Use access key #2 to skip to page content.

TheGarcipian (34.14)

What’s CAPS got against me, TDRH, abitare, Gtrinvestor, StatsGeek & Allstar13913?

Recs

0

December 19, 2008 – Comments (16)

The Man is keeping us down!

I’ve been away for about 6 weeks from the CAPS blogosphere. Once I recovered from my Election Night celebrations, I concentrated on protecting those assets that I’d not already protected (namely work and continuing to secure that paycheck!). I’d intended to post a long blog about where we are now, but it’s taking me longer to compose than I’d expected. In the meantime, however, I’ve noticed a disquieting aspect of the CAPS system. For some inexplicable reason, the Accuracy calculations for some people are woefully (and ironically) inaccurate.

And some of “those people” include yours truly.

Ok, I’ll admit it: I’m anal. I keep stats on all sorts of things, including my CAPS Score, Accuracy, Ranking and general progression. It’s one way to know if I’m improving as an investor (but certainly not the only way). For example, here’s a graph of my Accuracy in CAPS beginning on 6/1/2007, when I started keeping these numbers:

TheGarcipian’s CAPS Accuracy

Since 22May2008 when I popped over 69%, I’ve been struggling to keep my Accuracy high. With the financial meltdown, it has slid just like everyone’s, I suppose. But now, I can’t seem to move beyond 63%, even though I’ve consistently closed winning positions, and even re-opened them to score more positive points. And since Accuracy affects a full one-third of my Rating, you better bet it’s important to me!  (Several of us have blogged about this aspect of CAPS; I accept them for what they are and will not go into the virtues or vices thereof).

So, I did a little investigation, taking a couple of hours and an Excel spreadsheet. First, I calculated my Accuracy by hand using the spreadsheet. According to CAPS own definition, “a member's Accuracy is how often that member has made correct predictions.” Thus, this is a simple ratio of those picks which are in the GREEN (positive) divided by the TOTAL number of picks a player has made, with two minor exceptions. What’s not clear to me is:

(1)   whether those picks earning exactly 0% are considered a positive or a negative or are ignored in the calculation; and
(2)   whether picks in the interim state of “Pending” are counted in the “total number of picks”.

In either case, for a sufficiently large number of picks, these minor factors are in the noise range.

What I found was this: as of 18Dec2008, my actual win record is 324 “in the money” picks divided by 499 total picks, which equates to an Accuracy of 64.93%, but CAPS reports I have only a 60.76% Accuracy. That’s a discrepancy of over 4%!

The Man is keeping me down!

Turns out I’m not the only one. For the large majority, the Accuracy seems to be computed correctly. Anything within a 1% discrepancy is probably due to round-off error or one of those two factors I listed above. But for some of us, the discrepancies are much bigger. Here’s the spreadsheet I created, picking some Player’s names at random and doing each calculation by hand:

CAPS Accuracy as of 18Dec2008

Most discrepancies are in the noise range, but for those Player’s in bold-italics font within the chart, the Man is messing with you!

So, Motley Fool employees, what’s going on? What’s so different about me, abitare, Allstar13913, Gtrinvestor, StatsGeek, and TDRH? Why the discrimination? Even dwot (at a -1.38% discrepancy) and DemonDoug (at a -1.23% discrepancy) could improve their standings considerably by having their true Accuracy represented; well, okay, not dwot because she is currently Supreme Queen of the CAPS Universe… :-)

I can’t believe it has anything to do with more recent picks weighing in more heavily than old/ended picks. How asinine would that be?  Accuracy is a literal weighing of your history as a stock picker; all of it (present & past) should be counted and weighed equally.

I suspect CAPS may have glitch in the software. Can someone at Fool HQ please look into this and get back to me?

For everyone else who’s not a TMF or CAPS employee, please REC this post for me so that we can get some attention paid to this matter. I know it’s just a game, but it’s my life!

16 Comments – Post Your Own

#1) On December 19, 2008 at 5:23 AM, DemonDoug (31.42) wrote:

remember that 0-4.99 caps points ended on one pick does not count toward accuracy (as i understand it caps removes those from the calculation).  If you have 0-4.99 caps points on an open stock that does count as a positive accuracy point.

and of course anything in the negative is counted against you.

does that change your spreadsheet?

Report this comment
#2) On December 19, 2008 at 8:31 AM, madcowmonkey (< 20) wrote:

Spreadsheet? We don 't need no stinking spreadsheet.

TMF must have though nobody was silly/anal enough to run a spreadsheet on 20 caps members accuracy rating. That is too funny. Go get em gar.

