if you really want to improve it, just buy the ES and hold. there is no way to beat that with a coin toss or other RNG process.
What you suggest is a non-trading method. i.e.: The trivial solution. You are absolutely correct: Any non-trader will outperform 95% of all traders and also will outperform this trading METHOD as is. The intention of the thread is to make this TRADING METHOD better however. Not to point out the trivial solution to the problem of trading. Your comment is very good in that 95% of people "trading" would be better off if they never "traded" in their life.
I want to add that improving the entry or the exit by 1 tic will make the method breakeven over the long run, including commissions, and improving the entry or the exit or both by 2 tics will make the method a winner. Why is the simple and obvious always the most overlooked?
trading non linear instruments such as options might be the pathway your looking for. the RNG process whatever you chose might offer no advantage but the combination of an unbiased direction decision process coupled with the non-linear nature of options could alter the rules of game. i favour short gamma positions but the opposite might be even better, depending on what you want to set up.
Let's keep the discussion centered in ES futures as the instrument for now. But if you start a thread with options as instrument, you can link it here.
What about keeping your entry methodology the same but putting in exit levels. You could close positions that are either down two points or up six ... or whatever ratio you want to test.
Just find 20 tossers, plenty in London, let them trade, monitor the net position and take the opposite side to the group. bingo!
What if the homeless guy, in between coin-tossing, got some newspapers from the trash and cleaned passing cars' windows? He could harass people to give him change, thereby covering your commission cost and bringing the system closer to b/e.