Comments on: Double-Digit Returns in Any Market – Update 2

By: Rick

Rick — Sat, 25 Jun 2011 21:39:23 +0000

Any updates?

By: John

John — Sat, 20 Nov 2010 00:55:38 +0000

AP:
“Another approach is to forgo rebalancing completely, and do “dollar cost averaging”. This assumes that you are willing to start relatively small compared to the maximum amount of margin you want to dedicate to the strategy and let your position(s) build up over time, sampling various points of the market history. ”

It seems to me that there are fluctuations about the mean, so we should sell when above the mean and buy to close when below.

My problem is in assessing the mean. What does everyone think of normalizing the closing prices to equal dollar amounts and using a moving average of their sum to ID entry/exit points?

By: AP

AP — Sat, 20 Nov 2010 00:51:38 +0000

@Teck,
Rebalancing requires discretion and tends to morph into regular “trading around a core position”, demanding time, energy, and attention. Spreading the positions over time would eliminate discretion and psychology. You are creating individual 50-50 initial mini-positions that can get out of balance as much as they want, and you don’t worry too much because each one is small. Over several years (that’s the time scale one needs to consider with dollar cost averaging) the individual mini-positions that went offensively against you would presumably have ranged back and recouped the losses, and since you sampled all the peaks and valleys and everything in between, there is no net trend that went against you; your average entry was more or less in the middle of the range, looking back. In that case you should have harvested the volatility without a trend. That’s the idea, and it’s only based on intuition; I don’t how well it would work in real life. Of course nothing would make this strategy work in a relentless, multi-year, straight-line trend. For example, if hyperinflation sets in and all nominal asset prices go up rain or shine, any variant of this pair-shorting strategy would be a sure loser, unless you turn it into a directional, discretionary play, which is not the point of the method at all, and you don’t need to be shorting the pair to do that. Anyway, in that case, you can switch to a go-long-the-pair strategy, hopefully in time.

By: Teck

Teck — Sat, 20 Nov 2010 00:27:30 +0000

@AP,
My initial thought is that this approach is actually similar to starting with a maximum position which your margin allows, and then rebalancing weekly/monthly, because the effect of the new 50-50 positions periodically would be to “rebalance” the short pair such that the ratio of FAS:FAZ would tend towards 50:50 each time. I would think that by starting small, one would actually lose out on the initial profit potential which an initial large position can potentially create. I would also prefer not to rebalance the short pair when there is no need to (the dollar cost averaging approach might result in this). Hence why not start with a near maximum position (giving yourself 10-20% of your margin as room to rebalance if needed) and only rebalance as and when the FAS:FAZ ratio gets too out of whack and only cut back on positions when you have used 100% of your margin?

Having said that what advantages do you see in the DCA approach? To me it is highly possible that I missed something..

By: AP

AP — Fri, 19 Nov 2010 19:39:57 +0000

Another approach is to forgo rebalancing completely, and do “dollar cost averaging”. This assumes that you are willing to start relatively small compared to the maximum amount of margin you want to dedicate to the strategy and let your position(s) build up over time, sampling various points of the market history. You start a new 50-50 position every week, or every month, etc. I am considering switching to something like that. After a sufficient buildup and hopefully some profits, you cut back significantly and start again. Anyone care to comment about that style of investing in this strategy?

By: AP

AP — Fri, 19 Nov 2010 06:10:08 +0000

@John, The automated backtesting in thinkorswim is done via Prodigio, but it is still not quite ready for prime time. Among other things, it doesn't correct for splits or distributions; makes it nasty to backtest meaningfully. There is a neat playback feature in the main thinkorswim platform, which you can use to backtest manually. Actually I might just do that when I have the time. Incidentally, you can forget about shorting 3x leveraged etfs at thinkorswim; it's not going to happen. You can of course do options, if you choose ones that simulate shorting with very little premium and reasonable spreads involved.

By: AP

AP — Thu, 18 Nov 2010 19:53:55 +0000

@John,
I don’t think it makes a difference on average. There is no edge coming from this strategy intra-day. You might want to submit limit orders and let the market come to you. I haven’t come up with a “rule” about that yet. It might pay to do some technical analysis about support, resistance etc. In other words, as you get fancier about the method it becomes increasingly like regular trading and less like “set it and forget it”.

By: John

John — Thu, 18 Nov 2010 18:54:57 +0000

Another question: at what time of day have you found it best to rebalance/close/open positions?

By: John

John — Thu, 18 Nov 2010 07:39:28 +0000

Thanks for the interesting comments that square with my experience.

I have a question: have any of you found a way to backtest this in an automated way. It clearly won’t work with Strategy Desk. Might it work with thinkorswim?

By: AP

AP — Thu, 18 Nov 2010 06:35:35 +0000

@AP, At the beginning of my reply above I wrote "Jeff asked:"; I meant "Teck asked:". Sorry about oopses.