Report this comment
#3) On December 19, 2008 at 8:59 AM, binv271828 (< 20) wrote:

Gar! Nice to see you man! Yeah, I have been absent a lot too. And yes, it is work too that is the culprit. Nasty paychecks getting in the way of CAPS :)

Yeah, that explains me. All of my closed picks have a >5 point gain (and one negative one). So it makes sense that there is no discrepancy between by implied and reported accuracy. But I think DemonDoug is on to why the other ones are off.

Report this comment
#4) On December 19, 2008 at 9:20 AM, goldminingXpert (28.71) wrote:

GMX 993/1182 correct in your spreadsheet.

If you take out my ~40 picks closed between 0 and 5 points, you arrive at 953/1142 correct... which yields 83.4% which is what CAPS reports. I believe the 0-5% is the problem here, though I'd love a poster starting in "TMF" to weigh in.

Report this comment
#5) On December 19, 2008 at 9:46 AM, icesword2 (60.87) wrote:

The >5% rule is a real thing, and it changed my life whenever someone graciously informed me about it.  Well, it changed my CAPS performance anyway.  I was at 0.32 or so in the ratings at the time.

Report this comment
#6) On December 19, 2008 at 10:23 AM, Pureunderperform (31.52) wrote:

  The CAPS calculation is correct. You have about fifty stocks you ended with a gain between 0% and 5%. Those don't count as either accurate or inaccurate picks. All picks with gains above  5% count ended or not, between 0% and 5% gain they only count if they haven't been ended. All negative score picks count.  At the beginning of today you had 497 total picks but for calculating your accuracy score you must subtract the 51 stocks you had ended between 0% and 5%, that gives you 446 stocks. You had 271 stocks above 5% or between 0% and 5% but not ended. 271/446=.6076=60.76%. There is no problem with the system.

Report this comment
#7) On December 19, 2008 at 12:29 PM, Tastylunch (28.57) wrote:

That's a pretty cool spreadsheet Gar. Pretty neat to actually see wins/losses broken down. I especiallylike the chart you made of your own accuracy.

How long did it take you to do that? did you record your accuracy at the end of each day? Or does CAPS have a historical record you were able to use somewhere?

Report this comment
#8) On December 19, 2008 at 3:42 PM, TheGarcipian (34.14) wrote:

Thanks to all of you for setting me straight. I didn't know about the "5% Rule". Ya'd think after being on CAPS for 2 years that I would've stumbled across that. But then, my wife tells me I can't see the butter when it's right in the front of the fridge. Well, that makes a heckuva lot of difference... This also explains why some of you are doing better than me in CAPS. But not anymore! You'd better watch your rearview mirror!

To quote Gilda Radner's lovable character Emily Latella: "Nevermind!"

Sorry TMF staff for the fire drill. I figured it had to be something like that, something to which I was oblivious. madcowmonkey, glad you enjoyed the laugh. That's what I'm here for. Heh. It certainly isn't too make any money in this stock market.

Tastylunch, I usually check my stats after the markets close. I've automated some Excel spreadsheets with VisualBasic programming that pull down prices from Yahoo! Finance for stocks I own, computing a running tally and comparing it against a set goal, which is usually what it would cost me to borrow that money on margin. This is sort of what CAPS does when it compares you to the S&P500, except my comparison (against margin loan rates) is tougher because it's higher. I am also working on automating the Accuracy & Score tracking I've alluded to above, but presently it's all done by hand. Once I got it set up, it takes me about 5 minutes a day to update. I don't think CAPS has a historical record somewhere (but I wish it did), so I've got a friend working with me to produce a "web page scraper" to pull info off of CAPS pages into Excel. If you're interested, I can keep you informed. Let me know...

Merry ChristmaHanuKwanzaakah to you all!

Report this comment
#9) On December 19, 2008 at 4:13 PM, TheGarcipian (34.14) wrote:

I'm feeling a bit sheepish now...

But I did just pick Brookfield Homes to tank last night, and it did this morning -- down 30%!

Report this comment
#10) On December 19, 2008 at 10:13 PM, madcowmonkey (< 20) wrote:

cin>>

sheepout<< 

Report this comment
#11) On December 20, 2008 at 12:55 AM, TheGarcipian (34.14) wrote:

Ha, I speak Programmer!  Funny & concise, MCM.  GIGO, I suppose.

To answer the question I posed in the title of this blog, it'd be: learn all the rules so you can play all of the game.

Report this comment
#12) On December 21, 2008 at 1:02 AM, Tastylunch (28.57) wrote:

Wow That's really cool Gar.

My excel spreadsheet is a lot less sophiscated, I update it manually. Yeah I'm interested if you feel like sharing. Not really sure I have anything of value to share in return though. I'll try to think of something

Me no speaky programmer:-( Although I do know GIGO, that sounds a lot like my bsuiness plans . :-)

Nice call on BHS man did that one take a dump on friday!

