Monday, June 16, 2008

COST and RSI 2

I'm going to stop posting these RSI studies pretty soon, as my model trading basket is almost complete, but here's is today's prime candidate, Costco.
Just a bit of history here. . .I spent 35 years in Seattle, the last 25 in a place called Kirkland. . .a suburb of Seattle and home of Costco. .which is why all their private label stuff is called "Kirkland". I'm actually member # 86000 so you can tell how old I am. I remember when the first store was opened. . .I was there. . .sorry, no pictures. Back then I was in the custom home building business and built homes for several of the Costco execs . . .believe me . .these guys know how to make money (and spend it).
OK . . .enough reminiscing.
This is a powerful bullish chart by anyone's standard and although buy and hold is the wise man's play, you know me, I prefer to trade it.
The RSI 2 settings for COST are (3,34,80,26,78). Still easily within a single standard deviation of our base model, but the interesting thing here is that the optimized RSI value is 3, not 2. . which is an indication that (this is a quiz)...................the stock tends to trend.
Contrary to our NEM study last week where the short side excelled, COST excels on the long side. . .again no doubt reflecting its overall bullish trend and its constant distant from the 200 WSMA (top most chart is weekly bars).
From trading standpoint, there aren't a lot of trades. 128 over 5 years equals 1 every 2 weeks, with the average length of trade 7 days for the winners and 15 for the losers. A couple stops, fixed and trailing, will knock the later number down substantially and also limit drawdown. Taking the long side only equals about 1 trade a month, holding for a week and then out.
The RSI study may also be a useful cycle indicator for option traders for buy/writes (.64 dividend) and calendars. Limited exchange participation here and although open interest is robust, the ATM spread tends to run .05-.10 on $2.50 strike increments, so not a daytraders fancy.
Probably best to focus on the long side of COST since, as with XLE, the short side is considerably more risky and less profitable. Still, the long/short equity curve contains no hiccups.

6 comments:

sysin3 said...

Good stuff, I appreciate it.

(and you might even be able to help Richard out sometime, lol)

Just for kicks, have a look at RSI(5) on SPY or QQQQ.

a quick eyeball says that might have promise.

klynn55 said...

Why the riddles, rsi(5) is fine, but when to buy and sell, when it's greater than 10 or less than 90, please explain for the unwashed! thanks.

klynn55 said...

"the rsi2 settings for cost are
(3,34,80,26,78)." could you explain those settings. i did a search for
rsi2 at the search box and still could not come up with an explain for the settings, thanks again, kerry

bzbtrader said...

Sysin,
Thanks, I'll check out the possibilities of the RSI5, but TS optimizes the Qs to RSI2, so I might have to adjust entries and exits further.

sysin3 said...

BZB,

curiously, I was looking at Telecharts RSI ...

it seems different from Tradestation ....

I always thought that RSI was pretty well-defined, but now, not at all sure about that.

Not in the mood to investigate tonight ;-)

bzbtrader said...

kerry,
There are no riddles here and I'm sorry you think there are. I'm trying to make things easy, not complicated. Please explain what you don't understand about the May27th post which detailed the system code and default values(2,20,80,20,82. Subsequent RSI2 posts were just variations on the RSI2 theme. You don't need
Tradestation to run the system tests. . .any robust backtesting program should provide the same results.