That's like the cab driver "day traders" in the 90's. Every day call the broker to buy at open and sell at close. They must have felt like effing geniuses.
I'm not surprised. The C2 model is inherently flawed: almost nobody with real trading skill or a genuine, durable edge will give it away for a few bucks when they could trade it themselves or leverage outside capital through a performance-fee model. The "systems" on there are therefore going to be either unproven/new/aspiring traders who will mostly bust out, or systems which have done well but the creators don't consider suitable for their own capital or OPM, due to poor risk profile, low or negative edge, etc.
Why would you do that? You have market prices, how is a closed position different from an open position?
Since some commodities/futures I stay into for years and having to do rollovers into different contract months which are seldom at same prices, one quarter can show nice profit in say coffee of $5k, then rollover happens and next quarter shows loss and being stopped out of loss of $5k. So based on monthly stats and quarterly up 5k and down 5k as profit made then lost and drawdown, but in total equity of the overall position of coffee it is a no gain/loss trade. Since goal is to capture 75% of range of nine year cycle, the above example does happen. What it points out to me is studying more of what is happening in chart patterns that can alert me to re-apply hedges when losing, when gaining to dance more options around position. Once a position gets to breakeven plus fees, protective stops never move with exception of rollover money distance to new protective stops. All my testing shows for huge profits to happen, stops do not move, but there are times where it will lose huge gains cause it didn't make the 75% of the move, but overall, this method has done nicely for me since late '91.
Maybe I am confused, but a roll is a pnl-neutral (mostly) event. Are you looking at synthetic futures return series or you actually are looking at something else?