Page 1 of 1
Sharpe > 2.0?
Posted: Mon Sep 03, 2012 12:50 pm
by Macro
Is it possible to build a medium to long term trend following strategy with a sharpe ratio > 2.0, or is this simply unheard of? I'm aware of the pitfalls of the sharpe ratio when concerning divergent strategies, however I'm curious. I think it would be possible, but one would have to blend it with shorter term pattern recognition and mean reversion strategies.
Posted: Mon Sep 03, 2012 1:46 pm
by wguan
you may be interested in the following graph from BlackStar. TF is great for shock absorption in a portfolio, but by itself, I think its tough to get sharpe of >1 consistently.
Posted: Mon Sep 03, 2012 2:03 pm
by BuyHigh SellLow
ed thorp says sharpe = 1 is as good as it gets.
Posted: Mon Sep 03, 2012 10:15 pm
by Aaron01
What is the reason you are seeking to get a Sharpe > 2.0, or is this just a hypothetical?
Posted: Tue Sep 04, 2012 1:10 am
by Macro
Curiosity, really. Most academics as well as practitioners in the industry generally place primacy on the Sharpe ratio over any other measure, hence I thought it would be wise to appease the powers that be. However, wguan's chart of Sharpe convergence helped put a lot of things in perspective.
Posted: Tue Sep 04, 2012 12:14 pm
by sluggo
Where have all the (>1 Sharpe Ratios) gone? As written here in Dec 2009*,
Here are some of the places they've gone
- Made tons of money, retired, closed their fund (cf. Monroe Trout)
- Closed fund to new investors, stopped reporting performance to database providers (cf. E. Thorp)
- Never reported performance to database providers in the first place (cf. Renaissance Medallion)
The guy who made the graph knew this. That's why he carefully put the qualifier "Actively Reporting" in the title of the graph. [/size]
* A fun exercise: see how quickly you can find the original. Google Advanced Search lets you Find Pages With This Exact Phrase.
Original graph from original source in 2009:
http://www.mebanefaber.com/2009/12/03/w ... dium=email
Posted: Tue Sep 04, 2012 6:37 pm
by ecritt
Sluggo,
You offer 3 examples of programs that enjoyed sharpe ratios greater than 1.0, but are not included in the sample used for the chart. I would counter with the suggestion that there are probably hundreds, perhaps thousands of programs that suffered sharpe ratios well below 1.0, perhaps even negative, that are not included in the sample. In other words, despite your observation, the sample is still probably too optimistic.