In terms of tracking the systems and their performance with the hypothetical portfolios, I will be completely mechanical. In my personal trading, I reserve the opportunity to deviate from the system, but I don’t intend to do that often, because I believe the discretion is better put at the design and testing phases of system development, and not at the execution phase. If I do anything in my personal portfolio, I intend to announce it on the site the night before execution, if at all possible. The interesting thing will be seeing if my occasional deviations add, or subtract, value from the system. For example, INP presented two opportunities to me, one to add value (by trimming when it traded at surplus to NAV) and one to subtract value (by panicking and selling under $90). I passed on both and stuck to the system.

There are various ways to reduce volatility in systems like Rotational, especially downside volatility. I considered stop losses, increasing the numerical diversification, a “faster” momentum measure, and forcing an asset class diversification regardless of trend strength. In the end, I decided to go with the latter three. The bounces in oil, gold, the Euro, and the long bond ETFs helped defray last week’s fall quite a bit. Also, the hardest falls could by had by the leading momentum players – but having more entries means that I’m more likely to have something moving up the ladder at the same time something else is moving down. Of course, in the long run these techniques to defray volatility will degrade CAGR, it’s a question of making a tradeoff that I can live with.

As mentioned in comment 2, above, the real danger in this system is a cessation of trending for an extended period. The whipsawing will hurt more than having one trend end while another starts (and gets picked up by the system).

One of the systems I’m tracking is partially based on buying extremes of fear, and another system I’ve tested (but discarded as too short-term and active for my tastes) buys these kinds of dips. As scary as it seems, there’s more money made buying panic than selling into it.

There’s nothing new under the sun, and the ideas are pretty old. Trading relative strength systems go back at least to O’Shaughnessy, and he also researched adding a valuation overlay. There are some other papers suggesting that the spread of relative strength in an asset class has implications for performance of them long or with a short extension. Similarly, there are lots of academic papers on asset class allocation, including domestic and foreign stocks, bonds, and commodities in the mix. Mebane Faber has some work on trend-following models in asset allocation, and I think he researched relative strength but instead went with long/cash in each bucket based on trend and applies leverage to juice return if desired.

I think I’m the only one writing about mixing relative strength and asset allocations with a small component of forced asset allocation, but I could be wrong. The odds are that somebody else is doing it and not writing about it. My momentum measure is also “faster” i.e. shorter than what’s found in the literature, because the limitations of most testing databases are total return series at month-end, and you can get some different results using daily bars to generate momentum measures.

From what I’ve seen, I like the results of the various combinations I’ve seen used by others. I think that the number of positions, some (but not a completely) forced cross-asset allocation, and a slightly quicker momentum measure will produce improved results.

I’m glad you appreciate the website! Good to have you as a member, and great questions!

I just started watching your rotational suggestions. It’s an interesting technique. I’m curious, are you a hold to the numbers and follow your schedule kind of guy or will you react to extreme market conditions, say like the just posted jobs numbers, and move up a scheduled reassessment? One of the things I have noticed within the short time I’ve been aware of sector rotational systems is that because of the weighting they can go down fast and hard once the sector goes bad and we all know how easy it is to get into a hole but hard to climb out.

Also, do you know any of the history of the rotational concept? Who was the originator and why do you expect it to work better than previous systems?

Thanks and I really appreciate your website, you’ve done a lot of work and thinking.

My testing didn’t include dividends, but a dividend-adjusted metric would be very close in most cases, and usually dividend yield isn’t enough to make a difference. If I ever have any situations where momentum is roughly equal but dividend yields do make a difference, I’ll point it out. That’s most likely to happen in the currency or bond asset classes.