Which system is better?

Which system is better?

  • 60% Win-rate/1:1 RRR

    Votes: 12 54.5%

  • 40% Win-rate/2:1 RRR

    Votes: 10 45.5%
  • Total voters
    22
Handle123 said:
I pretty much scalp Indexes, currencies and energies, and I average down on each trade, so my losing percentages have to be extremely low to do so. What is not specified in two methods are the "mean" drawdown is, meaning what the average expected drawdown to be. And also what the breakeven percentages are, so if system is 40% winners, what is the 60%, are they all losers or mixture of losers and breakeven trades. Of course all may have different definition of what a winner is as well.

You can have a very good method where there's 30% winners of 1:1, 50% breakevens and 20% losers and clean up by averaging down on each trade. On breakevens you could be adding at every two ticks against position and exiting them at original entry tick plus one tick, so if trade when against the position seven ticks, could have added on three trades, so breakeven trade gives +7, +5, +3, +1 . Of course the Math goes into hyper-drive of "what ifs" trading this way, and backtesting has to be deep and lengthy to provide a better edge.
More...
I heard guys who average down blow up sooner or later.( LTCM, Barings Brothers and Sumitomo Copper.)
 
Expected number of consecutive losses = LN(Number of trades)/-(LN((1 - Win Rate)))

For a 60% win rate you would use 0.60 in this equation.
 
cornholio said:
Is there a formula that can figure out what the maximum number of consecutive losses your system can have?
More...

The number of consecutive losses is of no use if your system has variable-size wins & losses, because that one trade that breaks the losing streak might just be a BE or win 1-tick, where your average loser in that streak is 4, 10, 20 or even 100-ticks.

What you need is to estimate your max. drawdown with a given confidence level. Look into using a MonteCarlo simulation, using the trade distribution out of your backtest (or live trading results). Use at least 1-year worth of trading for each iteration of the MC sim. Then use the average & stdev of the max DD of these runs to define your max.DD with a given confidence level.
 
I ran MC analyses on sample sizes of 100 trades. In order to do that, I had to generate trade lists. In so doing, I had to make assumptions about the standard deviation of the trades since the information isn't provided. I suspect the OP intended for the SD of both systems to be the same. However, the SD is an important characteristic of the trades and would play a role in a MC analysis which I believe is the way to determine which system is preferred.
 
dom993 said:
The number of consecutive losses is of no use if your system has variable-size wins & losses, because that one trade that breaks the losing streak might just be a BE or win 1-tick, where your average loser in that streak is 4, 10, 20 or even 100-ticks.
More...
Really?(rhetorical) That's good and makes sense I guess. Because It doesn't make sense for you to get 10 losses in a row if you have an edge. Do you agree? If your strategy is based on flipping a coin then yeah I think you can get 10 heads/tails in a row easy. Idk... I didn't go to MIT.;-)

What you need is to estimate your max. drawdown with a given confidence level. Look into using a MonteCarlo simulation, using the trade distribution out of your backtest (or live trading results). Use at least 1-year worth of trading for each iteration of the MC sim. Then use the average & stdev of the max DD of these runs to define your max.DD with a given confidence level.
More...
Thanks man. I heard some other guy talking about this monte carlo thing but I didnt know what the hell it was about. Will definitely look into now. Cheers!
 
dom993 said:
What you need is to estimate your max. drawdown with a given confidence level. Look into using a MonteCarlo simulation, using the trade distribution out of your backtest (or live trading results). Use at least 1-year worth of trading for each iteration of the MC sim. Then use the average & stdev of the max DD of these runs to define your max.DD with a given confidence level.
More...

Well from the looks of it, you look like your a real professional trader so I want to ask you: Trading is just a game like Black Jack?

(Lol, I just learned Monte Carlo is a casino.)
 
Cornholio, no matter what else, keep your wallet in your pants. Preferably a zippered pocket. Or, better still, leave it at home.
 
cornholio said:
It doesn't make sense for you to get 10 losses in a row if you have an edge. Do you agree?
More...

No, I don't agree. Edge or not, regardless of the win% your trading system has, it will eventually get 10 losses in a row (if anything, it will get there the day your edge is gone, but even though the edge might not be long-term gone, it might be temporarily away).

If you don't already know this (attributed to Keynes) : "Markets can remain irrational longer than you can remain solvent. "
 
You must log in or register to reply here.
Back
Top