Requiem for a drawdown

I lost 6% of my money this week.  Apparently this is referred to as a drawdown.  The standard treatment for drawdowns is “stop trading”: just do paper trades until things start working again.

The last 12 things I bought have all been losers (although IAU might come back soon).  20 of the 26 things I’ve sold since Feb. 22 were losers.  I do not know how to invest in a choppy sideways bear market.  I would actually have been better off just sitting in SPY even though it was declining.

The basic problem seems to be that I made too many bets that were all predicated on the same iffy investment thesis, because I liked the sound of it.  Although I am a “talking dog on the Internet”, I am an investment bear at heart, so theories that the market will soon drop a lot are overly attractive to me.
    New rule: In my spreadsheet, at the bottom of the stop-loss column, show the expected account value if all stops trigger, rather than the percentage-change that clearly wasn’t scary enough.

			│	Daily % gain						│	Stop		│	Realized Gain
Symb	size	when	│	Fri	Mon	Tue	Wed	Thu	Fri	│	Beg	End	│
TZA	19%	MR14	│	-0.7	-7.4	-6.0	-6.9	-8.8	-11.3	│	-11.1	-11.1	│	−11.1%
VXZ	11%	MR15	│	-6.5	-10.7	-10.4	-11.5	-12.3	-11.4	│	-12.3	-12.3	│	−12.6%
TYD	11%	MR15	│	-0.6	-1.9	-3.8	-4.2	-5.6	-6.3	│	-7.6	-7.6	│
TYP	10%	MR16	│	-0.6	-6.0	-5.7	-7.2	-11.8	-13.2	│	-18.3	-15.2	│
UDN	12%	MR22	│			-0.3	-0.8	-0.6	-1.2	│	-2.7	-2.7	│
RA	11%	MR22	│			-0.2	-2.2	-4.0	-1.6	│	-8.4	-8.4	│	−3.4%
IAU	12%	MR24	│					-1.1	-1.0	│	-3.9	-3.9	│
SPY			│	0.0	1.4	1.1	1.4	2.3	2.8	│			│
me			│	0.0	-2.7	-2.6	-3.3	-4.9	-5.6	│			│
(Average)			│							│			│	−9.4%

What a sea of negative gains!  I bought lots of highly volatile stocks that have large negative betas.  SPY went up, so all of those went down simultaneously.  I sold half of TYP which raised its effective stop floor, but not by much.  For the others, the stop depth didn’t change all week because I’m using weekly pSAR.

I bought RA because it was going up—it immediately stopped; then I sold it, so of course it went up afterward.

I bought IAU because it broke new highs.  It immediately pulled back, but traditionally gold stocks go up in the spring so this one might come back, especially if all the G8 currencies decline simultaneously, which might well happen.

Of the stocks I sold last week, only DRETF is worth more now than it was then, but it’s too difficult to do swing-trading on DRETF because it’s a foreign stock (no sell-stops, no online trades anymore).

New Trend Trading Rules

Still being worked on.  Here’s what I have so far:

1. Company’s last earnings report was over two weeks ago and next earnings are not expected for at least two more weeks.
2. Last week’s PPO(4,10,7) > 0.5%.
3. Last week’s PPO rose at least 0.2 from previous week.
4. Last week’s ROC(8) < 4%.
5. Last week’s ADX(6) has +DI > −DI.
6. Last week had some price that was in the bottom half of Bollinger(12,2).
7. Purchase price is no greater than last week’s close
8. Purchase price < 110% of this week’s pSAR(0.2,0.2).

Use weekly pSAR as the stop price.  Keep until stopped out (about two months).  For SPY, this would have made 30% over the last three years, while just holding SPY for that time would have produced a 2% loss.

New Swing Trading Rules

These also need more work:

1. Stock released earnings last night.
2. Earnings were better than analyst estimates.
3. Stock rose over the week preceding the earnings release.
4. Stock fell today.

Buy tomorrow, use daily pSAR for stop, keep until stopped out (about 8 days).  If stock becomes eligible as a trend-trade while being held as a swing, lower the stop price to weekly pSAR and keep going.

Week of 2011 MR 18

I had to lower my loss floor!  Note that the bottom of the graph is now at 97%.  Things looked much better on Wednesday, when SPY was down to 100.4% and I was up to 99.9%; only one more good day and I would be beating SPY year-to-date!  Alas, it was not to be.  Now I have to decide whether to discontinue trading and maybe lie low for awhile.