Report this comment
#13) On December 31, 2008 at 5:41 PM, BravoBevo (99.97) wrote:

Maybe this should be a different post, but with all the brainiacs here I thought I'd throw open a question: What is with the weird post-closing price gyrations?

There are some days (now more frequently than before) where after everythng has settled and the cats and dogs are going to sleep when, WHAM, something happens and there is a 200 or 300 point swing (either positive or negative). I think that over time these wild swings balance out (or I hope so), but I don't quite understand how it happens.

I have noticed that when it happens, it affects everyone but in adverse ways. For players with about as many open "thumbs up" as open "thumbs down", it doesn't seem to make much difference. But for players who are heavily "thumbs down"and players who are heavily "thumbs up" it makes large differences in opposite directions. Based on watching these results happen several times, I guessing that internally within the bowels of CAPS it sets a random S&P500 "closing price" that skews everyone's score on each stock overnight. 

What's happened to me several times is that in the mornings, I've looked at a relatively dormant stock that CAPS says I have a, say, 5.5% gain (net of S&P500). So I might set that stock to close "on the open."  Let's say that the S&P500 opens flat, and the first trade of the day in my stock is at the same price as yesterday's closing price (i.e. no change).  I would have expected to have closed the position with a 5.5% gain. But continuing with my hypothetical, what really happens is that the CAPS's S&P500 score takes a major change (i.e., a jump or dive) that doesn't happen with the real S&P500. And I inadvertantly close my position with a 1% or 2% gain or a 1% or 2% loss.

Has anyone else experienced or noticed these post-closing changes? And has anyone else been burned by relying on CAPS data only to find randomness throws off the results?  The same weirdness sometimes happens when I try to close a position that didn't trade at all on a given day, and when the CAPS S&P500 lurches one way or the other, it changes significantly from the CAPS result showing at the close (4 PM Eastern time) even when the actual S&P500 hasn't moved much in the last couple of hours of trading that day.

Sorry for the long ramble . . . I just didn't know if anyone else noticed this. 

Report this comment
#14) On January 01, 2009 at 10:05 PM, dex10picks (27.74) wrote:

I TOO ONLY RECENTLY DISCOVERED THE 5% RULE after two years... it should not be burried in the FAQ like it is. It should be in the basic description of accuracy not in a separate FAQ

Report this comment
#15) On January 03, 2009 at 8:33 PM, TheGarcipian (34.14) wrote:

dex10picks,
Agreed on that point: the 5% Rule should not be buried in a separate FAQ.

BravoBevo, I've not experienced what you've described here, not exactly. But I do know that the CAPS results do not percolate through the system until about 5:10pm Eastern time, a little more than an hour after the market's close. It takes some time to compute everyone's standings, I suppose, and things keep changing until that time. However, I have noticed some weird pricing on some of the CAPS picks I have. For instance, I picked SSG on 10/14/08 to go down (red-thumbed it). Giving CAPS the benefit of the doubt and assuming the start price would have been set on that day PLUS or MINUS one day gives me a starting price anywhere between $100-$140 per share, yet it now shows my start price as $72.95/share. This price adjustment must be due to a dividend payout on 12/23/08, and I guess we just have to hope that these calculations are done right. However, the stock price on 12/22/08 (the day before the dividend payout) was just about what I'd paid for it on 10/14/08, so you'd think the "adjusted CAPS starting price" after the payout would be comparable to the $80-$87 per share range it held on 12/23/08 (the day of the payout; i.e., after the payout). Yet, CAPS has me now with a starting price of $72.95, 10% lower than what I think it should be. Now, I'll just have to wait longer to get some positive points...

I've seen other minor "disturbances in the Force", but am never quite able to put my thumb on it. Perhaps there are other things I don't understand or am not aware of with respect to the CAPS system.

Good luck, and should you happen to identify any quirks (especially if they're repeatable), please let us/them know so they can be fixed. Thanks,
--Gar

Report this comment
#16) On January 03, 2009 at 11:20 PM, BravoBevo (99.97) wrote:

TheGarcipian

Let's assume that Monday morning (1/5/2009) all of the markets open flat from Friday's close (1/2/2009).  If that were true then come Monday morning your CAPS score will see an immediate dropo of between 50 and 100 points because your picks are roughly 3:1 bullish (thumbs up). Meanwhile, abitare's score will jump at least 100-150 points because he is 6:1 bearish (thumbs down). My score will drop at least 100 points.  And all of this will happen if NOTHING changes from Friday's market close.

The more someone is mostly bullish or mostly bearish, the more volatile will be the after-market closing swings. The closer to 50/50 a CAPS player is, the less noticable the post-closing swings will be (because the thumbs up and the thumbs down will be "balanced").

 

Report this comment

Featured Broker Partners


Advertisement