Predictive Model Output - Mar 23, 2007
Last week, the models were all bullish, but I wasn’t. In (additional) punishment, I will be quiet and let the models do all the talking. I should trust them more than my own judgement.
The S&P 500 closed at 1436.11.
The five-day “potential” model score is 80 on a scale of 1 to 100. A good even-odds score for next Friday would be 1441. Odds are 75% that the index is over 1422 at next Friday’s close. I do not model for a minimum or low aka “safety” measure over five days.
The twenty-day “potential” model score is 69 on a 1 to 100 scale. A good even-odds score for the close four Fridays from now would be 1454.
The twenty-day “safety” model score is 95 on a 1 to 100 scale. A good even-odds low point or minimum for the next four weeks would be 1420, the odds of falling below 1364 are less than 2%, and the model gives 0% odds of falling below 1293 at any time in the next four weeks.
The 250-day “potential” model score is 50 on a 1 to 100 scale. A good even-odds score for a year from now would be 1570.
The 250-day “safety” model score is 84 on a 1 to 100 scale. A good even-odds low point or minimum for the next 250 sessions would be 1389, the odds of falling below 1293 are less than 6.1%, and the model gives 0% odds of falling below 1077 at any time in the next 250 sessions.
My vote to the Blogger Sentiment Poll by TickerSense is “bullish.”


March 25th, 2007 at 1:36 am
Greetings Bill- do you have a link to a post giving an overview of the mechanics of your predictive model?
March 25th, 2007 at 8:19 am
http://marketthoughts.com/zs20051229.html
The model is a bit different from then, but essentially the same - multivariate regression, but now exclusively on technical indicators for the S&P 500.