As a follow up to our earlier post “Let’s Talk About Professional Funds“, here is a great graphic from the guys at BlackStar showing the unsustainability of a Sharpe Ratio over 1 for longer than short periods of time (and this is for CTAs that still exist!!) :
how about longterm sharpe ratios of 42,397.57?
http://www.covestor.com/rankings/portfolio?perf...
Mr. Sykes illustrates my point. If I put $1 million with him today and he reproduces those Covestor results over the next five years, I'll have over $67 trillion. That's trillion, with a “T”. The combined GDP of the United States, China, and every member nation of the European union would barely cover half of that.
That's a great study and I have taken the liberty to reference it in my newer blog post “A robust story: Cockroach vs. Cheetah” http://bit.ly/5Kvsiq – good timing… The study/chart above illustrates the point perfectly, that a robust system will ensure survival – but does not look that attractive (ie Sharpe ratio <1) whereas systems that look very attractive (Sharpe ratio >1) are likely to fail in the long term.
Is there a link to the whole study/article from BlackStar?
PS: Isn't Tim Sykes just running some sort of “pump and dump” scheme (scam)?…
I've no idea what Mr. Sykes does, but it isn't compounding 3,000% returns a year. The richest man in the world, Warren Buffett, got there by compounding at less than 20% per year, with a Sharpe Ratio of about 0.5, and he wasn't born rich.
I think you need to make a distinction between ridiculously high returns with a >1 Sharpe starting from 10k and the same starting from 1MM. The prior over two years could return extreme results, while the latter would inevitably start distorting market patterns and become the target of some quant's terminator project. There's a point of diminishing returns as a previously successful strategy starts to move price, not just from a scaling perspective but also from a model risk perspective. It would be interesting to see more data on how systems break as their AUM grows.
WEB wasn't born “rich?” Surely you jest! Or, at least exaggerate.
WEB was a privileged child, the son of a Congressman. Why did WEB get started in the financial markets at such a young age? Because his family had access to the markets through a stockbrokerage business. By 1940s terms, he WAS “rich.” Maybe not top 1% rich or top 0.01% rich, but bloody well f-ing off to start with.
WEB didn't get as rich as he did by compounding 20% – he got that way by BUILDING BUSINESSES.
The biggest flaw most people commit when looking at WEB's returns is comparing apples to oranges. I don't really give a hoot how much the “book value of BRK” has compounded. You know why??? Because my broker doesn't give a hoot about the BOOK VALUE of my trades, combined with my non-stock investments and personal business. I also don't buy whole businesses and have power over their direction and management.
The only proper comparison of WEB's investing ability to the average person's is to take the compounded returns of WEB's and BRK's positions, counting ONLY those positions where they have no controlling stake. To my knowledge, nobody's ever done this comparison.
I take it back. Warren E. Buffett was BORN AT LEAST TOP 1% RICH.
Daddy was a four-term Congressman who owned an investment business?
Granddad owned a grocery store?
Chart that up on a histogram of net worth and that's “rich.” Top 1% or higher. No doubt.
I meant he wasn't born a billionaire, like most billionaires.
Going from the top 1% of wealth, to the top 0.001% of wealth, isn't nearly as impressive to me as going from the median to the top 10%. But maybe that's just me?
My second point – that WEB got that way in a way you didn't describe – is just as important. I think next time Forbes comes out with their “World's Richest” list, you should do a research project on the members of it.
Last time I went through the top 20 or so was a couple of years back, but there were three overriding themes in the list.
1 – born privileged,
2 – political connections, and
3 – building businesses.
WEB's “stock-market investing” expertise, which IMO has never been properly quantified on an apples-to-apples basis, probably mattered very little in his accumulation of wealth. Maybe if someone does the study on a proper basis, we'll see what that compounding really is.
Mebane, interested in that idea? Sounds vaguely similar to “alpha clone” doesn't it?
