Use access key #2 to skip to page content.

Is a CAPS based Fund plausible? Attempt at Quant based strategy

Recs

51

January 10, 2010 – Comments (48) | RELATED TICKERS: IND , EXMCQ , F

Greetings fellow fools and best wishes for 2010!

I guess I feel obligated every time the year rolls over to do a blog – it pricks your conscience to take so much and not give back( Also Hans ie portefeuille keeps bugging me about doing a blog !). Given the fact that I am so irregular in the CAPS blogosphere – I have a feeling that this post will be a long-winded and maybe at times meandering one – but hopefully you will be able to keep up with it – read the high points and take away something which maybe of use.

THE CHANGING FACE OF CAPS
==========================

Its been 2+ years now for me over here. And what those 2 years have been – not just here – but from every which way I look at it!  Personally, from CAPS – what I  value the most are my online friends – something which I never bargained for.

However more importantly , with the changing leaderboard, the advent of fresh new talent ( Hans ie portefeuille, checklist34, bullishbabo, Ultralong etc), from the “Peak Oil and Commodity” blogs in early 2008 to the “Gloom and Doom” ones in late 2008 , to the Technical Analysis and Chart based “Bear Turn” call by GV ( Also I will defend my friend – who people keep ridiculing for some reason – he called the “turn”, he didn’t trade it to perfection – 2 very different things IMHO), and now the current either Ultra-bullish or Bearish polarized crowd.

Things change – yet they remain the same!  My friend FloridaBuilder (FB) is coming back to blog, although Michael ( EverydayInvestor) has left, and so seems James (TDRH). But Tasty(Lunch) keeps churning out great blogs all the time – and who can’t feel for Binve – who thru his new fatherhood status and the excitement and sleepness nights – still  produces excellent fundamental and technical analysis in a class by himself. Abit continues to blog in his inimitable fashion, kdakota630 is there with his Peter Schiff ones and dwot still comes up with her gems.

DOES CAPS WORK?
==========================

There’s an amazing blog which FB wrote last year where he had asked rhetorically “Are we still an investment website?” in face of the End of Days drivel by “u-know-who- soon to be I guess 8-9-10 or something like that”

FB has been vindicated a year +from there – but he was wrong – dead wrong – short term.( I would also mention that FB’s personal strategy has always been long-short – so its more of a philosophical debate – than quantitative measurement) The volatility that the markets went thru – its was simply impossible to keep your cool. And most people thinks- its not over yet still. Who knows?

One struggle to me personally has always been – and I think its becoming increasing a law of diminishing returns – to dissociate value from the maze in CAPS. I personally think I was lucky. I think in my first week – I stumbled into Florida’s blog – and kinda challenged him on his views – and he was patient to explain and open up his encyclopedia of knowledge to me and others here. Abit – singlehandedly – was responsible for at least defensively orienting a bunch of people in Summer 2008 – including yours truly. GV, Checklist ( he had a Rolaids blog – I think) convinced a lot of folks about the impending bottom in Mar 2009. My friends in Stinkyfeet – kept chugging away at small cap investing – and we went long with the Piotroski screener in Jan-Feb 2009.

I mean does it need to be this difficult or an outcome of circumstances?

WHAT CAPS IS MEANT TO BE
==========================

To makes things interesting and not get too insipid – I will have a tirade against the Motley Fool later here . The Gardner brothers – set this up IMHO – with one and one goal only – use collective intelligence for stock rating – and improve their own – myriads of stock advice services.

But the simple question is  can you blindly pick a 5-star rated stock from MF list  and put money on it – and see it perform ? The answer to that is as they say  “complicated”. Personal corroboration on my own portfolio – 5 stocks bought in mid/late 2008..all still negative - heavily. But then again – you’ll say –“ hey man – you are still a certified idiot from a remote village in India – of course your choice is bound to be pathetic”. So of course in my infinite wisdom – I asked the Chimp ( you do not know who he is ? What do you mean – he throws the best poo around – by self admittance amongst the builder community in Florida. Look at FB’s avatar BTW J ) He told me – he was using it in conjunction with a TA based method he followed. So then I started this journey down the path of learning TA – and of course GV comes along the way – and turned everything on its head!

