Who says the Commitments of Traders data has lost its relevance? I've just updated my latest signals page table with the results of my newly revised trading setup for Japan's Nikkei Average. Like all my other setups, this one trades mechanically solely based on the COT data, as reported weekly by the U.S. Commodity Futures Trading Commission. The data tells us how trillions of dollars are positioned in the major markets.
My Nikkei setup, which has been in cash since the beginning of the month, trades when the small traders and large speculators (the big investment firms and hedge funds) hit extremes in their futures and options positioning. It combines two of my best signals from the small traders and large specs. With starting equity of $100 in 1995, the setup ended up with $885.60, compared to $79.70 for the Nikkei itself. It also beat the Nikkei by 69 percent since 2003 after a per-trade friction of 0.2 percent for commissions and slippage - even though the setup was in the market less than 50 percent of the time. The setup also scores quite well in measures of statistical robustness, as you can see from that table. That includes a pretty impressive 97 percent score on my series of 16 walk-around tests. I came up with the walk-around test as a way to see how well "neighbouring" setups with slightly differing parameter values do as compared to the original setup. It's a way to see if the setup does well solely because of a statistical fluke. The result means that in 16 such tests, 16 different measures scored 97 percent as well as the original setup.
I'm looking forward to trading this setup because the Nikkei is uncorrelated to other markets and thus adds diversity to my portfolio. Also, there is a relatively new ProShares leveraged short ETF for the Nikkei, EWV, that lets me trade short signals, along with the existing funds for the long side. Good luck this week!
TAGS: Nikkei, ProShares, walk-around testing, COT, Commitments of Traders, market timing, trading system development, CFTC, Commodity Futures Trading Commission, COTs Timer
4 comments:
Alex watch the Nikkei tonight and tell me tomorrow that it is not correlated to the US. Just like the Nasdaq goes with spx and djia, they all fall and rise together; maybe some not as fast/far but they all move together.
Hi IC,
Sure, those markets often move together, especially in shorter-term periods. But take a longer view and you'll often find greater divergences. My own study of the data is based on the 1995-2007 period.
Regards,
Alex
Are you using COT data from Nikkei futures to model Nikkei price changes or price changes in the EWJ ETF? They are not that well correlated (R^2 about .2 daily and .6 weekly or monthly), as EWJ tracks the MSCI Japan and not the Nikkei.
Hi Anonymous,
I am using Nikkei price data.
Regards,
Alex
Post a Comment