It pained me to sell DRETF, but SPY has been down for three of the last four weeks and DRETF has a positive beta so “the market” will drag it down.  I sold at a small profit; hopefully I can buy it back someday.

TZA, VXZ, and TYD all gapped up at the open on the days when I bought, so my stops ended up being much more expensive than planned.  EWV and TYP were bought during the day to avoid this.

For TZA, I had planned to sell off chunks during the upward spike that was supposed to occur (but didn’t), then I would buy up some replacement shares when the price came back down.  Instead, only the “buy” triggered so now I have lots of TZA with an uncomfortably-deep stop depth.

I bought EWV because “The average stock in Tokyo will go down after the Fukushima earthquake” seemed like a sure bet.  The Nikkei did indeed go down, but it was sickening to watch so I didn’t want this stock anymore after the first day.  Also it seems that EWV has become detached from its index since the quake and is trading at values that are abnormally low.

			│	% gain: close stop																	│	Realized gain
Symb	size	when	│	Fri			Mon			Tue			Wed			Thu			Fri		│
DRETF	11%	DE16	│	6.0			6.7			6.4			3.9			5.2			5.8		│	+5.8%
HGSI	8%	MR11	│	-3.2	-6.5		-2.9	-6.5		-4.4			-4.9			-6.1			-6.4		│	−6.4%
TZA	20%	MR14	│	new	-9.3		-2.2	-12.2		0.7	-12.2		3.8	-12.2		3.1	-11.8		-0.7	-11.1	│
VXZ	11%	MR15	│				new	-6.4		-3.7	-11.7		-0.2	-11.7		-3.3	-11.7		-6.5	-12.3	│
TYD	11%	MR15	│				new	-2.4		-1.7	-5.0		0.8	-5.0		-0.2	-5.0		-0.6	-7.6	│
EWV	11%	MR15	│							-2.5	-6.9		4.8	-0.7		-0.4			-3.0		│	−1.2%
TYP	11%	MR16	│										2.3	-7.6		-0.2	-7.6		-0.6	-18.3	│
SPY			│	0.0			-0.6			-1.7			-3.6			-2.3			-2.2		│
me			│	0.0			-0.1			-1.0			1.1			0.0			-1.2		│
(Average)			│																		│	−0.4%

On Friday I switched from use daily-pSAR to weekly-pSAR for stocks that I intend to keep for awhile.  This was a huge increase in stop-depth for TYP.

No stocks sold last week are worth more now than they were last Friday.

All of the stocks I currently own started rising on FE 22, so they might all stop out on the same day, leading to a horrible 5% overall loss for such a day.  I’m sitting on 47% cash, but I don’t want to buy anything until I can cut my risk.  Thankfully “the trend is my friend” and the expected direction for TZA is positive.

Week of 2011 MR 11

Another bad week.  My lowest balance of the year to date.  I lost 1.6% this week, while SPY lost 1.2%.  Officially, SPY’s +0.1% gain last week means that we are not yet in a bear market, but for practical purposes I think it is walking like a duck and quacking like one, too.  Obviously I do not yet have any skill in bear-market investing.

 Notes on individual stocks   

(See also the table below.)

DRETF: Released its earnings on FE 24; they were excellent.  Like clockwork, the stock price went up for 8 days, then went back down to maybe 3% more than before the earnings release.  Not a problem: I didn't buy this as a swing trade.

PROV: Released its earnings on JA 24.  Price fell JA 25, recovered JA 26, then jumped for five days.  Then six days of no movement, then seven days of slow uptrend.  Two days of correction, four days of uptrend.  Five more days of no movement and then finally downtrend.  This might be an Elliot Wave 1-2-3-4-5-a pattern, but I don’t quite see it.  The 10-day price channel suggests maybe 1-2-3-4-3-4-5 with the 2 and 4 being horizontal and a not yet visible.  If I had been paying attention to either the news or the price channel, I would have realized that it was too late to buy this again for the current cycle.  Once again we see about two weeks of strong uptrend after good earnings.

