Friday, 19 December 2008

Bah, Humbug! S&P 500 Bearish Through Holidays

Interesting week. You'd never know from the news that we're in the middle of a powerful rally. Gold stocks and real-estate ETFs have doubled in the last four weeks. I've been playing some of these rallies with discretionary trades while I take another look at my COTs Timer trading setups using detrended price data and Monte Carlo testing.

My first setup that's come out of that process - for the S&P 500 - gave a bearish signal for the start of this week. Seemed a little crazy when the Fed opened the floodgates Tuesday by slashing its target rate to zero, but the index ended up closing about even for the week. Not bad. So what about the much-anticipated Santa Clause rally? Sorry to be a Grinch! The setup is still bearish for two more weeks. It finally goes to cash on the open of Monday, Jan. 5. Bah, humbug! See my spreadsheet for this new setup on my DYI page.

This afternoon's Commitments of Traders data has given my other setups a few new signals this week: cash for the NASDAQ 100, BKX U.S. Bank Index and platinum; and bullish for natural gas and the Nikkei. See my latest signals table for more details on these and my existing signals.

Please keep in mind: I haven't yet subjected these other setups to testing with detrended data and a Monte Carlo test. I have no idea how many I'll keep and which I'll need to replace. These tests are a key way to gauge the risk that the setups came from chance or if they represent an actual market effect. For this reason, I'm temporarily combining those signals with technical indicators for trading as explained in this post.

Have a good weekend and wonderful holidays. See you back here next Friday with the next COT update, or maybe sooner if I finish testing setups in gold, the next market I'm updating.

TAGS: SPX, S&P 500, NASDAQ 100, gold, natural gas, platinum, Nikkei, BKX, Bank Index, crude oil, COT, Commitments of Traders, derivatives, Black Swans, market timing, trading system development, CFTC, Commodity Futures Trading Commission, COTs Timer, out-of-sample testing, walk-around testing


Anonymous said...


"The setup is still bearish for two more weeks. It finally goes to cash on the open of Monday, Dec. 5. "

This seems to infer that you will hold the position for a minimum of 3 weeks? Is this correct? I did not see where there was a 3 week hold rule.

Also, what did the 2nd sentence mean that it goes to cash on the open of Dec 5?

Thanks Alex,

Dan said...

I agree on the S&P 500 but, strictly from a technical stance (COT's aside for the moment). I'm not sure of your comment about the S&P ending flat being "not bad" - the Fed effectively targets rates to zero and the S&P ends flat for the week! That used to be a 120-150 point rally back in the good ol' 1400 point days. The action this week was pathetic. It can't break 920 and the trend line has been broken and confirmed today. I think we're retesting and breaking the the lows as the economy and the dollar reel like a 10th round fighter on the other end of Rocky Balboa's gloves. Now Obama wants foreigners to fund even more spending to boost our economy with his initiatives? It's frustrating and sad.

Alex Roslin said...

Hi Dan,

I agree there's a strong risk we retest the lows and the action wasn't a good sign. I meant "not bad" from the viewpoint of the new setup.


Alex Roslin said...

Hi Mit,

Oops - that should have said Jan. 5, not Dec. 5. I've corrected that line. Thanks for bringing it to my attention.

One week is the only minimum holding period, just because the data comes out only once a week. I can tell how long this setup will still be bearish because of the trade delays of the two individual signals.


Tim said...


SO if I understand, you can tell the future combined signal by looking at this weeks individual signals - that I get. BUT, don't you have a metric built in that would override, so to speak, if the market stays lower on the short side or higher on the long? It just seems counter intuitive that a lagging signal would be nimble enough in the type of fast moving market we have been in. I see where the stops will help.

Also, what day do you close out positions?

Alex Roslin said...

Hi Tim,

I've tested various price filters but always found they degrade results substantially. I haven't tested an extensive set of them, however, so it's possible something could improve the setup.

But note that David Aronson's Evidence-Based Technical Analysis found that none of the 6,000-plus technical setups he tested were robust enough to pass a Monte Carlo test.

I close out positions on a weekly open after a new signal. In other words, if I'm short, I'd go to cash or long if I get a cash or long signal.