Friday, 25 January 2008

What's For Lunch? Us?

We're not out of the woods yet, my friends. Are we bear lunch? I dunno. But I can tell you this: this afternoon's Commitments of Traders report ain't very promising for you bulls out there. See my Latest Signals page for all the grim details. My COTs U.S. Composite Equity Index is flat from last week: -0.9964 this week, up a teeny-tiny baby-step from -1.0001. This is based on the readings from my trading setups built on the COTs data for the NASDAQ 100, S&P 500, Dow Jones industrials and Russell 2000. A -1 reading for the index means all four setups have just flashed a sell for execution on next week's open of trading. So even though I'm not convinced we're bear catnip yet, the market might not be ready to bounce, either. (For you newbies, see the explanatory links on the right for more info on how all this works.)

My COTs Agriculture Composite Index (based on my setups for sugar, wheat, soybeans and corn) is also a real downer these days, but bouncing a little better. It's up to a -1.62 from last week's near-record -2.13. Don't know how heartening that is, but hey, it is what it is.

But not all commodities are busted up. My HUI Gold Bugs Index setup is giving a second renewed bullish call, based on the "dumb money" small traders sharply cutting their net long position as a percentage of the total open interest. In silver, the large speculators (also the dumb money) are building up their net long position to somewhat excessively bullish levels, but so far so good; they haven't reached the historic extremes I'm looking for in order to flip me to bearish silver.

The other interesting area is crude oil. The wrong-way large specs have dramatically reduced their net long futures and options position in the past two weeks as crude has corrected. It's given me another renewed bullish signal for crude, one of the better setups I've seen so far in the COTs data, if the past results on that table mean anything.

Of course, it's impossible to say how all these setups will perform in future. We can only guess based on various statistical measures of their robustness. That's really true for all trading systems, whether mechanical or discretionary. Incidentally, if you read something about historic market patterns but there's no mention of backtesting and tests for robustness at all, be skeptical. And that covers a lot of pretty well-known analysts out there!

Speaking of all this, I plan to include some new measures on that table soon that will let you compare them with other kinds of setups being traded out there. I have to say the COTs stand up really well. See you back here early next week with some more details on today's report. Have a good weekend!


Steve said...

I have been reading your blog for awhile now and like the no nonsense approach of picking a some reliable market teas leaves to read and not trying to give some explanation of why the market or a sector is moving the way that it is. I try to do the same thing using a set of momentum indicators. I find a lot of what you write about lining up with what I see on my charts. If you are interested, take a look


Alex Roslin said...

Thanks, Steve!

john said...

Would you care to exchange links with Successful Trading Tips at

Just let me know at

Kind Regards,
John Bougearel

Peter J Barban said...

Hi, Alex.

I have a small account at Scottrade that I'm using to follow your signals. Here's my last Friday breakdown.

Nasdaq 100: 18% ultrashort
Russel 2000: 14% ultrashort
S&P500: 15% ultrashort
DJI: 2% ultralong
Nikkei: 2% long
HUI: 18% long
Silver: 14% long
Oil: 15% long

Next week I'll give you the end of month results.

Alex Roslin said...

Hi Peter,

Thanks for your message. I'm flattered by your belief in the signals. A few comments: please read carefully my portfolio allocation information on my Glossary page and in the notes to the table on the Latest Signals page. My own risk management rule is no more than 20 percent one way (long or short) in any one highly correlated sector. Portfolio sizing and other risk management controls are vital in trading, perhaps more important even that the system you use to get your signals.

Second, I don't know when you've entered those trades, but as I point out on this blog in a couple of places, the setup confidence intervals and past results are based on entering when the signals are given, not at subsequent moments. The performance even on those signals is impossible to predict, but if I enter at some point afterwards, that can be a bit of a crap shoot. I consider the trade to have been done on other grounds (e.g. technical price action) and thus my exit may have to occur on similar grounds (i.e. not waiting for a contrary COTs signal).

Good luck!