MTSC: Released earnings FE 03.  Price jumped FE 04, then a slow rise for five days.  Then the price jumped again FE 14, wobbled for four days, 2-day correction, 7 more days of uptrend, then down.  Another stock that I bought too late in its cycle.  Anyway, there were about 11 good days after earnings release.

BWC: This thing has been rising since October.  It might not be done yet, but my stop was too tight and I got thrown off.  At least I got to sell one quarter at a good price.

PCLN: Bought because indicators were positive, but now the chart suggests a rising wedge.  Perhaps I could buy it again to catch the last little uptrend before it heads downward, but I’m not feeling very lucky right now.

CALD: What a lucky break!  This was a “best of a bad lot” purchase.  It was supposed to have a -3.4% initial stop, but Monday’s closing bid was -5.3% so I had to start with that.  Tuesday: rose 10%!  Then dropped to 7% so I sold ¼, then returned to 10% for the close, so I put in an order to sell another ¼ at Wednesday’s open, and a limit order to sell another ¼ at 13%.  The limit order triggered on Wednesday, but the price fell so I got *only* another 10%.  The final ¼ got stopped out on Thursday.  Total realized gain: 8.4%.  Why can’t all stock purchases be like this?

FARO: What an unlucky break!  My -4.2% stop experienced major “slippage” and ended up selling at -8.1%, which just about wiped out all those gains from CALD.  I suppose it serves me right, using a trade-stop on a NASDAQ stock (which Schwab warned me does not work well).  But at least my trigger price (and actual trade price) were within the official range of the day.  With bid-stops, I have twice gotten stopped out by a bid that officially never happened.  So I think I’ll stick with trade-stops for now.  Maybe I should treat “listed on NASDAQ” as a demerit when selecting a stock to buy?
      FARO reported its earnings FE 23.  The stock rose and/or wobbled for 9 days, then fell.  Now the candlestick chart shows a piercing line, but I’m not buying it.

CBS: Rebought because I sold all my quarters last week for 6% gain and yet CBS still hadn’t fallen yet.  It’s been rising since December but the uptrend seems to be on hold for now.

SEMG: A new stock (issued last November), so still rather opaque.  Starting rising in January for no obvious reason.  The “new uptrend” I thought I was buying has apparently fizzled out.

HGSI: The only stock besides DRETF that I still own.  On Wednesday they announced that the FDA has approved their new medication for sale.  Thursday: stock jumped 13%.  I bought it on Friday, using a limit order to get it for 1% less than Thursday’s close and then giving it a 6.5% stop.  Its low for Friday was only 69¢ above my stop price.  I hope I still have this one on Monday.  Maybe it will consolidate for three days and then go up again.  Size is only 8% because that’s all the cash I had.

Summary of price action

			│	% gain: close stop																	│	Realized gain
Symb	size	when	│	Fri			Mon			Tue			Wed			Thu			Fri		│
DRETF	10%	DE16	│	10.6			11.2			11.3			8.3			5.4			5.3		│
PROV	11%	MR03	│	0.3	-1.7		0.3	-0.9		0.8	-0.4		0.5	ppo		-0.8			-0.1		│	+0.4%
MTSC	11%	MR03	│	2.1	-2.6		0.4	-2.6		1.4	-1.3		0.7	ppo		-2.0			-1.7		│	−0.2%
BWC	11%	MR04	│	-0.1	-4.1		-0.1	-4.0		1.6	-2.7		2.3	-0.7		1.6	-0.1		0.1		│	−0.1%
PCLN	11%	MR07	│	new	-5.7		-0.9	-4.8		-0.4	-3.7		-0.2	-2.8		-1.4			-1.6		│	−2.8%
CALD	11%	MR08	│				new	-5.3		9.3	2.4		8.9	8.5		8.8			8.7		│	+8.4%
FARO	11%	MR09	│							new	-4.3		-4.0	-4.2		-8.0			-5.6		│	−8.1%
CBS	11%	MR09	│							new	-3.0		0.8	-3.4		-0.8	ppo		-0.2		│	−1.2%
SEMG	11%	MR09	│							new	-3.9		0.7	-2.8		-2.4	pat		-2.3		│	−2.9%
HGSI	8%	MR11	│													new	-6.5		-3.2	-6.5	│
SPY			│	0.0			-0.9			0.1			-0.3			-2.1			-1.2		│
me			│	0.0			-0.2			1.2			0.9			-1.1			-1.6		│
(Average)			│																		│	−0.8%

The thin green line for PROV is due to an oopsie.  I thought I had entered a conditional order: “If MTSC rises by 4% then sell ¼ of my MTSC shares at market price”.  But what I actually said was, “If MTSC rises then sell some of my *PROV* shares”.  Since MTSC shares cost 6× as much as PROV, the number of shares I entered was only ⅟₂₅ of my PROV holdings.  When I discovered this, I entered the MTSC trade manually, but the piddling PROV sale had already gone through.