Anyway , although the illiterate fool that I am , sometimes I manage to read up some things called “papers” – especially if they come from a certain , “school” in Chicago. So Joseph Piotroski wrote this paper on Value investing – which really appealed to me in concept –  so I pitched in early 2009 – all Piotroski  High Scoring picks – with a CAPS score of 3+ for Stinkyfeet. The results have been much better – although some have been negative from a CAPS perspective – you made money on all of them – some were big winners ( TCK on my own call list is one – I would have liked to pick it right at the bottom on 3/3/09 – except I had the pick from Dec 08 – so I wanted to bank accuracy!) This is however from a perspective where 80-85% of stocks have gone up since – that means on a Mean Adjusted basis – you need to have at least 92+% accuracy to be considered a Non-Random event ( err ….a small little piece of statistical jargon – somehow which found a way into my ears – called test of sample proportions. Hans ie porte beats this actually on his Biotech calls – hands down! )

THE POINT and CALL TO ARMS!
+++++++++++++++++++++++++++++++

Anyway the long-winded point that I am belaboring here is simple – everyone seems to be using CAPS as an adjunct to their strategy at best – or ignore/entertainment/blowhorn purposes at worst.

This includes the Fool staff also.  Other than the yearly blog/article on showing that CAPS ratings seem to be rank-ordering on an Alpha ( ie Market Index – SPY adjusted basis) and the paper ( Yale was it – not linking – Hans will read this – and do a much better and thorough job than me – I am pretty sure) – most of the stuff spewed out is individual opinionated perspective. Like my friend GV – who’s opinion you may disagree – but analysis you can’t discount – there are people on the MF who come and actually lay down an analyzed perspective – but like GV – its again personal.

CONTEXT and a Proof-of-Concept
----------------------------------------------------

The latest blog by Jake

on the risk ranked performance of  CAPS ratings – I am hoping you read it.

5-Stars: 65.86%4-Stars: 58.64%3-Stars: 53.97%2-Stars: 61.36%1-Stars: 23.28%SPY: 26.35%

The question is – was that just an FYI – or something actionable? Lets take a look.

On  01/01/2009 – there were 4735 CAPS rated picks ( Using the screener)

(a)    4 and 5 stars: 1942 stocks

(b)   1 and 2 stars: 1826 stocks.

  

The issue is any strategy ( even with obvious filtration like liquidity thru volume etc) –those are a lot of stocks. With the exception of the Vanguard Total Stock Market Index ( based on Wilshire 5000 I believe) or the Russell Index funds ( 2000 stocks) – very few funds have or can have this large a holding let alone individual investors.

Part I:

So I took a step back and did the following:

1.       Known Fact: In one of the old blogs – it has been proven that IF AND WHEN they work – Stocks that are transitioning into 4 star provide the best entry points in a CAPS based long strategy

2.       Took all 4 stars ( with some liquidity filters) that were not so as of 2/1/2009 but so in March 2009 …..

3.       Devised a LONG only portfolio in Smartmoney.com – with a buy date of 3/6/2009 – ie assuming the BUY TIME AS GRANTED ie known beforehand!

This way I found 103 stocks and by dropping a few index ETFs was able to restrict it to the 100 limit for Smartmoney ( Jake/Chris …guys do something with the screener – I had to do hand entry on each stock – there’s no compatibility outside of the fool !)  This is manageable – but still a lot for an individual portfolio. Anyway lets look at the results

Portfolio Gain/Loss   Win % age is 98% ( only 2 stocks lost money –QCOR and UFCS)

Additionally I checked fallout/rebalance would have been minimal : 82 stocks  are still rated 4+- so high survival rate.

The Actual Portfolio Curve

 

 

The results obviously beat the pants out of the above numbers – however the timing makes a big difference. S&P itself is up about 70% from that date. Additionally I tracked stocks which recently in Mar 2009 slipped into 1 and 2 star status – that portfolio also would be up about 70% - very similar to what was presented above – ie even laggard stocks which the MF community went against in Feb 2009 – met the market.

