Friday, 1 February 2008

New S&P 500 Setup: Bullish Signal for Feb. 11

Hi folks. I've just taken another look at possible trading setups based on the Commitments of Traders data on futures and options positioning on the S&P 500. I managed to find a new trading setup that I really like. The results are posted on the Latest Signals page. I think you'll agree they're pretty good. Especially interesting is that, while they've been on a bearish signal during the recent market rout, they flipped to bullish with the Jan. 15 COTs report. The setup uses a three-week trade delay, so that means I'll execute the buy for the open next Monday, Feb. 11. My UltraShort ProShares S&P 500 position has been a great hedge for my portfolio, based on my previous setup's bearish signal. Also worth nothing is that my Russell 2000 setup flipped to bullish a few weeks ago, and with the seven-week trade delay for that setup, the buy is to be executed for Monday, Feb. 18. So all this suggests the bottom really is in around here - although we still need to wait a little longer for the ideal buy point, from the perspective of historic relationships between trading positions and price action.

I've also updated my Sample Spreadsheets page with the new S&P 500 setup so you can download it. See you later this afternoon with the latest signals from today's COTs report. Thanks again to reader Dave for his help locating this new S&P 500 setup. One of the great unexpected benefits of creating this blog is meeting readers to share insights and brainstorm about this fascinating data!

6 comments:

Peter J Barban said...

Alex,

I think your new SP500 set up is a big step backward. Consider this:

Setup, Profit/wk, Drawdown, RAP
Old, 1.43%, 7%, 20.4
New, 1.17%, 25%, 4.7

Those are the numbers I care most about and use in portfolio construction, especially RAP (risk adjusted profit). You new setup offers a much higher drawdown for less profit per week. As an investor I can't see any benefit to switching to the new setup.

Alex Roslin said...

Hi Peter,

Thanks for your question. You're right, the new max drawdown and weekly profit are higher with the new setup, but those kind of stats tend not to be very reliable as measures of a setup's actual reliability. The max drawdown in particular is usually considered suspect because it's based on a single event.

More important numbers that I feel make the new setup vastly superior are these:

Old setup Sharpe ratio: 1.7
New setup Sharpe ratio: 3.4

Old setup confidence interval (profitability): 99.98%
New setup confidence interval (profitability): 99.999%

Old setup confidence interval (beating market): 93%
New setup confidence interval (beating the market): 99.97%

I've found these measures are consistently the best tests of a setup's robustness. The new setup wins hands-down.

All the best,
Alex

Alex Roslin said...

Oops - I misspoke a little in that previous post. Should have said the new max drawdown is higher in the new S&P 500 setup, while the weekly profit is lower. Apologies!

Alex

Alex Roslin said...

One more thing about the new SP500 setup - I plan to soon replace the weekly profit figure with a better measure - Compound Annual Growth Rate, a widely used stat in the business.

By this measure, the new setup is also superior to the old one for this market: CAGR of 18.7% compared to 15.8%.

Regards,
Alex

Peter J Barban said...

I look forward to seeing you publish the Sharpe ratio and CAGR of all your setups. That would help me a lot in choosing my allocations. I'd also like to see the Standard Deviations so I can compare CAGR/StDev.

I understand that it is very difficult to maintain a CAGR/StDev above 1 for a portfolio, even the Endowment funds at Yale and Harvard can't do it.

Alex Roslin said...

Hi Peter,

Thanks for your interest. Will be doing that fairly soon. Still need to update some of the setups first though.

Regards,
Alex