|
|
发表于 2007-3-23 20:52:20
|
显示全部楼层
摘自MONEYBEE的论坛
rudel 05.04.2005 11:55 PM
Dear all
Let's imagine that moneybee achieves a high rate of consistent predictions on the various markets' moves.
If that would ever happen, I would imagine that word will spread out, and more and more people would follow moneybee's predictions.
In that case, would'nt moneybee become a self-fulfilling prophecy?
Ie, if moneybee predicts that some stock price will rise, and all 'moneybeers' follow that prediction, most probably that stock price will rise accordingly.
If predictions are for a stock to decrease its value, 'moneybeers' will sell or short that stock, and of course, its price will drop...
Either way, predictions will be magnified in their magnitude. I'm afraid that those movements will invalidate moneybee's market models, and any predictions out of it will become rubbish.
In my opinion, markets models only work good, when and if, there aren't lots of people following their predictions.
So, I'm a little skeptic on this moneybee concept. I was told that markets are zero-sum processes, and if one wins, another must lose...
Or, silly idea, are we lending our computer processors so i42 can assemble enough computer power to build models for their exclusive use?
Any toughts?
BabyJoe666 29.04.2005 03:54 PM
No, this is not possible, your creating a feedback situation where the input is outnumbering the output.
So, for a while, it might be correct, but MoneyBee does not know that all the people are following what is does, so then the feedback starts, and suddenly the prediction quality will be off charts(negative)
简单的来说, rudel提出moneybee如果预测得准，那么将形成自反馈系统。对这一点我也同意，在投资市场，效率必然导致失效。所以rudel认为这将导致moneybee的预测模型退化，预测质量下降。但是两年过去了，好象没有发生这样的现象（当然，我们不知道从2002年以来所有的moneybee预测数据，也就无从知道moneybee的演进过程，PS：i42作为组织者似乎是一个隐士。）
从 BabyJoe666的跟帖来看，他/她好象是不同意这个看法的，但是最后几句没看懂，谁给翻译一下？ |
|
发表于 2007-3-22 12:32:30
