$50k - I'd reserve 30% ($15k) for a common cash reserve, and start 7 LD-AIM programs with the remaining $35k.
Each would would cover 5 Sells (0 Safe / 10% Minimum) and have a beginning Portfolio Control of 12,240. Those 5 Sells allow the price to appreciate 69% before selling out (assuming no Buys), and would provide a minimum ROCAR of 35%.
Your Actual $ Buy for each program would be $5k. The Virtual:Actual ratio would be 1.45:1. So for a $10 stock you actually buy 500 shares and carry 725 virtual shares.
I would also make sure that the 7 stocks are not highly correlated. This protects your shares cash over normal times and lowers overall risk. I've learned from experience that one sector goes up and another goes down. It happens all the time with me.
Another alternative might be to perhaps allocate 36K stock, 14K cash, diversify across perhaps stocks (IWRD), commodities (GSG) and REIT's (VNQ) and run 2 LD accounts each. One with a narrow share price range before all-in or all-out, along the lines of Steve's choices, another with a wider range.
The wider LD in this example has you out of the stock after the price near doubles, and all-in after the price halves.
Less so for the buy side, but for the sell side there are quite a few close sell price trigger levels in the two LD's which could be averaged into a single combined trade in practice to keep trading costs down.
The narrower LD would have you capturing price volatility around the current price levels whilst the wider LD would help ensure you didn't exhaust cash (or stock) too soon.
Market Data 
Markets 
AI
