|
发表于 2010-6-4 04:55 PM
|
显示全部楼层
This Monday market was closed so don't count as red.
I still doubt the way you forecast whether green or red the next day. I'm still thinking you are using data from a huge rally (09/03-10/04) as your main data base. Frankly speaking, even if you use all years' data, it's not a good way: the market goes up as a whole, 70+% time bull, 30-% time bear, so if using all data, probably the result is still bullish. One way to get a more neutral result, is to use data half from bull market, half from bear market. But, even in that way, we still get quite wrong results, probably. Why? Because market is so complex that it's not like rolling dice, and a single day's move is more random than one week's move.
I'm interested in 方法论 on stock study, and I think we should think if a method works then use it. Thanks. |
|