bostonbubble.com

GenXer · Joined: 20 Feb 2009 Posts: 703

Some of the greatest 'sucker' rallies happened during Depressions/recessions. What does it mean? Absolutely positively NOTHING. If we try to predict the future by a rising (or falling) stock/price chart, we will always be wrong, unless we happen to guess correctly (a pretty slim chance for that exists). The only thing we know is that macro fundamentals are crap and getting worse if anything. There is really no way it can be spun to prove that the bottom is here. If anything, calling bottoms is near impossible, and amounts to making pretty random guesses about the future movement of the economy.

admin · Posted: Fri Aug 28, 2009 5:00 pm GMT Post subject:

mpr · Joined: 06 Jun 2009 Posts: 344

mpr · Joined: 06 Jun 2009 Posts: 344

BostonObserver · Joined: 09 May 2009 Posts: 8

Regarding the assertions made below, the devil is in the details... check out the YoY sales data in the table found here:

http://seekingalpha.com/article/158901-mixed-housing-data-clearing-up-the-picture

Granted, the data refers to % change in unit sales levels (not price change) but the message is quite interesting nonetheless.

Homes priced under $250,000 seem to be where the turnaround is taking place -- can you even buy a SFH for <$250,000 in Boston?

GenXer · Joined: 20 Feb 2009 Posts: 703

See if this works. At the start of the week, note the Dow's price. Try to predict what it will be in 1,5,10,15,20 days. Do that at the start of every week. See how close you can get. You need to analyze the data right, because there may not be statistically significant difference on the order of days, but there certainly is on the order of weeks and months, and certianly years. Its just that days/weeks are much easier to test and to accumulate statistical data on guesses vs. the actual price. I propose that the longer ahead you try to guess, the more wrong you will be. This is a simple result of a classical Random Walk model. While being totally wrong simply because it does not allow for deviations much larger than several sigma, but even if we take this model for the stock market, (keeping in mind that deviations are much larger, leading to a much larger impact (or a possible future deviation), we can see mathematically that the longer you try to predict ahead, the more wrong you will be.

Such predictions will amount to nothing more than guesses. While for the classical Random Walk model you can bound your prediction by confidence interval, no such thing is possible for the actual stock market, and larger moves simply amplify this effect, so that long term it is impossible to predict the behavior of the market.

Real estate index, while NOT exactly the stock market is more akin to bond market. While bond market is NOT as extreme as the stock market, it is nevertheless extreme enough to ditch the Random Walk model. So in other words, even Treasuries are volatile enough so that any predictions as far as 6 months ahead will miss the mark. Granted, you can get close by chance, but the more times you do it, the more likely you are to be wrong. Yes, you can guess right several times in a row, but that does not come from superior analytical skills. If you practice your prediction skills with any economic quantity which behaves as a random process you will see that your prediction ability based on 'fundamentals' is nothing more than wild guesses, which only appear to be based on 'data'. There is simply too much data, and the economic variables respond very quickly to big changes which are impossible to predict, and which can be analyzed only in the hindsight.

So, my usual suggestion - don't take yourself too seriously trying to predict the future, and don't bet your savings on it (i.e. by trying to call bottoms when we know we do not have the ability to do so without incurring a lot of unnecessary risk).