Thread: Sentiment Survey Talk

I did a little research with this year's sentiment survey and found some interesting results. I will probably write about it in Monday's (12/18/06) comments, but I wanted to share this cool development.

I made various automatic allocations based on the bulls to bears ratio of our Sentiment Survey.

Here's the interesting one...

- If the Sentiment Survey had a bulls to bears ratio of 2.00 or higher (overly bullish) I went 100% G fund for the following week.

- If the Sentiment Survey had a bulls to bears ratio of 1.25 or lower (overly bearish) I went 35% C, 35% S and 30% I fund for the following week.

- If the Sentiment Survey had a bulls to bears ratio between 1.26 and 1.99 (neutral) I just kept the same allocation as the prior week:.

The results thru 12/15/06: An account gain of 25.15%.

Trying another option:

Using the same criteria above, except instead of keeping the same allocation as the prior week, I moved to a 50% G, 50% C allocation when the bulls to bears ratio was between 1.26 and 1.99 (a neutral reading):

The results thru 12/15/06: An account gain of 21.16%.

As a comparison, here are the current totals of the funds thru 12/15/06:
(editted for poor format)
Want more?

- If you used 100% G and 100% S instead of the split (35/35/30) allocation, the gain is 27.74%.
- Using 100% G / 100% I, yielded 23.95%.
- Using 100% G / 100% C, yielded 23.49%.

In all instances you would have made 16 interfund transfers during the year and, get this, you would have spent 28 of the 50 weeks in a 100% G fund allocation (no risk)!

Interesting, huh?

I have a similiar spreadsheet set up using rokids weekly tally and the account thread.
Using the leaders, it gives out an averaged allocation and a Bull vs Bear percentage.
Individually the leaders are kind of hard to follow. But, as a group they are easy to follow and their direction is more pronounced.
If one or two are wrong it doesn't matter a lot, because the group survives.

It's kind of fun to listen to the funky doom and gloom of the outside market experts and then watch our leaders make them eat their words.

Spookey ain't it, must be a atmospheric phenomenon.......

One of our readers recommended a book to me last year called, The Wisdom of Crowds. It's very interesting and I think you are on to something Spaf.

The sentiment surveys are good contrarian tools used at extreme levels (which is sort of what I am doing above) but as a "crowd" we would not want to use the total data's average as a contrarian indicator, but as 'wisdom of the crowd' just like you are doing.

The concept, like you said, is that some will be right, some wrong, but as a whole they are a third entity that is very wise. Just like the market is wise and seems to know things even before they happen. I find sentiment and the pyshological angle of the market fascinating.

I was supprised how easy it was to do. It takes me less than 5 minutes to tally the leaders allocations and update the spreadsheet. We are not going with the heard, we are going with the best of the heard.

The only draw back is that some post too close to the 12:00 EST deadline. You can talk to them about that.......

I also use the "best fund" as a factor in the average, kind of like the olive in the martini!

The "Tin Box" in the Market Talk weekly and Daily Yak thread will show their allocation average and how they appear as a percentage of Bulls v Bears.

My thinking; why reinvent the wheel!

Tom are you going to back test the data further than 1-1-06? Not sure how long the survey has been around....

Jeff

very interesting...sounds potentially like a marketable strategy/system

Originally Posted by ATCJeff

Tom are you going to back test the data further than 1-1-06? Not sure how long the survey has been around....

Jeff

I have sentiment data going back to June 2005 so if I get a minute this weekend I'll add it to the spreadsheet and post the results.

To me, this is a confirmation of what we kind of already knew. But now we have a short-term, basic verification.

Maybe we can get this allocation added to the tracker for 2007 so people can follow it.

Originally Posted by tsptalk

I have sentiment data going back to June 2005 so if I get a minute this weekend I'll add it to the spreadsheet and post the results.

To me, this is a confirmation of what we kind of already knew. But now we have a short-term, basic verification.

Maybe we can get this allocation added to the tracker for 2007 so people can follow it.

adding it to the tracker would be ideal

Originally Posted by tsptalk

I have sentiment data going back to June 2005 so if I get a minute this weekend I'll add it to the spreadsheet and post the results.

To me, this is a confirmation of what we kind of already knew. But now we have a short-term, basic verification.

Maybe we can get this allocation added to the tracker for 2007 so people can follow it.

Great idea.

Tom are you going to back test the data further than 1-1-06? Not sure how long the survey has been around...

The half year results from June 2005 (when the sentiment survey began) thru December 2005 were not as dramatic - Perhaps because it was new and fewer people took the survey.

- A bulls to bears ratio of 2.00 or higher = 100% G fund for the following week.

- A bulls to bears ratio of 1.25 or lower = 35% C, 35% S and 30% I fund for the following week.

- A bulls to bears ratio between 1.26 and 1.99 = same allocation as the prior week.

The results 6/13/05 - 12/30/05: An account gain of 6.02%.

Or it could just mean that there were fewer monkeys, but equipped with more darts.

(Right? :-)