Let's assume using a system with a P/F of 2, at 50% win-rate, for 35-ticks win / 17-ticks loss. Expectancy is 35/2-17/2 = 9-ticks per contract per trade (ignoring commissions & slippage).
Using $10 as tick-value, it takes trading a total of 47 contracts per day to reach the goal (47*90 = 4230). Given that we are limited to 4 cars / trade in this 35-ticks target category, this is going to take 12 trades per day, either 3 trades x 4 instruments, or 4 trades x 3 instruments.
When using 4 instruments, and no overlapping trades in the same instrument, we'll have a max. number of 16 contracts in open positions at once, which translates into finding a broker allowing for RTH margin of $1000 per contract or less to stay withing the 25K limit of the account (on a very bad day, 100% losses, the total loss for the day would be 17-ticks * $10 * 48 = -8,160, which on a 25K account is compatible with having 16 contracts open using 1K margin or less per contract - but that requires reloading the account overnight to keep trading beyond the 1st trade the next day).
The main challenge remains in designing the trading system that will meet or exceed the performance parameters assumed in the 1st paragraph, for 240 (days/year) * 3 (trades/day) * 4 (instruments) = 2880 trades per year.