From Niederhoffer's wiki entry: "Niederhoffer Investments returned 35% a year from inception through 1996, when MAR ranked it the No. 1 hedge fund manager in the world. In 1997, Victor published a New York Times bestselling book, The Education of a Speculator. In statistical terms, I figure I have traded about 2 million contracts, with an average profit of $70 per contract (after slippage of perhaps $20). This average is approximately 700 standard deviations away from randomness.[5]" So, my question is why would anyone expect to make more than $70 per contract per trade (after slippage), if by doing so, with scale, you could be on par with the world's #1 hedge fund manager? Not that VN reached that level only by trading futures, but this is what he says he made in futures trading specifically. I'm working on something that shows a profit of over $100 per contract per trade, before slippage in both 6E and CL, with over 500 trades per year combined and I'm wondering to myself what is the point of trying to make the average trade profit larger when doing so only increases the probability that I am over-optimizing and, given what VN says, trying to make more than that is unnecessary, not to mention highly unlikely, in the long run?
Interesting, trades with about $70 potential are what I look for [NQ]. I chose that as a minimum after looking at the tradeoff between frequency of trades [I don't want to be all that busy if I can avoid it] and overall gain for the day. Increasing the gain per trade is not the path to greater profits. Increasing the frequency of trades while taking smaller gains increases profits right up to the point where it runs up against the cost per trade.
There is absolutely a tradeoff between frequency and size of average gain. I always go back to the Niederhoffer stat because I think that with all his resources, if he could not find something sufficiently compelling in the futures market to make more than $70/contract per trade, then there must be very few opportunities of that magnitude in the markets period and the ones that exist must have a lower frequency, so that if you were to try to make $140/contract per trade, you'd have only half (or fewer) as many opportunities as the $70/contract per trade strategy, as well as increasing the risk of over-optimization.
This whole topic of "realistic per trade profitability" is bullshit for 2 reasons. 1) In logic_mans quote it is pressumed Neiderhoffer "knew" which trades would produce average gains and which trades would produce outsized gains. Nevermind that (consistent) 100% accuracy does not exist in trading. While I agree there is a relationship between frequency and average gain, I would argue that frequency(number of trades means time spent in market aka duration)=risk, risk=profit OR loss potential. Therefore frequency(duration) is an opportunistic necessity for producing an above average number of outsized gains. 2) Overall, this "realistic profit" is no different and just as INVALID as using a win-loss ratio. You take 12 trades. 10 of them are $70 gainers. 2 of them are $350 losers. you have a win-loss ratio of 5:1 with average gain of $70 per contract... it has no meaning as to the profitability(neither before or after taxes) or the abilities of the trading system or the trader. Trade On!
"Frequency" as it has been used here is confusing. There are two ways to increase the number of contracts traded: increase the number of setups traded, or increase the trade size. The former is mostly determined by market conditions, what the market gives us. The latter is a function of liquidity, capital, and leverage. If market liquidity is good enough, we can trade more contracts (and make more profit) by using more capital or more leverage. As for the profit per contract traded, that is a mixture of trader skill and luck.
I don't see anywhere in the quote where VN says he knew which trades would be winners or losers. Obviously, the ways in which a trader could arrive at that $70 figure varies, but the question I am posing is what is a realistic expectation for an individual trade for someone who trades with a frequency approximately equal to VN during that time. $70 might be out of reach for someone trading even shorter-term, who might be happy with $10, while it would be poor performance for someone who holds longer-term and expects $1,000. The question would then be to what extent there is a linear relationship between time in the trade and a realistic expected profit, so that if the trade for $1,000 has to take ($1,000/$70) or ~14 times as long as VN's typical $70/contract trade. So, if I knew VN's average trade duration (let's say it's 10 hours), then I could have asked, "Is it reasonable to expect profits per hour of greater than $7 per contract per trade, since that was the profit-per-hour for a guy considered one of the best traders of all time?". I agree, that would be a better way of stating it.
Right, but increasing the number of contracts traded won't alter the profit per contract. I do think that the profit per minute of time spent in the market is important, though, and that should have been a part of my initial question about what is realistic to expect.
None of it matters because 1) gross and/or net is not a factor for win/loss ratio. 2) the point has been made "realistic per trade profitability" is a useless statistic, just like win-loss ratio. Trade On!