Ok…now you are wondering – am I on steroids or something – this is not a valid backtest – its contrived data. No question however – keep the premise in mind – here are the takeaways

(a)     Stocks that are transitioning( using 1 month or similar look-back) into a Buy in a Bull environment perform well

(b)   MF bogies the Small Cap and Mid-Cap universe – hence would mimick similar characteristics. Here's the breakout for the portfolio tested

 

(c)    There is outperformance in this 1 test window

Part II:

Used it the other way around. Chose the absolute worst day to buy stocks 9-19-2008. Did the same algorithm and found 27 stocks that qualified.

Here’s the Equity curve for holding that portfolio thru 9-19-2008 to 3/5/2009 – possibly one of the worst times ever to hold stocks ( S&P  was down 46% same period) . This portfolio would have been down 15% !!!

Equity Curve

I have not tested any short strategies – because I believe like Piotroski – shorting on a broad-based levels do not work – there it has to be much more focused and timing driven.

The Motley Fool Index Fund/ETF Experiment:
-------------------------------------------------------------------

You know what – something I learnt in an obscure class called Probability – Law of Large Numbers – seems to have unfailing habit of working time and again. On I think the very first blog in Nov 2007 – showing CAPS ratings seemingly  working – I proposed the Fool should start a Quant based Index fund on it. Of course 2 years have passed since and the Gardner brothers these days are more interested to sending me “personal invites” to participate in this “new option based strategy” – which will be my nirvana – BTW I have their first Fool Investing book – I think there was a chapter on Options in it – if I rejig my memory – something monikered “circus of freakish delights” comes to memory

So Hans and I were chatting on CIL ( coveritlive – the chat room from the Circle Community – you can access it from www.iamgv.com) about ways to construct a viable quantitative selection strategy for a MF Rating driven Index Fund/ETF.

The WHY:  The ratings performance needs to be back-tested on a real portfolio term basis – much more exhaustively and openly and regularly. A once a year dynamic cohort ( ie allowing stocks to move in and out , every which way , possibly unknowingly introducing bias – without analyzing) – study seems lame to me.

The need is obvious – Solarisking, Tigerpack –all are experimenting with this concept. However I still think – both of these initiative – although laudable – have major selection bias.

Someone needs to run this on an open platform basis ( To showcase how good people –seem to have disappeared into the ether over time in Caps – look at players CAPS1StarIndex, CAPS5StarIndex  - I am not original by any means. They were I think 2 of my FIRST favorite players – I dropped them once they became inactive)

The HOW:

(a)    Run an automated Index fund on the myriads of tracking websites available

(b)   Open up the MF database to true Search and Optimization possibilities. The current screener is – well words are flooding – but better left unsaid

(c)    Invest in analytics and share it. Talk to SAS Institute or Oracle – make the MF database open and available for better strategy building and testing.

(d)   Have someone continuously test these ideas and blog about it – rather than a crowd coming and regurgitating the same market news over and over.

PARTING THOUGHTS
-----------------------------------

The goal should be to use the system more effectively for actual monetary participation on a low-cost basis.

There is an element of adaptive market timing –embedded in the system –if MF is bearish – it would auto-switch to cash and vice-versa. I am currently working on this – as usual excruciatingly painful – hopefully that can be another blog based on people’s interest levels.

All the best and take care!

 

 

48 Comments – Post Your Own

#1) On January 10, 2010 at 11:12 PM, anchak (99.86) wrote:

2 other stats about the portfolio

Heat Map:

Beta

Report this comment
#2) On January 10, 2010 at 11:33 PM, portefeuille (99.56) wrote:

the paper.

The CAPS Prediction System and Stock Market Returns (pdf)

Report this comment
#3) On January 11, 2010 at 12:43 AM, streetflame (30.54) wrote:

"1.       Known Fact: In one of the old blogs – it has been proven that IF AND WHEN they work – Stocks that are transitioning into 4 star provide the best entry points in a CAPS based long strategy."

It reminds me of the "Dogs of the Dow"/"Foolish Four/Five" mechanical strategies which have fallen apart.  Only the 3 -> 4 star outperformance has occurred over a much shorter span of time.  In my opinion the effect has a greater chance of fading over time than 5 star outperformance (the "expected outcome", not something located by hunting through the data).

Good post though.  I will read it in more detail later.