The symbol pat indicates that I sold SEMG because of a bearish candlestick pattern (kicking).  This is supposed to be a “high reliability pattern” but was not confirmed on Friday—meanwhile I had already sold.

Again this week, SPY is shown as if I bought at last Friday’s close.  New this week: me is shown as if I bought “shares” of my own account at Friday’s close.

Where do I go from here?

Only 18% of my account is invested right now.  There’ll be enough cash on Monday to buy something, but there seems to be nothing worth buying.

I looked around at ETF daily news for ideas.  They think VXZ might be good, but it fell on Friday, as did LTPZ.  Since I have declared a “bear market”, TZA seems like an obvious choice—but its pSAR stop is too shallow; I’ll wait until the price jumps out of its doldrums.

I found HGSI via AmericanBulls.  They also suggest GOK and PIR, but neither of these seems ready yet (PPO still negative).

Symb	From	To		Actual		MR04	MR11
MMS	FE18	MR01		+0.0%		2.0%	0.9%
PIKE	FE23	MR02		−3.4%		-2.5%	-5.9%
CXO	FE23	MR04		−1.2%		-1.3%	-4.8%
CBS	FE25	MR03		+6.3%		6.0%	5.8%
LRN	FE25	MR04		+0.8%		-0.6%	-3.4%
BND	MR01	MR03		−0.4%		-0.2%	0.3%

This week I started keeping track of stocks sold last week, to see whether they should be rebought.  I had hopes for MMS, but its run ended.  I rebought CBS, but then its uptrend ended.  BND is up, but not really enough.

I guess I’ll let Monday’s money sit for a day.  On Tuesday there will be massive quantities of cash coming in, enough for about four purchases.  Hopefully there will be something to buy then.

Stock-picking report

Today’s task: choose a stock to buy with the money that will be rejoining cash tomorrow morning (actually, it has already arrived in my account).

Screen:

• Schwab Equity Rating = A, B, C, or D (but not F)
• MACD = Recent cross above signal line
• Price compared with 50-day moving average = Moderately above
• Price compared with 10-day moving average = Moderately above
• Parabolic SAR = Bullish

There are 39 matching stocks in Schwab’s universe.  Of these, 19 did not have gains on Friday so I ignored those.  Here are the other 20:

Company		Yearly gain		Indicators			Chart
Symb	Name	Hold	MACD	ADX	PPO	pSAR	Trend	Sign
ADS	Alliance Data Systems	35%	25%	35	+0.06%	12%	Hold
AWI	Armstrong World Industries	30%	2%	24	+0.03%	10%	Hold
CALD	Callidus Software	70%	85%	16	+0.09%	3%	Hold	Spinning top
CROX	Crocs, Inc.	146%	28%	20	+0.11%	7%	Hold
DKS	Dick's Sporting Goods	52%	45%	20	+0.05%	6%	Hold	High wave
DPZ	Domino's Pizza	25%	14%	17	+0.09%	7%	Hold
ELMG	EMS Technologies	34%	5%	14	+0.59%	6%	Hold	Spinning top
GCO	Genesco Inc.	48%	51%	18	+0.20%	7%	Avoid
KR	The Kroger Co.	4%	8%	33	+0.04%	5%	Avoid	Long white
LINC	Lincoln Educational Svcs	−36%	−4%	12	+0.39%	8%	Buy	Confirmed
MPX	Marine Products Corp.	−3%	−15%	14	+0.17%	7%	Hold	Spinning top
MCRS	MICROS Systems, Inc.	57%	6%	24	+0.26%	5%	Hold
HK	Petrohawk Energy Corp.	1%	−18%	18	+0.09%	6%	Avoid	Long white
PII	Polaris Industries	58%	52%	21	+0.07%	6%	Hold
PCLN	Priceline.com, Inc.	99%	106%	18	+0.24%	6%	Hold
REGN	Regeneron Pharmaceuticals	49%	2%	27	−0.01%	5%	Hold
SPN	Superior Energy Svcs	84%	35%	15	+0.27%	8%	Hold
TLEO	Taleo Corp.	34%	−5%	14	+0.03%	7%	Sell?	Doji star
TMK	Torchmark Corp.	31%	9%	20	+0.01%	3%	Buy	Confirmed
WYNN	Wynn Resorts, Ltd.	93%	29%	25	+0.25%	6%	Hold	Spinning top

