Obviously there is black swan risk when you increase size which he did not hedge against. That being said. I could see $ 70/contract net profit in any futures you wanted to trade. Also remember the more contracts you trade, the lower your cost will be. For example you could lease a seat. Finally, if you have a profitable system, you can then get funded by larger investors so there is no need for home run trades, instead you just need to out perform the market and not have big draw downs. So, in futures markets, what we can call black swan events are holding through known news events, or holding when unexpected news hits the markets.
Here's a novel thought... Why not determine what is "realistic" based on YOUR OWN TRADING! It doesn't matter what VN or ANYONE was or is capable of. So you come up with a "realistic" (whatever the hell that means to you) per contract number ACHIEVED by someone else. Quick anecdote... On Sun, Feb 19, 2012, French cycling enthusiast Robert Marchand set the worlds record for longest distance cycled in one hour, riding 24.25 km (15.1 miles). Interestingly, the record does not even come close to nor threaten the 48.7 km (30.8 miles) record achieved by Czech cyclist Ondrej Sosenka in 2005. The Czech was 29 at the time of his record. Marchand was over 100 at the time of his record and is credited in the over 100 year old category. Additionally, the Union Cycliste Internationale (UCI) created an entirely new record category! Who the fuck cares what some other trader was or is capable of!! It's illogical_man. Trade On!
I am surprised at some guys dismissing this question. I think it is always good to benchmark oneself, irrespective of the kind of success one achieves. Per trade profitability data over a long period of time is very scarce to find, especially for people who trade big size. If one works in an institution, probably one can get an estimate of one's performance against other traders. But for independent traders, even if you have your personal trading history for a few yrs, its always a good idea to benchmark oneself against another. For people who deride VN's achievements, they underestimate to a large extent hindsight playing a role of 20/20. His rise and fall is not too different conceptually from rise and fall of great ancient civilizations - its just that he saw highs and lows of his life over 20yrs, whereas great ancient civilizations saw highs and lows over say 500 years. Anyways to answer OP's question, I think scoring 70$ per contract average profit over 16 years is a fantastic record by any standard, especially considering the fact that he was trading a large number of contracts.
Well, benchmarking is fine if you have proper verifiable metrics for which to measure yourself against. And you won't find that asking on ET or comparing yourself to what a HF manager says or the media somehow implies that he or she does. And that is the point I was trying to make. So, you have to test your strategy for yourself using different holding timeframes. And that is the very best benchmarking you can do.
This post refers to the $70 per trade profit he estimated, and that probably is what he has done. Around my statistical analysis the portfolio simulations are between $60 and $200 with 50%-60% win percentages. This article particlarly is confusing winks and anybody who doesn't follow that this isn't a discussion about w/l, is that they are discussing what the "net profit per trade" has been, and in terms of real results, this is the system's expectancy. $70 makes him the best, huh? As far as I've come along with my trading, the systems I use are robots to me. Turn them on, after proper parametrization, and wait a few months and see. Your robot is your style, and when it comes to trading, planning the strategy must have at least this much expectancy before slippage, and after trading for a long time the average profit came to $19 per trade in reality so this is a third of his performance....right.
The relationship is unlikely to be pure linear; as I understand it, common quant finance models of price evolution rely on the sum of a linear term (drift x time) and a sqrt (time) term (volatility x sqrt(time)).
While there is some relativity in the idea of benchmarking and it's not an apples to apples comparison, the reason you want to do it is so that you don't set unrealistic expectations or overoptimize your system just to jack up the per trade results. Would you tell someone who was a professional baseball player, "Sure you can hit 100 home runs in a season. It doesn't matter that Barry Bonds, on steroids, only hit 70. Don't limit yourself to what Barry Bonds could do."?
Well, when I posted the thread, I was hoping some of the better posters would chime in with their own experience. There's been some food for thought provided and that's about all I expected could happen. I happen to think the question of capital productivity is paramount. If your capital can't earn a return above its opportunity cost, you may as well be doing something else with it, after all.
Would that imply that if your per-trade returns should show an increase correlated to the square root of the amount of time in the trade? Interesting correlation to check.