QUOTE (A.steve @ Jul 1 2008, 03:47 PM)

You will note that it wasn't me who suggested a fixed way of identifying a turn in sentiment -but the moving average trick, to me, sounds like reasonable circumstantial evidence that sentiment may have turned. I don't think it infallible and certainly wouldn't trust that indicator alone - but it does seem relevant to me. We've the phrase "
One swallow doesn't make a summer" - but that doesn't stop us using swallow sightings as corroborating evidence for changing seasons. While you argue the position that speculative risk can be eliminated by making entirely random investments, I think that there are broad factors that influence the market... not dissimilar to seasons. I can't predict if it is going to rain on any particular day - but I would far prefer to leave the storm gear behind when walking in summer than winter.
I think I know why analysts do so badly when trying to predict the market - and it is the same as the reason why I won't back my hunch that builders' share prices will fall with a spread-bet: the time-horizon. Pundits
need to continuously show that they are 'on the ball' - because, essentially, they are salesmen and their objective is to close deals - not to establish an academic high ground over the long term. The existence of leverage undermines the decision they should be making as the "tortoise" investor tends not to reap large rewards... it is far better to be the "hare" and get regular bonus payments for each sprinted section - even if the customer, ultimately, looses.
You say that you think that the computer makes better trading decisions - but, also that you advocate random selection of investments... which I think to be contradictory. If the investments are really random, you don't need a computer - just a cork board, a copy of the FT and a dart. The only reason I can see for automating the decision process is to roll it out to those who don't understand the reasoning behind it... which, if you can pull it off, will be a coup for the earliest user - and a disaster for everyone else... rather like a ponzi scheme.
I think that the elephant in the room today is credit availability. This is the metric I'd most want to measure and understand - I believe that if such a metric could be established and understood, it would identify a clear investment strategy that would yield significant benefits for investors party to it. I'm not sure it exists - and that's probably why it would be effective in the near future. This metric, however, will not be a tool I can pick up - instead, it will constitute all available information... and, maybe, some less available information too.
I'm standing shy of posting this on Wilmott - for several reasons. For one, I'm not a regular reader - and for another, I get the impression that the forum is concerned with operational issues surrounding quantitative specifics - whereas I'm expressing something closer to a belief or a philosophy. I don't have a mathematical framework - so I don't have a hypothesis to be refuted. I do have an anecdotal that I predicted the collapse of the TMT bubble several months before it collapsed (I had no idea when it would go pop - only that it must) and, more recently, I anticipated the stock market falls of 2008 in 2007. This has been quite different to quantitative analysis or technical appraisal of markets - it has been to do with establishing popular fallacies and anticipating how they are skewing the market... metrics and quantitative techniques, I think, only yield clues - the real skill is in putting these clues together - and that, I don't think, can either be done automatically - or entrusted to another who is likely to have contrary ulterior motives.
"I think I know why analysts do so badly when trying to predict the market"
Let me give you a clue - it's because you/they/I can't!
"You say that you think that the computer makes better trading decisions - but, also that you advocate random selection of investments... which I think to be contradictory"
What I say is this. I buy a portfolio of diversified shares (high yield) that I buy and hold forever. I don't try and second guess the market and trade those shares because I can't.
I haven't thought about actively trading for a few years but all this talk about beating the market has got me interested in mechanical investing again. My point about using a computer is that it cannot be swayed by emotions - you and I have discussed the fact that the tech boom occured as as indicator that the market is not efficient and humans are allowing their emotions to get in the way. A computer wouldn't suffer from this.
I think if we are to propose market beating strategies (risk adjusted), we at least need to show some evidence that it is indeed just that (net of trading costs). For example, I think Validea's price/sales investor
http://www.validea.com/home/home.aspoutperforms the S&P. But this example doesn't take into account costs ans doesn't show me risk adjustd returns. I will see if they can give me that.
"If the investments are really random, you don't need a computer - just a cork board, a copy of the FT and a dart."
Or buy the index
"I think that the elephant in the room today is credit availability"
The elephant is happily chatting with the 700lb gorilla
"I'm standing shy of posting this on Wilmott - for several reasons. For one, I'm not a regular reader - and for another, I get the impression that the forum is concerned with operational issues surrounding quantitative specifics - whereas I'm expressing something closer to a belief or a philosophy"
There are all sorts of people there (including authors of books) happy to discuss anything.
"I do have an anecdotal that I predicted the collapse of the TMT bubble several months before it collapsed (I had no idea when it would go pop - only that it must"
But this is the biggest problem. You don't know when the bubble will pop. I thought housing was overvalued in 2002 - look how wrong I was