2017年5月3日

I Am Reading


日期:2017/04/30

I am reading Walter Heiby's Stock Market Profits Through Dynamic Synthesis, before I prepare for my challenge match with J. Weisenthal at Central Park handball courts today and it's beautiful with 350 pages of tables and 65 graphs. It contains daily data on S&P, advancing, declining, unchanged, ny volume, active issues volume, active issues price, odd lot purchases, sales, odd lot short sales. Monthly data of mutual fund, purchases, monthly debit and credit. It's an admirable effort, and he thanks his father, his mother, and wife for taking an active part in entering all the data and giving him energy for the 40 years at least he's worked on it, (thereby falling short of Larry and the signer's 65 years). It brings to mind the panoply of eminents in Cradles of Eminence: Childhoods of More Than 700 Famous Men and Women: "My wife Eunice Heiby pushed the buttons of our first adding machine, my father who gathered many of the stats which were fed into our clrculators, he also helped perform many of the early calculations that became the basis of the growing body of knowledge that is now known as dynamic sythesis." His "mother, Laura that I owe whichever of these traits of tremendous energy and perserverences". He is very much like Galton on eminence.



Heiby's numbers and ideas are very valuable. They hold up very well except for the odd lot data. The monthly data are so sparse relative to market movements that any sort of tweaking will yield any conclusions so not as useful as the daily and the ideas. He like almost all his followers has no idea of multiple comparisons and the amazing ability of a few cuts of random data to come up with seemingly incredible results. If you cut 10 times from a normal distribution randomly you are likely to find a 1 in 1000 probability emerging, with exact figures given by the exponential distribution and best obtained today by simulation. So all his conclusions about when to get in and out are worthless and couldn't beat buy and hold but it's a noble effort, deserving to be in the valhalla of our field and aside from the high correlation of advancing and declining issues with price moves itself, the daily data is a nice template for further research.

沒有留言: