06/30/2012 two additional notes on position sizes

Just two short points:
A cash only account trying to follow the large model would begin with ten equal sized positions, and would NOT re-invest dividends.  The reason for this is that over time each stock will have different amounts of gains, and the cash from the dividends would tend to limit the imbalances in the position sizes.

A margin account trying to follow the large model would begin with ten equal sized positions, and would not re-invest dividends either.  But the amount invested for each new stock would always be the same: 10% of the highest value the account has achieved.  The total drawdowns would be the same, but the recovery would be faster and of greater magnitude.

06/30/2012 short and simple

Small Portfolio	XLF & IAU	7.87%
Position	Date	Return	Days
GCI	7/14/2011	13.48%	352
CSGS	10/3/2011	36.71%	271
NLY	10/25/2011	9.53%	249
KBR	10/27/2011	-14.79%	247
VG	10/27/2011	-38.91%	247
BT	1/4/2012	6.38%	178
SAI	5/30/2012	9.39%	31
XEC	6/5/2012	9.80%	25
DECK	6/15/2012	-8.29%	15
FCX	6/25/2012	6.40%	5
S&P	Annualized	1.16%
Small Portfolio	Annualized	7.26%
Large Portfolio	Annualized	8.19%

After talking to some friends who were trying (unsuccessfully) to follow the model, I’m simplifying the blog.

They all basically asked for long-only, un-timed, and un-margined stock ideas.

So I’ll be sharing my long-only, un-timed, and un-margined stock ideas.

When I started sharing this, I wasn’t reporting my short and hedging ideas for the first few months, and it seems that was probably the best way to go.  Right now all my margined plays have totaled a 13.76% annualized return – but it’s been a choppy sideways market for the past year, and long term a long only position WILL eventually grind out an advantage.

I do consult two sources for market timing – my friend Len and the site www.effectivevolume.com – and I’ll leave timing in their capable (and superior) hands.  My goal is to create a model that can be used by slow pokes who don’t short the market and don’t use margin.

For you slow pokes out there – the next scheduled stock rotation is this coming Friday, July 6, 2012.

The small portfolio will still be open to end of day trades, about 3:30-4:00.  Those should average about a trade a month (and cost me a grand total of five minutes of my day).  The large portfolio should average a trade every one to two weeks.

My one day trading experiment this week was exciting, but unworkable.  I was on vacation this week and don’t plan to STAY on vacation.  I have a job that I enjoy. I don’t have time to day trade.  And this blog is for folks like me.

As far as the world markets are concerned, the fact that I’m dropping hedging from the model doesn’t mean I’m not hedging myself.  I don’t believe this hogwash in Europe, and I don’t have any grand hopes for a political miracle in this country either.  And for you folks who fondly remember Ronald Reagan… even if this were 1980, we would still have two years of bear market left to go.

Before this is over I expect the S&P, my small portfolio, AND my large portfolio, to fall underwater at least one more time.

Tim

06/29/2012 realtime trade

Covered the 50% short position on XLU for a 0.87% loss and replaced it with a 50% short position in XLK.

06/29/2012 new format

Condition	Bear Market
Small Portfolio	XLF	7.29%
10% Margin	XLF (long)	-0.01%
50% Margin	XLU (short)	-0.32%
Position	Date	Return	Days
GCI	7/14/2011	12.94%	350
CSGS	10/3/2011	32.99%	269
NLY	10/25/2011	8.62%	247
KBR	10/27/2011	-17.86%	245
VG	10/27/2011	-40.73%	245
BT	1/4/2012	3.66%	176
SAI	5/30/2012	7.13%	29
XEC	6/5/2012	1.02%	23
DECK	6/15/2012	-9.88%	13
FCX	6/25/2012	0.75%	3
S&P	Annualized	-1.11%
Small Portfolio	Annualized	6.76%
Mousetrap	Annualized	5.49%
Margin	Annualized	11.23%

The full model is 110% long and 50% short – slightly long for a bear market.

I overstated the 3.25% gain on the XLK short earlier today.  The actual profit was 1.25%.

In any case, the Margin account represents short-only trades from 5/31/2011.  At that time the settings were either no margin (in a bull market) or 100% short (i.e. fully hedged in a bear).  After a good bit of review I found that a better default setting for a bear market is 100% long and 50% short.  Because of the recent dip the model is now 110% long and 50% short.

The full range of long settings in the past decade were from 55% to 130% long, and from 40% to 90% short – depending on market and account shifts.

These will not be adjusted on a daily basis, and my goal is to make as few trades as possible.

The ten fundamental selections will be rotated on a set schedule (the next rotation target is 7/6/2012).

The two margined ETFs will trade in real time at the end of the day. 

Tim

06/28/2012 end of day trade

Covered the XLK short for a 3.25% profit.

Replaced it with a short in XLU.

Current model ratios are 110% long XLF and 50% short XLU.

06/28/2012 and... back to financials

In my first and only attempt to daytrade, I'm back to XLF, financials.
Still plan to allow end of day trades on the realtime model -- but no more fakeouts in the beginning of the day.

06/28/2012 realtime note

Realtime version of my model shifted from XLF (financials) to XLV (healthcare) on the long position.

06/25/2012 correct ratios

The correct ratios for a hedged position are
110% long (XLF)
50% short (XLK)

For a cash only account the setting should be

60% long (XLF)

I increased my long position to 110% this morning, and have set orders for tomorrow to reduce my XLK short position to 50%.

The Mousetrap holdings are a bit tricky, since they are rotated on a fixed schedule (next rotation will be sometime close to July 6th).

