Why CTAs prefer trend-following over mean-reversion?

[d08]

Neither is my native language english but it means that if you enter at tops and bottoms in the direction of the previous trend then you're a trend follower, opposite trade is reversal to the mean (mean=average).

 

[OddTrader]

http://www.investopedia.com/terms/m/meanreversion.asp

Anyone else other than me would think one is similar to long gamma trades, while the other short?

http://quant.stackexchange.com/questions/946/what-type-of-investor-is-willing-to-be-short-gamma

 

[Deuteronomy_24_7]

reversion to mean is cheaper to execute (passive execution)

Passive execution != 'cheaper to execute.' Possibly passive execution results in better markouts (ie, fills versus the market price X minutes after execution) but that's an empirical question that could go either way. There's no free lunch. Passive execution involves adverse selection (ie, risk of being run over by informed traders).

 

[globalarbtrader]

Passive execution != 'cheaper to execute.' Possibly passive execution results in better markouts (ie, fills versus the market price X minutes after execution) but that's an empirical question that could go either way. There's no free lunch. Passive execution involves adverse selection (ie, risk of being run over by informed traders).

Fair comment. I'm measuring execution cost here purely on the fill price vs the mid. If you're trend following, at least on a short term basis, you'll normally have to pay the spread each time, or miss out on the trend. But this isn't really CTA land - they can take the whole day to achieve fills, which means using a combination of passive and aggressive execution with algos.

GAT

 

[d08]

Fair comment. I'm measuring execution cost here purely on the fill price vs the mid. If you're trend following, at least on a short term basis, you'll normally have to pay the spread each time, or miss out on the trend. But this isn't really CTA land - they can take the whole day to achieve fills, which means using a combination of passive and aggressive execution with algos.

GAT

That's what I do as well. Being passive achieves better average price but on slow periods I won't fill.

 

[Deuteronomy_24_7]

Fair comment. I'm measuring execution cost here purely on the fill price vs the mid.

Which is reasonable, assuming you also account for runaways -- 'marking' unfilled bids to the offer side as of the day's close, or a similar convention.

 

[learner88]

Interesting thread.

I think mean reversion is more for sophisticated quants. CTAs are historically not quants. If the CTAs started off as floor traders, they aren't likely to be educated enough to do quant stuff.

I don't think commodities markets are unsuited for mean reversion strategies. Maybe the younger, more educated CTAs will be able to employ mean reversion strategies in the commodities markets.