Friday, 8 May 2009

Bearish Commercial Traders Refuse to Give In to Rally

Yay, the bear market is over! Phew. I survived. So did you. We're all fine. Never thought I loved bulls so much as right now. I love them so much you hate them compared to how much I love them. (If you have kids, you probably know the reference.) What's that? It might not be over yet? Oh. Great... Man, you just have to know there's a monster selloff coming here. My stops are tight, but I'm still trying to cash in on this rally. But what I know. How about the people who matter? What word from the Commitments of Traders reports? 

For you newbies, these are the Commodity Futures Trading Commission's weekly reports on derivatives positions in 100-odd markets, from the S&P 500 to the Canadian loonie, gold and frozen H1N1 bellies. I've just updated my latest signals table with the calls from my trading setups based on this data this week. Some highlights from this Friday afternoon's data update:

- S&P 500: What in God's name are the commercial hedgers up to? The smart money folks - who tend to be correctly positioned at key market junctures - still loath this rally, friends. It's just bizarre. And spooky. Today's data has them reducing their net short position somewhat, but nowhere near enough to matter in the broader scheme. They're still stuck at a depressive 0.73 standard deviations below my moving average. Yes, that's up from the previous week's 1.45 below average, but not by a lot. As I explained in a mid-week post, my currently under-water bearish signal - based in part on the commercial net short posture - is now close to hitting its stop level, which is 935.74 (for a 7.77-percent loss). I'm not in that trade because I adopted this new setup only after the signal took effect on April 20. Next week could prove critical once we get up to that level - which also coincides with the next major resistance at the January highs.

Another twist: as you'll notice from my latest signals table, my S&P 500 setup has gone to cash, to be executed in three weeks' time. This is based on the "dumb money" small traders, whom my setup fades with a three-week delay. They've gone and heavily cut back their net long position as a portion of the total open interest. They're now 1.25 standard deviations below average, down from 0.02 standard deviations above average last week. This drop is more than enough to trigger a bullish signal from the small traders. And since the two signals that make up my setup now won't be in agreement in three week's time, the setup goes to cash.

BKX Bank Index: You'll see a new setup added to my latest signals table for U.S. banks. It easily outperformed the index - including a 36.2-percent gain in 2008, when banks lost 50 percent - while being in the market just one-quarter of the time. The setup is especially robust from a statistical viewpoint in part because it combines three different groups of traders. This is the first market where I've found the most reliable results to come from a triple signal. (In other markets, three signals tends to degrade the results.) You'll see all the details on the latest signals table (including parameter values in the notes). Note that there's no data for the bank index in the COT reports. My setup is based on following traders in the three-month Eurodollars contract (not the currency, but rather a contract that follows interbank lending rates - an important measure of global liquidity.)

What's the setup saying now? It's been in cash since Dec. 8. The small trader signal (based on trading on the same side as the small trader total open interest when it hits certain extremes of volume) is now bullish. The other two signals in the setup remain bearish, but they could be showing some of those "green shoots" of life we keep hearing about. The large spec net position, which I'm also trading alongside, is now just a hair from flipping to bullish. And the large spec total open interest, which I'm trading on the same side as too, has shot up in each of the past two weeks. It's now 0.6 standard deviations above the mean, up from 0.39 last week. It needs to hit 1.25 above for that signal to go bullish.

- Crude oil: [NOTE: This section is corrected from my original post Friday. The following information applies to the crude oil COT data, not to gold as I had written in haste in my original post. Very sorry for the mix-up, and thanks to an alert reader for bringing this to my attention.] Little change in the crude oil data since last week. The commercials are still strongly bullish - though down in their positioning to 1.23 standard deviations above average, from 1.64 the week before. That's not a major change. They need to fall to 0.2 standard deviations above for my signal to go bearish. Meanwhile, the small traders, whom I also trade on the same side as in my crude oil setup, have jumped back into bullish territory this week. Their net percentage-of-open-interest position is up to 0.59 standard deviations above the average - up from 1.08 standard deviations below last week. I should note, however, that my crude setup will be going to cash next week, for a single week, before returning to bullish. That's because the commercial crude traders got quite bearish the week of March 24. With a four-week trade delay, that change in positioning will take effect next Monday's open of trading.

Sorry to run out of time before I got to my gold setup. Not much to report in that market anyway this week. I should also note that for my gold setup, I've gone back to using my previous parameter values, as you'll see from my updates notes on my latest signals table. Sorry for switching things back and forth like this, but results from some extra testing of setups using signals from three groups of traders caused some changes in the order of my best setups. Have a great weekend and a Happy Mother's Day! Tune in early next week for a portfolio page update.


Ariel said...

Alex, I enjoy reading your work every week. But would your results be different if you used the ES COT data, not the SP (S&P 500) data? For the ES, the commercials went net long as of 05/05. My trading mentor only looked at the COT data for ES, and I have the impression the volumes are higher there too - so I have the impression it is the right one to use.

Ariel said...

Just a note to make sure I said the correct thing in my comment. That the ES commercials went net long, as of May 5.
Also FWIW, their net short positions were declining over the few weeks prior.

manatrader said...

Typo in the signals link points to "blogstpot". Great blog, thanks

Alex Roslin said...

Hi Ariel,

In my backtesting, I've found the most reliable signals come from the SPX data. Not to say you can't also find reliable signals from the other data.


Anonymous said...

Hi Alex,

Do you think today's NG double topping and then big run down is it for this trade? The fundamentals in NG over supply still shows bearish...


Alex Roslin said...

Hi Pete,

I'm not sure how much stock to put in the double-topping, but I did get stopped out of my natural gas positions yesterday on a breakdown through a TDST support line on the 15-minute chart. I think natural has still looks potentially great for a trade if and when it rights itself. It hasn't broken down on the 60-minute or daily charts. Plus, you've got to love that insane bullish divergence in the MACD going back to early last fall. Mind you, I think trading just off oscillator divergences is a recipe for disaster. But I think it suggests natural gas could eventually going to shoot up big-time.