This by itself doesn't mean a whole heck of a lot. The changes aren't anywhere near enough to flip my trading setup for the S&P 500 - which has been bullish since Aug. 24 - to cash or bearish. As well, the correlations between this data and next-week S&P 500 prices are so small it's hard to base decisions on them. (In fact, the small trader net position as a percentage of the total open interest has a slightly positive, albeit tiny 0.18 correlation with next week's S&P 500. Meanwhile, the commercials have a negative -0.27 correlation. So much for the popular wisdom that we should trade alongside the commercials or fade the small traders based on their weekly COT position changes. Doh!)
What is more interesting, however, is what's going on in the three-month Eurodollar contract. (That's the interest rate, not the currency.) This widely watched measure of global liquidity is putting up a caution flag. The large speculator total open interest has suddenly taken a nosedive. That's important because this indicator has a strong 0.61 correlation with next week's U.S BKX Bank Index. My setup based on the Eurodollar data remains in cash for the eighth consecutive week. It takes a position only when three different groups of traders agree in their positioning, which hasn't occurred. But this change in the data could spell more turbulence for the coming week.
Also, my setup for natural gas will go bearish with a one-week delay - i.e. on the open of trading the week of Oct. 5. That signal will not last long, however. One of the components of that setup - a signal that trades alongside the small trader total open interest - has gone bullish this week. That component, however, trades with a two-week delay, so it doesn't impact the coming signal for Oct. 5. It's just to say that the setup won't remain bearish for long and will go to cash or bullish after a single week of being bearish. Interestingly, both components of this setup have strong correlations of nearly 0.60 with the price of natural gas the following week. Yet, the two groups of traders are giving very conflicting reads of the market in the latest COT report. That, again, shows the importance of relying on more than one group of traders when attempting to interpret this data.
Hope you did okay this week and that you have a great weekend. Tune back in early next week for my portfolio update. And please keep your comments coming about the CFTC's proposed changes to the COT report. Read my post for more details ("Save the COT Report!") and a comment sent in by researchers in the Netherlands (scroll down to the end of this post).
10 comments:
I like your eurodollar signal. It's something that I keep watch of. But as of late, it has fallen off my radar. Thanks for bringing it up again.
Alex, bonds have never been my strong point, so when you say that the large speculator open interest has taken a nosedive on that Eurodollar contract, what would that mean say, for the long bond? Yields up or down? Thanks
thanks for the analysis
it has been a great tool to study
have you plotted the weekly changes against SP500
Hey Alex,
One trick I read in a book regarding cot data is to add up all the cot data of the e-mini sp, the full contract dow, and so on.
Did you also do that or do you just analyse one instrument?
Alex,
Just a tip to promote your website...
I follow some blogs + people on twitter related to trading (as I also blog on automated trading systems). Every time they mention COTs reports I send them a link to your blog as it is the best I could find on the net - and they usually like it very much. You could build more awareness of your blog by reaching the tchwatter (chatter + twitter) on COTs. You could use the real-time twitter search to look for all people twittering with specific words such as "COT" and dropping them an @ reply mentioning your website. I am sure they would appreciate and you would build up your readership...
As an example I just recommended your blog to @droskill and he replied "@JezLiberty Excellent find - thanks for forwarding!"
Just an idea...
Jez
Thanks, Jez!
Hi Bjorn,
Thanks for your question. I haven't done that, but it's an interesting idea. See my FAQ page for more.
Regards,
Alex
Hi Simonsays,
If the large specs are bearish on the 30-year bond, that tends to be bearish for the bond - which means the bond yield would go up.
Regards,
Alex
Does your natural gas trade delay take into consideration the rollover of contracts with a $1 contango in the forward month contract? Would that not suggest the bearish position be taken on the day the contract rolls over instead of a delay?
Hi Roger,
That's not a factor that's involved in my backtesting. I suppose one could include it, but I didn't and yet found good robust results. The system or a setup could of course be further tested with such a factor - or price-based filters - but the downside is that can actually reduce robustness. I'm not saying it would for certain, but I can say that when I've experimented with some additional filters in some setups it did. In any case, I wouldn't include that in my trading if I hadn't included it in the backtesting.
Regards,
Alex
Post a Comment