Hi Matthew,
I paid the subscription fee for my new system, but it is showing a negative amount in the p/l per unit. Please note that this needs to be fixed since I have had more winning trades than losers and my winners are doing much better than my losing trades.
Thanks in advance,
— Brad
Given the math, this is not necessarily an error. I am happy to look, though. But please tell me which system you are talking about.
Brad, this phenomenon has been discussed several times. C2 computes the P/L per unit as the unweighted average of the P/L per unit of the trades. In this case:
- The 20 PALM options were sold at 0.30 and covered at 0.35, so that is $5 loss per unit.
- The 5000 PALM shares were bought at 16.36 and sold at 17.53, so that is $1.17 profit per unit.
The overall P/L per unit is now computed as the average of -$5 and $1.17, which is negative.
The alternative is to compute it as a weighted average, which in this case would be [20 * (-$5) + 5000 * ($1.17)] / [20 + 5000] = $5950 / 5020 units = $1.19 / unit.
I think the last method is better for many purposes and it has been argued several times on the forum.
The two methods will yield identical outcomes only if the number of units is uncorrelated with the P/L per unit of the trade, which is generally not exactly true.
So can I get a weighted ave for my system? The former way seems silly if I use non naked options to boost my roi.
<<The alternative is to compute it as a weighted average, which in this case would be [20 * (-$5) + 5000 * ($1.17)] / [20 + 5000] = $5950 / 5020 units = $1.19 / unit.
I think the last method is better for many purposes and it has been argued several times on the forum.
The two methods will yield identical outcomes only if the number of units is uncorrelated with the P/L per unit of the trade, which is generally not exactly true.>>
I know this had been discussed before and I can see both sides of this. Whilst it is true that an unweighted measure does provide a good comparison across systems it is surely only the case if it is consistently trading the same instrument ie all stocks or all futures.
Where it makes absolutely no sense at all is where you trade multiple asset classes in one portfolio, I have a fund myself that can trade stocks as well as futures, so I can be trading several thousand shares or 1 or 2 futures, clearly this makes any p/l per unit calculation a nonsense unless it is weighted.
Perhaps the answer is to simply provide both weighted and unweighted side by side thereby giving the vendor the benefit of the doubt that having only one measure won’t provide.
Jon
Yes, I saw one of your systems, and it made profit after commissions but the profit per unit was negative $ 20. Therefore the current way C2 does this will not work well to promote good trading systems to possible subscribers.
I think it would be nice either to allow the vender to choose or to have both side by side with maybe a way that the user can be informed of the differences.
— Brad
There is also another problem: Both versions of the P/L per unit fail to indicate what happens if you trade 1 contract on each trade in systems that have multiple legs per trade. If you trade 1 contract per trade then you would typically follow only the first entry and the last exit signal of the trade, and only these signals determine what your profit on the trade is. However, both of the above formulas include the profit of the intermediate signals too. So one would actually need 3 versions of the P/L per unit:
1) the current unweighted version (although I would not mind if it is omitted)
2) the weighted version, which is simply Cumu $ / total # of units
3) the first entry & last exit version, with one unit per signal.
On top of this one could do a breakdown per asset class. One could also cluster trades with the same underlying, but I can imagine that it is a hell of a job to program that.
Even within one asset class there can be huge differences in price (e.g. $0.50/sh or $500/sh). Another interesting measure would be the average profit per invested dollar (which has also been suggested previously).
I understand where you are coming from, my profit per trade on C2 has been in the "negative" from day 1…while my "profit factor" has been very good…
obviously my profit per trade does not avg a negative figure, or my equity curve, total return, etc would be below par…
It would be "nice" for C2 to fix it… but for now I just live with it(I also have other "gripes" about C2 as well, but hey, overall it is the best web site out there).
I have had subscribers regulary subscribe, so when they looked at my stats, they could see the profit per unit figure was in error.
Good luck on getting this fixed, if they fix yours, please send me a private message so I can work on getting my fixed as well.
-David