For “Yearly gain”, the Hold value is how much I would have gotten if I bought the stock a year ago and just held it; the MACD value is how much I would have gotten if I spent the last year buying the stock whenever MACD(10,30,9) turned positive and selling whenever it turned negative.  Green means MACD yields *more* than just holding the stock; orange means at least half as much; red for even less.  The purpose of this indicator is to tell me how effective MACD (and therefore PPO) is as a predictor for when this stock will produce gains.

The Hold value comes from Google while MACD comes from Best Charts.  The “Indicators” all come from StockCharts.com.  The “Chart” columns come from AmericanBulls.com.

ADX is Wilder’s Average Directional Index or DMI(14).  Green is for values between 25 and 35 (“strong trend”), orange for 15-25 (“weak trend”), red for others.  Schwab says 25 means “buy”; Craig Ferguson says 25-30 is good while 50 means the trend will end soon.

PPO is the Percentage Price Oscillator or PPO(10,30,9).  Positive values indicate a rising trend in recent days.  To avoid bad stuff that has happened in the past, I don’t buy stocks whose PPO is less than +0.10%.  I have not defined an orange range for this indicator.

pSAR is the Parabolic stop & reverse calculation.  I use pSAR(0.2,0.2) for my stop price, so its value is how much risk I would be taking if I choose this stock to buy.  It’s hard to know what colours are appropriate here.  In recent days I have done well with ENTG and TNA, which had 6% stops, so now I’m thinking that maybe 7% could be reasonable. Most 2% stocks have not done well except CBS which did great.  Most 3% and 4% recent stocks did poorly, but troubles in the Mideast interfered.  For now I’m using orange for 2%, 3%, 4%, and 7%, green for 5% and 6, red for others.

Trend indicates whether AmericanBulls thinks I should own this stock or not, based on their automated analysis of its candlestick chart.  Their values are Wait, Buy-if, Buy-confirmed, Hold, Sell-if, and Sell-confirmed.  Sign is today’s candlestick pattern if they recognize one.  I treat “Confirmed” as a sign since they get it from the chart.

* * * * *

Clearly, PCLN is the winner.  It has no red indicators and only one orange.  The runner-up would be GCO because its only red indicator is from AmericanBulls, which I’ve just started using and I don’t know yet how much trust to put in them.  CALD, DKS, DPZ, and PII have only a single red indicator which might improve in a day or two, while CROX, MCRS, and WYNN are not going to get better any time soon.  LINC is an interesting case: so far I have not tried purchasing any stock that has lost value overall for the last 12 months, but if ADX and pSAR improve then this could be a candidate.

I have put in my purchase order for PCLN, with a 6% stop.

Week of 2011 MR 04

Another week where I had less money on Tuesday than at the start of the year, then things mostly picked up by Friday.  For the week, I lost 0.8%, while SPY eked out a 0.1% gain.

This week’s big problem was “the other shoe dropping” from last week.  CXO cratered, triggering my sell-stop on this huge bet.  Because of a snafu with Schwab’s StreetSmart.com system, only half the shares were sold; CXO later recovered and so my loss was reduced by the time PPO turned negative on Thursday.  Still, without CXO my realized gains would have been positive for the week.