Your points are interesting, and notice that I'm not disagreeing with you. But the topic is sustainable Sharpe ratios and annual rates of return. I simply pointed out that the richest man in the world happens to be an asset manager with a track record that is widely accepted as being superior to most. His CAGR has been below 20% and his Sharpe ratio has been below 0.5. These numbers are much lower than what many market participants believe is realistic.
I recall Buffett has stated that he views it as possible to compound at 100% per year with small scale assets.
If you look at early performances of “famous” (industry-wide) hedge funds, they tended to do high double or triple digits their first year or two and then settled back to the low double digits.
I don’t disagree with you that (1) WEB is one of the richest men in the world, (2) his track record is “widely accepted” as superior, (3) his CAGR is below 20 and Sharpe below 0.5, and (4) many feel that sustained higher ranges of compounding are possible.
I do think that mentioning (1) through (3) in one sentence is misleading, because it might draw someone to the false conclusion that one who is NOT rich now can get “ultra-rich” with his kind of track record. Buffett started rich and politically connected, and built his wealth the way MOST of the wealthy build it – through building BUSINESSES.
On what level of gain through market speculation is achievable long-term, check out the CS/Tremont Hedge Fund index when you get a chance. The granddaddy of categories, “Global Macro,” compounds at under 20% long-term. Studies suggest that HF indices have systematic errors that overstate their actual returns. I’m inclined to state that mid-20s to low-30s is probably the 99th percentile of sustained CAGR you’ll see from anything that is both survivable and scalable.
On Sharpe ratios, ignore them. They’re worth than worthless. Examples follow.
The CBOE PutWrite index, based on a totally mechanical strategy, has a Sharpe at RF=5% of 0.53, INCLUDING both October 1987 and October 2008. Going from November 1987 through September 2008, it has a Sharpe of 0.81. Is that Sharpe of 0.53 “good?”, seeing as how it came from an idiot play?
The CBOE PW has a better Sharpe at Buffett, who compounded (arguably! VERY arguably) his investments “near” 20%. The 23 1/3-year compounding of the CBOE PW is 10.3%.
Similarly, I have found in backtesting that the highest compounding for my systems is very often at Sharpes that are sub-optimal compared to what system parameters suggest. This seems to “disprove” what some mathematicians and HF managers seem to have “proved” about the Sharpe, but their “proofs” are really just a function of their flawed assumptions about the statistical properties of their return streams.
I know of one HF that went from +100% in 2007 and accepting an award in January 2008, to a -100% and “out of business” by spring of 2008.
Well said, excellent points. Personally, I think the Sharpe Ratio was designed in good faith but has since been “gamed” by asset raisers who artificially depress the denominator with creative accounting and/or substitution of infrequent catastrophic risk for normal volatility. Your point about the hedge fund indices is absolutely correct. With respect to backtesting, optimizing for Sharpe ratios almost always yields sub optimal terminal wealth and (terminal wealth / max drawdown). In other words, you pay a premium, in the form of opportunity cost, for a high Sharpe ratio.
What are the risk free rates of return used to compute the Sharpe Ratios over these time periods? A pet peeve of mine is the fund tracking sites use a constant, in many cases ca. 5%, even in 2009 when it was 0.2% or less. IMO, We have to live with market conditions, so should the those tracking performance. Another point, compare Sharpe or any other measurement on AUM size, the larger the AUM, typically, but not always, the lower the ratios. Just can't move around as easily with bigger size. Lastly, I think Paul Tudor Jones put the whole issue point on when he suggested that a truly good trader produces annual returns twice his draw downs. That I think is the real perspective for an investor to measure not only reward/risk, but also knowing if the risk profile fits his own.
TheRTTrader (Twitter)
it's easy to have those kind of returns when you just pump and dump based on their own alert subscribing newsletter…
it's easy to have those kind of returns when you just pump and dump based on their own alert subscribing newsletter…