Very interesting. A few pertinent points from Kaufman;
1) He advocates a raw stochastic as the entry indicator, as there is less lag than RSI or MACD. The long form formula (don't want to mess with subscripts) is
Stochastic = (Today's close - min last n days) / (max last n days - min last n days)
He tests look back periods (the n days) up to 20 days, so fairly short term.
2) Entry based on stochastic difference. He tests +/- 40 to +/- 60
3) for adjusting position size he advocates using Average True Range as the measure of volatility. His logic is it factors in highs and lows of the day, as opposed to just the closing price. For someone like me who desperately wants to get away from the screen, I'm not sure how good that would be.
What I like so far is he takes you on a journey of exploration, rather than saying this is how to do it. At the end you have clear ideas of what should be factored in, how to do so and test / measure the results. In other words he gives you the tools to design your own system to suit your needs. Perfect for me, because I can't stand that How to Make a Million Bucks in 5 Minutes a Day garbage.