Yes, I'm still long MXP, albeit with only 1 contract. The maximum position size in my portfolio is 3 contracts if the forecast were at its maximum value.
Thanks! That sounds about what I have as well, my max position is 5 though. I guess we just hang in there.
I'm long 7 lots, so you lot are pussies Not sure what my max is since that piece of legacy code has broken (nice way of encouraging me to work on pysystemtrade more quickly), but I was long 12 and I sold 5 contracts on Monday. GAT
Is there any way to improve the quality of a momentum signal for an instrument by using more variables? For example gold and equities combined signal.
Multi variables can be simple if you stick to a linear framework. Another variable is like another rule.
Can you please transfer some money to my trading account such that I can hold a similarly large position? 500k USD would be fine, thank you very much.
Well right now the sharpe ratio is sort of the prediction quality. There are other ways to measure, ie you can set it up as a linear regression problem and come up with a measure of the return x days for each signal you have generated. X days will depend on your trading speed. Then you can look at Rsquare
The current way how @globalarbtrader and others on this thread implement momentum is by looking at the Donchian channel and calculating the position of the current price within this channel. That is used as momentum signal. This is repeated for various lookback periods and a weighted average is calculated. I'm not sure how Sharpe ratio could be added to this calculation.