			│	% gain: close stop																	│	Realized
Symb	size	when	│	Fri			Mon			Tue			Wed			Thu			Fri		│	gain
DRETF	10%	DE16	│	6.4			7.0			8.6			9.7			10.0			10.6		│
MMS	11%	FE18	│	1.0	-0.3		1.1	0.1		0.2			0.2			2.1			2.2		│	0.0%
PIKE	11%	FE23	│	-1.8	-7.9		-1.4	-6.2		-3.6	ppo		-1.3			0.5			-2.5		│	−3.4%
CXO	38%	FE23	│	1.5	-2.8		-2.5	-3.7		-2.6	-3.7		-1.8	-3.8		-1.7	ppo		-1.3		│	−1.2%
CBS	11%	FE25	│	5.4	-3.7		5.9	0.7		5.1	3.8		5.9	4.3		6.4			6.0		│	6.3%
LRN	11%	FE25	│	2.0	-4.8		-0.9	-3.9		-0.1	-2.7		-1.5	-2.7		0.8	ppo		-0.6		│	0.8%
BND	11%	MR01	│				new	-0.4		0.0	-0.4		-0.2	-0.3		-0.6			-0.2		│	−0.4%
PROV	11%	MR03	│										new	-3.9		0.3	-2.7		0.3	-1.7	│
MTSC	11%	MR03	│										new	-5.3		1.5	-4.1		2.1	-2.6	│
BWC	11%	MR04	│													new	-5.7		-0.1	-4.1	│
SPY			│	0.0			0.6			-1.1			-0.8			0.8			0.1		│
(Average)			│																		│	−0.1%

Formatting changes this week: I’ve decided that subtracting 0.04%/day doesn’t help, because it makes the realized gains inexplicably larger than the last closing or stop price.  So now the table shows plain old percentage gains like everyone else uses.  The “size” column indicates the size of the bet.  My usual size is 11% of the account.  DRETF is a little smaller in order to have a round lot.  The pink half-bar indicates that half of CXO was stopped out on Monday, locking in my loss.  SPY is shown for comparison, but I didn’t own it this week.  Its “gain” figure assumes that I bought at last Friday’s close with no trading cost.

In the StreetSmart.com system, I can now specify a stop for a stock that I don’t yet own.  The notation new indicates that I placed an order to buy the stock, so a stop is specified but there is no closing price yet.

I have decided to switch from MACD to PPO.  Basically, PPO is MACD divided by share price, so it is more comparable between stocks.  The notation ppo indicates that PPO turned negative, so I entered an order to sell at market price the next day; in this case there is no stop because I’m selling regardless of price.

Notes on individual stocks:

DRETF: Doing great!  I have no idea why.  Schwab now says this thing trades on “TSX”; previously they said “OTC”.  This may mean that they are now holding real D.UN shares on my behalf.

MMS: Shook me off, then rose.  PPO turned briefly negative; now it’s positive but only 0.02 and I have a rule not to buy at less than 0.1.  Still, I should recheck next week and consider buying this again.

PIKE: Sold it because PPO turned negative, suffering only half the loss I had been facing when I got up last Wednesday morning to see it already down 6% before I could apply a stop!  Slightly better than the 4% stop-loss I had originally planned to give it.

CXO: Released their earnings report last Thursday, which was a stinker.  Went up on Friday because everything did, then started sagging because of poor earnings.
      New rule: Write down the next earnings date when buying a stock.  Don’t buy stocks that will release earnings within the next week.  If you’re still holding a stock when its Earnings Day arives, sell at the previous close.  For CXO this rule would have cut my loss in half (or prevented it entirely since earnings were so soon after purchase).

CBS: Ran out of quarters to sell!  PPO is still positive; recheck in a few days to see whether it should be bought again.

LRN: The uptrend ended so I didn’t get much.  Price never quite got high enough to sell some shares and lock in gains.

BND: I chuckled to myself while buying this: what does it say about the market’s mood that a bond fund is my best purchase?  It rose enough the first day to justify selling some, but I didn’t sell because it had risen only to break-even after trading costs.  Then it fell.
      New rule: Don’t buy things whose stop depth is so shallow that the first profit exit has no profit.

PROV: Bought it again because it did well after throwing me off last time.
      New rule: After the first day, use an automatic sell-limit to ensure that brief peaks above my strike price will get caught.  I think this rule will eventually catch some profits, but it wouldn’t have helped with PROV or LRN (which hit the peak its first day).

MTSC: Doing great!

I discovered a new website today: AmericanBulls.com.  Despite its name, the site is neither bull nor bear.  Instead, it provides free candlestick-pattern analyses for all US and Canadian stocks!  $35/month to use their screener is a little steep, but the free service seems useful as a cross-check against other screeners.