Friday, 12 September 2008

Data Says... Short Banks, Long Commodities

Woah. Another rough week. How long have we been saying that? Seems like forever. Trader positioning as reported this Friday by the Commodity Futures Trading Commission suggests the gloom for equities shows no signs of letting up. My BKX Bank Index trading setup turns bearish as of Monday's open. Also ominous: my 10-year Treasury setup goes bearish as of Sept. 29, as reported in last week's post. See my signals and results table linked at my latest signals page for my updates based on this afternoon's Commitments of Traders weekly data.

On the positive side, five of my six highly correlated commodities markets will be in the bullish lane as of Monday, when gold goes bullish. However, all's not sunny in commodities, either. My gold setup just got a new bearish signal from this afternoon's data; my gold setup has a two-week trade delay, which means execution of the signal is on the open of Sept. 29.

As well, natural gas, which isn't highly correlated with any of of the other commodities markets I follow, has just gone back to cash after two weeks on a bullish signal.

Election '08: I call it a toss-up between Osama and McCain, if my signals based on the Commitments of Traders data is any guide. My completely unscientific call is based on the fact that my five U.S. equities setups are neatly divided as of Monday's open of trading, with two bearish, two bullish and one in cash. Recall that the Stock Trader's Almanac folks say that a bad September and October tend to presage a loss for the incumbent party.

New copper setup: Also on my newly updated latest signals table are details of my new trading setup for copper. It's based on my best signals from two groups of traders: the commercial hedgers and the small traders (the little-guy futures and options players). Curiously enough, the best signals I found in both cases trade on the same side as the commercials and the small traders when their net positions hit extremes of bullishness and bearishness as a percentage of the total open interest. I know, I know, the small traders are normally seen as the "dumb money." But in this case, like in many others, that's not always the case. My combination setup puts on a trade only when both the commercials and small traders agree. Otherwise, it's in cash, as it's been 48 percent of the time since the data starts in 1995. Yet, it beat the market handily as the data on that table shows - for example, beating copper by 122 percent since 2003, when trade friction of 0.2 percent per trade is taken into account for commissions and slippage. Also very positive, I think, are the results of my out-of-sample and "walk-forward" tests for this setup. I came up with the walk-forward test in order to see how the setup compares to its "neighbouring" setups - i.e., those with slightly differing parameter values. A large drop-off in performance when those values are varied suggests a setup's strong results are likely just the result of a statistical fluke and probably won't be repeated in real-life trading. This setup scored 95 percent in my walk-around tests - meaning 16 neighbouring setups scored on average 95 percent as well as the original setup in 16 different measures. For more details, see the notes on my latest signals table.

Hope you have a relaxing weekend, and see you back here early next week with a portfolio update.

TAGS: gold, copper, Treasuries, natural gas, 10-year Treasury, BKX, Bank Index, out-of-sample testing, walk-around testing, COT, Commitments of Traders, market timing, trading system development, CFTC, Commodity Futures Trading Commission, COTs Timer

1 comment:

BobR said...

I hope it was a slip of the tongue (or fingers):"I call it a toss-up between Osama and McCain." If not, this is really not necessary.