There are two logical solutions:

1) Adjust the size of each stock I rotate into based on the correct long exposure for the full portfolio.
2) Adjust the number of stocks I hold based on the correct long exposure for the full portfolio.

Since this is trading I'll try for a third alternative...

In any case, a word about the portion exposure on the model.  In behavioral finance traders try to mean revert the broad market and end up doing momentum trading on their account.  If the market dips, they'll want to buy the dip -- but they end up piling on as they make money or selling positions as they lose money.

Both of those approaches actually lose money.

I've simply reversed the process.  I do mean reversion on the account and momentum on the market.

The math works.

But the proof of the pudding is in the eating.  It's a theory; it backtests well -- only live trading will prove the concept.

So, the Mousetrap is basically complete.  It's time to simply let some more months grind past.

Tim

06/25/2012 after the open

Finally fixed the math that was nagging me.

The purple line is the correct formula for a bear market sector configuration (which we have now).

The greenish-gold line is best for a bull market configuration.

In any case, current confiuguration calls for:

100% long Mousetrap stocks.
10% long XLF
50% short XLK

What was nagging me was the problem that I will have to adjust the long positions AND the short positions in a bear market, instead of just one.  I'll have to work out an easily readable way of presenting it but for now I've added a 10% long position in XLF and dropped the XLK short position to 50%.

06/25/2012 on second thought... (pre-market)

On some further analysis, a partially hedged position has a better reward/risk ratio – especially in bear market sector configurations:

I am cancelling my order to cover the XLK short and to buy XLF.

Right now the model has been fully hedged, but in bear market configurations a half hedge is sufficient: approximately 50% of the long positions.

It’s not EXACTLY 50%, but I’m not going to do the math this morning.

I am still planning to exchange PDLI for FCX, as long as they don’t gap away from each other.

Just to clarify a bit...

The hedged model isn't exactly calling for 50% short XLK, but rather 55% short XLK and 10% long XLF (in addition to 100% long the fundamental selections).

The problem is that the long and short percentages adjust both by the state of the market AND by the state of the account.

The question is, what's the best way to PRESENT this on the blog in a tradable manner? Do I actually present both long and short margin positions, or just the net difference, which would be 100% long ten stocks and 45% short XLK (55% short - 10% is 45% net short).

AND I'm on vacation! Not the best time to do a bunch of changes.

So, eh, 50% short XLK is close enough for government work (as my old father in law used to say...)

06/23/2012 margin and stock changes

Condition	Bear Market
S&P Target	1280
Small Portfolio	IAU & XLF	4.14%
Margin (short)	XLK	5.03%
Position	Date	Return	Days
GCI	7/14/2011	4.01%	345
CSGS	10/3/2011	34.73%	264
NLY	10/25/2011	11.81%	242
KBR	10/27/2011	-17.97%	240
VG	10/27/2011	-42.86%	240
BT	1/4/2012	2.79%	171
PDLI	3/7/2012	8.91%	108
SAI	5/30/2012	7.04%	24
XEC	6/5/2012	-1.91%	18
DECK	6/15/2012	-6.23%	8
S&P	Annualized	-0.71%
Small Portfolio	Annualized	3.89%
Mousetrap	Annualized	4.66%
Margin	Annualized	9.39%

NOTE: selling PDLI and buying FCX on Monday.

ALSO: covering the 100% short position on XLK and adding a 10% long position in XLF.

Both are explained below…

Any trading system has a number of elements that need to work in harmony:

1)      The technicals (i.e. money flowing into a stock through investor interest)

2)      The fundamentals (i.e. money flowing into a company through business)

3)      Exit criteria

4)      Position size

I’ve worked on each of these in turn and have finally reached the fourth: position size.

Red is the S&P, blue is fixed, green is variable.

In tests of the basic sector model, a fixed size at 100% invested outperforms the S&P, but a variable sized position, ranging from 78-130% invested has less of a drawdown and better returns (i.e. less risk and more reward).

Back tests are one thing – and live tests are quite another.

In any case, I’ve thought long and hard about this, and there is no good way to properly apply the variable investment allocations in the Mousetrap investments.  They will remain as before: 10% of the size of the portfolio, each.

What will change is the way I report the margin.

Instead of:

Margin (short)

XLK

5.03%

It will become (on Monday morning):

10% Margin

XLF (long)

5.03%

That’s a little more information than you’ve been getting.  Right now I’ve simply been hedged: 100% long the Mousetrap stocks and 100% short XLK.  The Margin will no longer be either fully hedged or in cash, but will instead either be partially long or partially short.

If the size of my portfolio on the variable sized model is supposed to be 90%, for instance, instead of selling one of the Mousetrap stocks I would be 10% short XLK.  However, right now the variable sized model is calling for a 10% long position on Margin.

At the current price it’s actually 9.31%, but I’ll round it to 10%.

In addition I will make a mid week post on the blog to show limit order settings to add or take away from the margin position.  I’ll start with 10% increments, but I may end up doing 5% increments.  It depends on how often the margin sizes would trigger a trade.  I really don’t want to trade more than I have to.

Also, by making limit orders I can set it at night and forget about it during the day. 

Finally, on Monday I plan to sell PDLI and buy FCX, Freep't-McMoRan C&G in the MINING industry.  They’ve been slammed in the deflation scare and look like a good bargain at the moment.  PDLI is in the drug industry, and looks to have gained some price in anticipation of good news from the Supreme Court this week.  Instead of selling the news, I plan to sell before the news.

I’ll be on vacation this week.  Have a fabulous week and good luck in the market!

Best,

Tim