Hi Allen, Re: Minimum orders vs AIM's proportional order sizes.....
I think I understand your thinking here. This could be done with a combination of minimum order trades and delayed trades. One would always do the minimum order trade when it occurs. Afterward, any sequential 'same direction' trade could be delayed either in time or price percent change. If it turns out to be a trend, then delaying 30 days or an additional 10% move would generally shoot well past the minimum order size.
AIM is a proportional control algorithm. The larger the divergence from Portfolio Control (PC) value the stock value is, the larger the order size will be. I love making trades at the "minimum order size" when they occur, especially after a long period of no activity. However, when the market makes a dramatic move, it's also fun to make a far larger trade based upon AIM's proportionality.
The main downside of doing only minimum order sizes whenever they are signaled is the addition of more trading cost. There could also, in a taxable account, be more transactions to report at tax time. The upside is that with minimum order sizes, even in the same direction sequentially, at least a small order will usually be traded near the price reversal extremes.
For AIM both histories should turn out about the same over time. Selling minimum amounts whenever they are signaled moves the "next buy" bar up with each transaction. That in and of itself is a good thing in that AIM makes money by recycling its cash. If one stores up a bunch of minimum order sizes and delays by some mechanism (your choice) and then executes a larger single order, AIM will move the "next buy" order price up accordingly. Either way AIM will take care of the situation.
The main issue with frequent minimum orders being trading cost is somewhat reduced when we make sure the minimum order size pushes the trading expense down to a level that is acceptable to the business plan. If 1% is acceptable and commission is $10, then your minimum trade needs to be $1000. If that $10 commission needs to be just 0.5% of transaction cost, then you're going to need a minimum trade of $2000.
Once the minimum trade size is determined by transaction cost percent, it hardly matters if the order is executed while the trend is continuing or when it has reversed. Technical Analysis (TA) is fun and interesting, but so is cash or shares in hand. I don't mind glancing at TA signals to get confirmation of AIM's intentions, but hate to over-ride AIM's hard work. I'm more inclined to follow the "set and forget" Good 'til Cancelled Limit Orders and go find something else to do with my time.
AIM is already efficient from a TA point of view. It almost always is selling above the 26 week moving average and buying below it. We can't ask for much more. But since it's not an all-or-nothing trading algorithm it can't capture all the profit or all the discount. Again, the term "proportional" comes to mind.
When one fills out the Capital Gain Schedule D section of the tax forms each year, one begins to see what AIM does from a very pragmatic vantage point. When washing and drying a car, it's good to wring out the towel once in a while. We don't get the towel all the way dry, but we improve its ability to pick up more water.
