The formula for positive expectation is: [1+(W/L)]*P - 1 meaning W/L = pay-off = average winner / average loser meaning P = hit-ratio = number of winners / total number of trades Swish please be aware that Sharpe Ratio is in most cases defined as: (Return p.a. of System or Trader - RiskFreeRateOfReturn) / Standard Deviation of Returns The modified Sharpe doesn't subtract the risk free rate of return from the traders yearly return. The idea is that the risk free rate of return differs from investor (institutional for example) to investor (retail investor)... Regards, Oliver
Oliver, Thanks for the post. What does standard deviation of returns refer to in your SR equation? Thanks, Swish
Hi Swish, the definition of Standard Deviation is: square root of the variance. That is, the standard deviation is the average squared deviation (variance) returned to standardized format You can access the formula for standard deviation for example in Excel. The StdDev gives you (in a normal distribution) the range for +/- 66% of all realizations of a random draw. For example if the yearly average return is 10% and the standard deviation is 5% than 66% of all years will show returns between 5% and 15%. Good trading, Olli
One thing to also keep in mind, though, is that the use of SD assumes and relies on the fact that distributions are normal, but it has been shown for many years that market distributions are not, that they are kurtotic (fat tails), meaning that extreme events occur much more frequently than would be accounted for by statistics assuming "normality" of distributions. Lots of arb firms adjust their models accordingly, and make their living off of this discrepancy. It is also important to remember that inferential statistics are just that, inferential, not absolute. There is a big difference (hopefully) between the risks you would take based on what you expect (or hope) that the market "should" do versus what you assume the market "will" do when you are talking about interpretation and prediction. It's that "will do" assumption that leads many traders to take "statistically based" risks that in fact, because of the kurtotic distribution of prices and returns, aren't really accurately statistically based, resulting in risk taking that is far greater than the trader assumes. I know that this discussion is asking about rates of return rather than market price distribution, but the same effect (and loss of accuracy) resulting from kurtosis occurs there as well. Good Trading, GO CUBS, and CONGRATULATIONS BADGERS!!!!! Jessie
1. Why would you need one formula to analyze a fund or a system? I don't think it's a good idea to formulate it all into one. All the numbers you can crunch has its own significance. 2. Balance is the key. You trade the way you feel comfortable.
Hi WDGann, if balance and perspective are needed - which formulas or views would you suggest to measure the performance of funds? Of course any individual utility curve would limit the comparability for others. Most CEOs nowadays seem to measure their success in terms of money being paid. Given the added transparency over the last 10 years, the race is still on
Numbers don't tell you anything by themselves. You could have someone up 200% per month because he's betting his entire fund on selling deep OTM options all the time. He may continue and then lose it all in one day. Whereas a guy bleeding 2% per month might be doing the opposite strategy and have a long-run expectation way ahead of everyone else. Even someone with a decade long track record of 25-40% a year could be a terrible person to invest with - look at LTCM or Victor Niederhoffer. My personal criteria for considering investing would be as follows: i) The manager must have at least 75% of their net worth in the fund. This shows that they are confident enough in their method to back it with cash. If someone is seriously wealthy ($50 mill+) then I might relax that to 50% of net worth in the fund. ii) At least 5% of the funds assets is the manager's money. This is to show that the person is not playing with an amount of money *way* in excess of what they are used to. If someone is good enough to manage $100 million, they should have made at least $5 million out of the markets, and kept it. This will exclude some promising new guys, but I'd prefer to let them have a good year or two then come back when they've made some dough. New guys should start small until they are proven. iii) Do you trust the person? Would they treat you fairly when the chips are down, even if that meant endangering their own position? If you would invest with them purely on a gentleman's handshake, then that's a good sign. If you find yourself wanting everything legally guaranteed in triplicate, then walk away. iv) Is the person a blowup risk? You want to see the manager show serious respect for the market's ability to wallop them. Someone who is too confident, who goes on about how "easy" it is, or how their models/system/experience has it all figured out, will most likely blow up sooner or later. A great example is Victor Niederhoffer - you could literally tell he was going to blow up just by reading "Education of a Speculator". This is especially important if they have a very good performance record over the last few years. v) They must show a deep understanding of the markets, based on actual trading experience of at least 5 years, preferably more. If someone understands the markets, then even if they are currently in a slow streak, they will eventually figure a way to outperform. Whereas someone who has simply found one method that works right now, but has no deep understanding, will be in trouble once their method stops working. vi) Are they a "monkey on a typewriter"? If someone is raising money, saying they "bet on the markets always reverting back to the mean", then you know they have no edge at all and have just been lucky so far. Same with the astrologer who has made 100% per year for the last 3 years, or the short-seller who raked it in from 2000-2002 but almost went bust in the late 90s.
Cutten- Excellent list. I would add that small funds tend to outperform the large funds, and guys just in their first few years usually beat those who have done it a while and are already rich. Basically, you want to invest with guys who are poor, want to be filthy rich and know this is their one and only shot. They will be willing to put in 120 hours a week to read everything. Once someone is rich, they don't care, and instead just want to preserve capital and grow it slower.
Hi Cutten, you are perfectly right that many managers, however excellent their track record, could happen to run into a "peso"-risk, losing huge amounts of money in a very short timewindow. I disagree with you concerning the personal investment a) nearly 50% of ltcm assets under management was personal money of the partners (highly leveraged through loans) - it didn't help in terms of making the fund more secure. b) any manager who has 75% of his net worth in his own fund doesn't apply the principal of spreading risk for himself. Already 100% of his yearly income depends on the funds performance and now he is risking 75% of his own assets? For me that doesn't reassure me, it tells me he is taking high risks. c) you relax the 75% investment when he owns more than 50mn - why? We all know that the more money one has, the more risk-averse (normally) he becomes - but is it sensible to let him reduce risk when he's already rich? Good trading, Oliver