I know that One Size does not always fit every fund. In another life with other software I was able to Performance Tune each one. Because Mutual Funds are like people, were all different. So does anyone Performance Tune Funds with the Configuration available in Automatic Investor.
Hi, Jim,
You've hit on one of the interesting philosophical areas regarding AIM. In the book, Lichello advocated dealing with one's entire portfolio as a single entity, remember a secondary goal of his being to keep things simple in the pre-personal computer age. One may find, however, as your post alludes, that AIM may work more effectively when each holding is run individually within it's own program. The whole portfolio approach will allow you to have a single cash reserve for the whole lot, giving you discretion as to what will be bought or sold whenever AIM gives a recommendation. AIMing each individually will require multiple cash reserves and so on. Lichello also wondered how this would play out if one holding keeps sucking money for more purchases, whilst others may be flush with cash. These and other arguments can be made for the pros and cons of each approach. Each may have adherents, though I think the dominant view is to run each separately.
In terms of adapting AI configurations on an individual basis, one tool that gives a good visual is the stockcharts.com zigzag function. As I recently posted for Jet Blue, the "standard" 10% buy & sell SAFE + 5% min trans on each side correlates to the 30% tracement of the zig-zag. 10+10+5+5=30. Thus a trip from a bottom to top would give a 30% gain on a LIFO basis. Not too shabby! It's a question of finding a balance where you keep your costs reasonable, get a reasonable number of trades and each trade being worth the trip. Using Jet Blue again, let me post some variants:
1) Our standard 30%:
2) Now we up the zig-zag to 40% which matches the AIM-HI of 10% for each of the values:
With these, not too different from the 30% so the larger range may be suitable for this particular stock.
3) dropping to a 15%:
Here we start to see more trades for less gain per trade. Worth doing? This is where AI's historical tester can help, but take with grain of salt as it's very easy to start changing the configs to find the "optimum" setting - but this is what would of worked in the past and may be absolutely horrid for the future. Also the results are predicated on the end-of-day prices, not the actual fill prices that you would have likely gotten.
So between the stockcharts and the historical testing we can get some idea of the "personality" of a holding. Then adjust the model for that holding accordingly.
With all the talk of optimizing AIM paramerters to an individual stock I say just look at a five year chart of C (shittybank) or anything else you can think of that had a big price move either up or down (how about APPLE?)
If you set narrow paramerters because the stock looks fairly stable and you want to catch small moves you will run out of cash much quicker and not be able to follow the stock down and take advantage of lower prices.
If the stock has a big move up you will be selling stock faster than you need to.
The "standard" settings are for average conditions that allow you to take advantage of the bigger cycles that happen over a few years time and not run out of cash to invest before the bottom or sell the stock to soon.
If you set up a bigger hold zone on a very volatile stock and it quiets down then you will have less trades than you expected.
Just my thoughts tinker on and ignore me because I am Toofuzzy
News 
Market Data 
Discover 
