Sunday 10 November 2013

Week of 2013 NO 08

IWM rose by 0.7% this week; my account fell 1.4%, matching its all-time low on AU 09.  The loss-floor has risen to -7.8%.  Once again we see that my trading algorithm has a nasty habit of rising slower than the broad market but falling faster.

Euro news: On Thursday, the European Central Bank unexpectedly cut its interest rate in half, to 0.25%.  Naturally, this caused the herd to shift some of its assets from EUR to USD, and so the relative value of the dollar rose sharply.  Bizarrely, the US market briefly rose to a new all-time high before plummeting as one would expect (the dollar rose, so the value in dollars of each stock share should fall).

US news: On Friday, the US Federal Reserve bought up $5 billion in Treasuries, its largest POMO purchase in quite some time.  It seems the sellers of those bonds then used the newly-printed cash to buy stock, causing the market to recover all its losses from Thursday.

End of week allocations:
% gain
size
Max loss
Sym  Buy  Fri  Mon  Tue  Wed  Thu  Fri Beg   End
 
URTY ⁰ OC17 +0.2 11 +0.2 11 +0.2 0
UWM OC24 -0.4 10 -0.2 10 -0.2 10 -0.3 10 -0.7 10 -0.7 10 -0.6 -0.7
FRED OC24 -0.3 9 -0.1 10 -0.2 9 -0.2 9 -0.3 9 -0.3 9 -1.0 -1.0
AGQ OC25 -0.4 10 -0.4 10 -0.4 10 -0.4 0
FNSR ² OC30 -0.4 11 -0.4 11 -0.4 0
TZA OC30 +0.4 11 +0.1 11 +0.1 11 +0.1 11 +0.1 0 -0.4 -0.4
YHOO OC31 -0.0 6 -0.0 6 -0.1 6 -0.0 6 -0.0 6 -0.0 6 -0.5 -0.5
FNSR ³ NO01 -0.1 9 -0.1 9 +0.2 9 +0.1 9 -0.1 9 -0.1 9 -0.7 -0.7
SRTY ¹ NO01 -0.2 10 -0.4 10 -0.4 10 -0.4 10 -0.4 0 -0.7 -0.7
URTY ¹ NO04 +0.1 11 -0.0 11 -0.2 10 -0.4 10 -0.4 10 -0.9 -0.9
SRTY ² NO07 +0.2 10 +0.0 10 -0.6 -0.6
URTY ² NO08 -0.0 11 -1.3 -0.8
Avail. cash 4 -7 13 13 33 23
IWM +31.7 +33.2 +32.7 +32.2 +30.0 +32.4
me -5.0 -5.1 -5.1 -5.5 -6.2 -6.4
floor -8.5 -8.5 -8.5 -8.1 -8.3 -7.8

URTY ¹: Bought with “unsettled funds” because half my money was tied up in already-sold positions (the position-size superscripts for Monday add up to 97%, plus there’s the 10% of money that I took out).  The unsettled-funds rule doesn’t seem to be very strictly enforced, but officially I was supposed to be promising to hold URTY ¹ until at least Tuesday (when URTY ⁰ and FNSR ² settled) but I actually held it until Thursday so no problem.

URTY ²: Price was below SMA(s) when I bought it (hence low stop-price), but closed above that cutoff so the end-of-day update raised the stop considerably.  If Monday is bearish then the stop will go back down.

Stock-trading robot

ATML: No trade.  Indicators were favourable for Tuesday, but price never got down to my SMA(q) limit and so the order expired unfilled.  For Wednesday, indicators were no longer favourable.

Ticker
Symbol
Buy dateBuy priceSell dateSell priceAcct Profit
ModelActualMAModelActual M  A  M  A 
UWM OC 23 OC 24 09:30 $78.25 $78.07 NO 07 NO 07 15:31 $73.23 $73.22 -0.59% -0.65%
FRED OC 18 OC 24 09:30 $16.11 $16.59 (Not yet)
TZA OC 30 12:00 OC 30 12:35 $20.20 $20.19 NO 04 14:00 $20.42 +0.12% +0.08%
YHOO OC 31 09:30 $32.43 $33.07 NO 06 09:30 $33.07 +0.10% -0.03%
FNSR ³ NO 01 09:30 $22.97 NO 08 09:30 $22.70 -0.13%
SRTY ¹ NO 01 13:00 $13.45 NO 04 10:00 NO 04 12:33 $13.16 $12.95 -0.27% -0.40%
URTY ¹ NO 04 12:00 NO 04 12:07 $75.94 $75.80 NO 07 11:00 $73.14 $72.91 -0.40% -0.41%
SRTY ² NO 07 13:00 $13.44 NO 08 10:00 $13.56 $13.51 +0.10% +0.02%
URTY ² NO 08 13:00 $74.88 $74.95 (Not yet)

UWM: Stopped out.  Nicked my stop-price, spent ½ hour below it, then zoomed back up.  Sigh.

FNSR ³: A perfect trade, for a change!

SRTY ¹: Another fumbled trade.  It seems Schwab runs much slower when the market is falling rapidly, increasing the likelihood that the robot’s orders will time out instead of going through.  I didn’t notice for a while so my sale was quite late and I lost 0.1% more than expected.

