Use access key #2 to skip to page content.

dalalsid (43.64)

Player Ratings on Caps are Bogus??



April 20, 2007 – Comments (6)

A player's rating indicates his percentile rank in CAPS. A player's score is the total percentage return of all his picks subtracting out the S&P. A player's accuracy is how often that player has made correct predictions.

This would mean that a player with large number of stocks will outperform a player
with small number of stocks even though the player with fewer stocks is doing better on average.

6 Comments – Post Your Own

#1) On April 20, 2007 at 12:52 PM, TMFKopp (97.55) wrote:

Not necessarily, a relatively small number of really good picks can land you a really good score. Granted if you want the #1 spot you may need to load up on picks, but keep in mind that there are a lot of ways to "win" with CAPS. I do keep an eye on my score, but I use CAPS more often to test out new ideas and strategies.

One player you may want to check out is TMFDitty. Compared to some other players he hasn't made all that many picks (I think 58?) and yet he's up there in the top 1%. I'm sure there are other really highly ranked players that have even fewer picks too, so it's definitely not necessary to have 200 or 300 picks. 

Report this comment
#2) On April 20, 2007 at 1:08 PM, adprintz (< 20) wrote:

Just to add to TMFKopp, CAPS is more of a stock picker's contest than a portfolio builder's contest.

Report this comment
#3) On April 20, 2007 at 1:33 PM, TMFKnave (98.50) wrote:

We actually see quite the variety in our leaderboard.  Go to the Players tab and click on Advanced Search and search for players with ratings from 95 to 100 and 20 or less active picks, for example.

Report this comment
#4) On April 20, 2007 at 1:53 PM, wcwhiner (29.20) wrote:

There are 362 players with 500+ scores and over 20 picks.  There are 9 such 500+ players with fewer.  In a probabilistic universe, if you want a high "score" in CAPS, you likely want a large pick list. 

Report this comment
#5) On April 23, 2007 at 9:08 AM, dalalsid (43.64) wrote:

Let us take TMFKopp's example of TMFDitty. If you take the top 10 of TMFDittys stocks and make a new portfolio (assuming the same timing for picking as TMDDitty), will that have a lower score than the current portfolio or higher?? In theory it should be higher.

I say we should take the score and divide by the number of stocks to have a better picture of overall performance.

Report this comment
#6) On April 23, 2007 at 12:50 PM, TMFKopp (97.55) wrote:

It's a tough question for me to answer since I don't have the CAPS algorithm in front of me. But, if by top ten stocks you mean his best ten, I would guess that the score would be darn near the same if not better.

Check out kesslerrk - he has 9 picks, 100% accuracy, a score of 307, and a rating of 99.15. In the TMFDitty example, he'd have the same accuracy and a higher score.

Along with that, though, I'd also refer back to what adprintz said. The rating system is really more about being an accurate stock picker than it is about being a great portfolio manager. That's why the accuracy even comes into play at all. Otherwise rankings could just be done on raw score.

Report this comment

Blog Archive

April (2)

Featured Broker Partners