Report this comment
#4) On January 11, 2010 at 12:51 AM, SolarisKing (< 20) wrote:

Column (6) of Table 9 uses the same regression framework as Columns (1) through (5), but with a dependent variable equal to the difference in daily returns between Portfolio 5 and Portfolio 1. Thus, the Alpha coefficient in column (6) represents the excess return earned by simultaneously buying Five-Star stocks and shorting One-Star stocks. This strategy produces a statistically significant estimated Alpha, corresponding to an annualized excess return of approximately 15%.

http://www.wallstreetsurvivor.com/Public/Content/MSN/Dashboard/AllStarPortfolio.aspx       7 mo. = +19%.

Good luck!

 

With that said, i really want to agree, in principle, that some of this can be done. 

Momentum Bias. See, another example of the need for a skelaton charm; for fools with zero site activity in a year. If you had a skelaton charm you could deduct those results to isolate momentum bias (i think).

Report this comment
#5) On January 11, 2010 at 1:08 AM, Tastylunch (29.54) wrote:

ah man this is sweet! both you and Binve have absolutlyl.killer posts today!

who people keep ridiculing for some reason

really? I thought most people had orgotten him around here. That's stupid people would bother with that. Even if you don't think TA is of value it doesn't mean  you can't respect a person who has put a lot of effort and herat into his/her craft. e.g. I personally think astrology is of no predictive value (zodiac signs etc), but that doesn't mean I can't appreciate the hard work an astrologer may put into it...

Besides if you think TA is dumb, then you should be glad people use it. I think Hans among others has proven yet again in his own way the best way to make money in stocks is to take advantage of other people's mistakes (aka "mispricing"s).So one should be happy that another holds beliefs they consider to be invalid...

I personally think QA has made a lot of TA obsolete, but I do think it's useful to know in the sense that it more or tells you the weather,

and I think its becoming increasing a law of diminishing returns – to dissociate value from the maze in CAPS

Absolutely. The Blogs badly need structure, I warned the CAPS team about this in december. I know they are very busy and probably understaffed but the problem is going critical with the hypontoad issue. I have communicated with  about half a dozen bloggers who have quit or scaled back because of this :(

There is so much noise and breadth of content it makes it difficult to find the new talent like stock7ii or the measuredinvestor. Some of us you know actually want to talk about things that can make us money. :)

I don't know if FB would be sucha  rockstar today if he debuted now instead of 2007...His awesome blogs might get drowned out.

WHAT CAPS IS MEANT TO BE

fully agree. I think the CAPS concept is proven to have merit but the scoring system needs to be tweaked for it to reach it's potential. For one thing there is zero pain for leaving incorrect calls open. I abuse this as much as anyone, but in the real world if you blow a call as a sell side guy like many do here without correcting it , you're toast.

Perhaps rl has too much pain, but I think we need a bit more to keep players (and thus ratings) honest. FWIW I trust one srra ratings more than 5.

hey man – you are still a certified idiot from a remote village in India

Anchak buddy you are too humble for your own good. :) You are one of the smarter guys on here, most likely ne of the very smartest. Furthermore you have an open inquisitive mind, a trait that i sorely lacking in too many investors.

everyone seems to be using CAPS as an adjunct to their strategy at best

Speaking of that CAPS does a particulary poor job at showcasing good momentum stocks. Guys like IBDvalueinvestin/EPSMomentum have little incentive to play given the focus on accuracy. Trend traders who often have in the low 50% accuracy range have no hope in CAPS.

as a result this leads too many CAPS players to redthumb  good momo stocks. Just because of the accuracy odds even though the risk/reward might actually say it's a good bet on the longside...

Why RINO and DRWI are three star stocks is beyond me....CAPS is a huge failure on these stocks.

my comment will continue below, this is getting long.

Report this comment
#6) On January 11, 2010 at 1:37 AM, Tastylunch (29.54) wrote:

guys do something with the screener – I had to do hand entry on each stock – there’s no compatibility outside of the fool

My goodness! That must ahve taken a really long time. Yeah the screener's getting better but it's got a ways to go.

% age is 98% ( only 2 stocks lost money –QCOR and UFCS)

WOW

Here’s the Equity curve for holding that portfolio thru 9-19-2008 to 3/5/2009 – possibly one of the worst times ever to hold stocks ( S&P  was down 46% same period) . This portfolio would have been down 15% !!!

even more impressive!