SRTY ²: Trade went through nearly perfectly, yet still I made 0.1% less than expected — presumably due to some leverage effect.

TRIX-based trading model

Results haven’t been so hot recently, so let’s review prior results for the July‥October period.  (The new TRIX-based model went into production on JL 15.)

Actual Model …
2013 2013 2012 2011 2010 2009 Avg    2008 2007 2006 2005 2004 Avg
AGQ +1.7% +1.9% +3.7% -0.7% +5.1% +7.1% +3.4% -1.2% +1.8% N/A N/A N/A +0.3%
ATML +0.1% -0.5% -2.8% -0.9% +2.4% +0.2% -0.3% -1.4% +0.4% -1.5% -0.9% -0.5% -0.7%
FNSR -2.0% -0.9% -1.7% +1.1% -1.0% +6.3% +0.8% +1.0% -1.3% +2.0% +0.9% -1.7% +0.2%
FRED -0.2% +1.0% -2.7% +0.6% +1.2% -0.1% -0.0% +2.8% -0.5% -0.3% +1.9% +0.6% +0.9%
SCHH +0.0% +0.0% -0.4% +0.5% +0.3% +0.2% +0.1% -0.6% +0.5% +0.7% +0.2% N/A +0.2%
URTY +1.5% +3.3% +2.1% +7.9% +4.2% +7.6% +5.0% +1.2% +0.7% +2.7% -1.3% -1.1% +0.4%
UWM -0.3% +0.7% +0.4% +1.6% +1.1% +1.4% +1.0% -0.4% -0.3% +0.5% +0.3% +0.0% +0.0%
YHOO +0.1% -0.2% +0.6% +1.2% +1.4% +3.3% +1.3% -2.0% +3.4% -0.1% +1.4% +1.0% +0.7%

total  +0.9%  +5.3%  -0.8% +11.3% +14.7% +26.0% +11.3%  -0.6%  +4.7%  +4.3%  +2.8%  -1.2% +2.0%

AGQ: Profits calculated using SLV and then doubled.  Data for SLV goes back only to 2006 or so.  Nasty stopout-with-slippage on AU 30 cost me 0.2% more than predicted.  Otherwise these trades went as per model.

ATML: Robot error in my favour gave me extra profit this year.  Also, the model’s new buy-limit rule was unfavourable this year but wasn’t in actual use for most of the period.  Maybe should have a “sell in May and go away” rule for ATML?

FNSR: Was supposed to buy MA 29, but actually bought JL 15 (so got 0.8% less gain for the period).  Half the years have losses during this 4-month period.

FRED: Was supposed to buy JA 30, making +1.0% for JL and AU, but I didn’t actually do that.  Was supposed to buy again on OC 18 but didn’t actually get started trading this thing until OC 24.

SCHH: No trades this year; results seem within range.  Past years calculated using VNQ since SCHH is so new.  No 2004 data for VNQ.

URTY: Was supposed to buy JN 25 but actually bought JL 15, so got 2.0% less profit than the model predicted for the period.  Otherwise these trades generally went as per model.

UWM: Was supposed to buy JL 03 but actually bought JL 15, so got 1.0% less profit than the model predicted.

YHOO: The new buy-limit rule was unfavourable this year but wasn’t actually in use (so avoided a loss of 0.4%).

total: Use the “Avg” value to fill in for N/A.

Overall summary: As usual, I installed this new model just as the market was turning unfavourable for it.  I just need to keep the robot going until the next favourable period.

MACD-based trading model

Actual Model …
2013 2013 2012 2011 2010 2009 Avg    2008 2007 2006 2005 2004 Avg
SRTY  +0.2%  -0.4%  -0.3%  +6.7%  +0.6%  +1.9%  +1.7% +12.0%  -1.7%  -1.3%  +1.1%  -1.9%  +1.6%

SRTY: This model only does well during years that have bear-market 20% corrections, of which there have been remarkably few during the Bernanke era.  During the period, I switched SRTY over to a TRIX-based model which didn’t work very well so I switched it back; also, due to robot issues there were several loser trades that didn’t go through.  As a result, the “actual” and ”model” results for July‥October 2013 aren’t really comparable.

PPO-based trading model

This model was retuned on AU 01, but that turned out not to have any effect on the trades during the period.

Actual Model …
2013 2013 2012 2011 2010 2009 Avg    2008 2007 2006 2005 2004 Avg
TNA  -0.2%  -0.1%  -1.8%  +1.2%  +4.5%  +9.0%  +2.6% +2.4%  +2.1%  -2.9%  -0.3%  +1.9%  +0.6%
TZA  -0.3%  -0.6%  -0.5% +13.4%  +0.0%  +0.0%  +2.5% +16.4%  +0.0%  -1.3%  +0.0%  -0.0%  +3.0%

TNA: Trades went as expected.  Difference in outcome is due to round-off error and/or leverage effects.

TZA: There was only one trade during the period, which ended with a stop-out.  The model watches IWM and then multiplies profit by 3, but leverage effects caused me to do much better than predicted.

No comments:

Post a Comment