It's in this thread: http://steadyoptions...-thru-earnings/
Did simulated trades with either puts or calls, but I lost money in the long run due to IV collapse. I've concluded that getting the direction right does not matter nearly as much as managing the potential profits/losses. In that thread, Jig mentioned using bear call and bull put spreads. I think that bull call and bear put spreads might work even better. You structure the trade so that the potential profits are at least 1.25 times the size of potential losses. If you get it right, hold until the max profit has been reached. If you get it wrong, get out immediately. Obviously you will need a little extra movement in the right direction when the max profit is larger than max loss, but if historical moves have exceeded the required move for max profit in at least 9 of the last 10 cycles, then it should be worth doing. With 50% accuracy, this strategy could be profitable. Options should also be very liquid (< 5% difference in bid/ask spread).
I've only done this on 3 companies so far: HPQ (+98%), CRM (-79%), ADSK (+84%). Our IV earnings candidates are also the best candidates for this strategy.