.  The loss-floor has fallen to -7.6%.

End of week allocations:

% gain size │ Max loss Sym  Buy │  Fri  Mon  Tue  Wed  Thu  Fri │ Beg   End   URTY OC17 │ +0.4 11 +0.4 11 +0.4 11 +0.3 11 +0.5 11 +0.5 11 │ -1.0 -0.8 FNSR OC21 │ -0.0 10 +0.1 10 -0.1 10 -0.9 9 -1.0 9 │ -1.9 -1.8 YHOO OC21 │ +0.1 13 +0.1 13 -0.2 12 -0.3 13 -0.3 13 │ -1.2 -1.2 UWM OC24 │ +0.1 10 +0.1 10 │ -0.7 -0.7 FRED OC24 │ -0.1 9 -0.2 9 │ -1.0 -1.0 AGQ OC25 │ +0.2 10 │ -0.6 -0.5 IWM │ +34.1 +33.9 +34.2 +33.8 +34.5 +34.6 │ me │ -2.3 -2.3 -2.1 -2.7 -3.3 -3.4 │ floor │ -3.8 -6.5 -6.2 -6.2 -7.1 -7.6 │

FNSR: Ouch!  This stock dropped 8% on Thursday, with no news other than that some VP sold a small fraction of his holdings.  Thankfully I had put only 10% of my money into it, so I lost only 0.8%.

[out]: I have withdrawn 10% of the money from this account, to put food on the table for my family.  I need a new job quick!

Next Monday: The robot has issued a “sell” order for FNSR.

Double-TRIX trading model

Tweak: previously, this model would buy at the opening price and would refuse to place an order if the preceding closing price was greater than SMA(q).  Now it places a limit order for the SMA(q) price, which means the purchase might fail to go through (another limit-buy could be placed for the following day if conditions are still copacetic).  This is a huge improvement in situations (such as the one YHOO was in last month) where the stock keeps rising and stays above SMA(q) for weeks on end, so I would never buy it.  Now the robot will buy as soon as price momentarily pierces the average on its low for a day, even if its closes are never below average.  Still, this wouldn’t have helped with YHOO, which didn’t even touch its average for a month.

I have added a new ticker for this model.  Parameters:

		(daily)
Model ticker:		FRED
Trade ticker		FRED
TRIX(q)		26
TRIX(s)		54
SMA(q)		2
SMA(s)		150
CHANDELIER(c⁺,a⁺)		32, 4.9
CHANDELIER(c⁻,a⁻)		25, 4.6
RETEST(w,m)		5, 1.005

Stock-trading robot

The model-data shown below include the effect of the new tweak, even though I didn’t install the tweak until Wednesday so earlier trades aren’t actually “wrong”.

To deal with the new buy-limit orders, the table below shows the “Model buy-date” without a time unless the limit is above the opening price, in which case the buy-time should have been 09:30 (if the limit is below the open but above the low, the purchase happened “sometime” that day).  The “Model sell-date” continues to show the time as 09:30 because those are market-price orders and always go through at the open.

To make the “Acct Profit” figures more accurate, I now adjust the Model profit based on the size of the position (the model assumes that the position size is always 10%, but sometimes I buy more or less to get an even multiple of 100 shares for non-ETF tickers).

Ticker Symbol		Buy date			Buy price			Sell date			Sell price			Acct Profit
		Model	Actual		M	A		Model	Actual		M	A		M	A
URTY		OC 17 13:00	OC 17 13:30		$75.24	$75.00		(Not yet)
FNSR		OC 22	OC 21 09:30		$26.10	$26.06		(Not yet)
YHOO		OC 23	OC 21 09:30		$33.57	$33.65		OC 25 09:30	OC 24 16:41		$32.40	$33.03		-0.48%	-0.26%
UWM		OC 23	OC 24 09:30		$78.25	$78.07		(Not yet)
FRED		OC 18	OC 24 09:30		$16.11	$16.59		(Not yet)
AGQ		OC 25 09:30			$21.10	$21.07		(Not yet)

FNSR: New model says I should have bought later (price was below SMA(q) at Friday’s close, but stayed above it all day Monday).  Buy-price about the same either way.

YHOO: Stayed above SMA(q) all day Monday and again on Tuesday, so should have bought on Wednesday for a slightly-better price.
      On Thursday, the robot had a problem because it didn’t have historical data for FRED.  So I fixed that and reran the daily update — which created duplicate OC 24 price-records for all tickers!  This caused the robot to say that I should have sold YHOO that morning, so I sold it after hours.  Later I removed the duplicate records and determined that the sale was supposed to be Friday morning — got a better price, anyway.

FRED: I characterized this model quite some time ago, but was waiting for a “Sell” signal before enabling it for trading.  Well, the Sell came and went and I didn’t notice!  So this purchase was late.

AGQ: Normal trade using new tweaked system.  Now sure why I got a better price than the official open.