So Hans and I were chatting on CIL ( coveritlive – the chat room from the Circle Community – you can access it from www.iamgv.com) about ways to construct a viable quantitative selection strategy for a MF Rating driven Index Fund/ETF.

didn't know Porte liked GV. Word on the street is GV took down CIL for some reason. what's goin on?

 However I still think – both of these initiative – although laudable – have major selection bias.

I agree.

To showcase how good people –seem to have disappeared into the ether over time in Caps

I do think that isnt entirely CAPS's fault. I'm not sure what more they could to  keep people active. Most elite players I would say seem to be active over a 6-18 month time frame then move on for one reason or another.

The HOW:

Re: HOW methods. all good ideas, but I don't know if the Fool would want to share that data for free....It would kind of be like giving away the goose...

There is an element of adaptive market timing –embedded in the system –if MF is bearish – it would auto-switch to cash and vice-versa....

Hehthat would be funny since the Fool has said you can't time the market for years now. But then again they did change their mind about mutual funds and ptions so anything is possible I guess.

re: comment 1

Anchak have yu tried this with data say from 2007? It would be interesting to see if it works as well in a low VIX environment. All those stocks have pretty high Betas. I haven't seen the Beta trade work this well in ten years...

 – hopefully that can be another blog based on people’s interest levels.

Yes Yes please do!

Report this comment
#7) On January 11, 2010 at 7:14 AM, Dobbes (< 20) wrote:

Great post.  Enjoying the paper you linked as well.

Report this comment
#8) On January 11, 2010 at 8:26 AM, Dobbes (< 20) wrote:

Also from the paper:

6. Conclusion
We find consistent evidence that CAPS picks yield information that predicts future stock market returns for individual stocks.

 

:-o

Report this comment
#9) On January 11, 2010 at 8:46 AM, caidencollett07 (29.28) wrote:

Would also be great if screener would have the tags to use with screens.

Report this comment
#10) On January 11, 2010 at 9:48 AM, TigerPack1 (99.02) wrote:

Anchak-

Good post.  Reweighting an index mutual fund each day, like CAPS does in their stats would murder real world returns, from high trading and tax costs.

The best design I can dream up, would be a hedge fund that shorted a select group of 1 and 2 stars, while being long 4s and 5s given some filtering by an experienced investor with tons of knowledge about which one's would truly tank or fly.  It would hold a net market neutral position, if you will, for overall exposure.  I would be willing to help pick them - perhpas portefeuille, anticitrade and others with some smarts could join to find 100 up and 100 down thumbs for a new group member "experiment."

For example, subracting a 30% performance from the 65% or so number, would have yielded something like 15%-20% annually for 2009 (based on 100% sold short, use the proceeds to be 100% long setup), without taking ANY MARKET RISK.  Perhaps you could create a similar return on a rolling compounded basis, no matter what direction the market is going!  I am confident MF could find plenty of institutions that would invest in a 10%-15% compounded net neutral fund, being well diversified, with lots of quant data and a group of intelligent stock weighters in a heart beat.  We could even set up our own hedgie shingle if it works, if they don't want to hire us!

As I posted on TMFJake's blog, the Value Line Arithmetic "average," equally weighted each day like the CAPS performance numbers, gained the same as the CAPS 4 and 5 star picks in 2009!  This would be a better and more "fair" evaluation measuring stick vs. the capitalization weighted S&P 500.

See my post #12 or so, with more information here...

http://caps.fool.com/player/tmfjake.aspx?source=iersitlnk0000002

Report this comment
#11) On January 11, 2010 at 9:56 AM, TigerPack1 (99.02) wrote:

Given the above example, perhaps the leveraged returns would have been closer to 30%+, after expenses, depending on exactly how you set it up.  I may have been thinking of a less leveraged 50/50 setup to get my 15%-20% estimated annual return, from a net neutral idea.

-TigerPack

Report this comment
#12) On January 11, 2010 at 10:37 AM, anchak (99.86) wrote:

First of all let me promise - to try and reply to each comment in detail - later in the day hopefully.

Quick thoughts: Do not take this strategy as THE answer or something. Its intent was to highlight the possibility and what MF and CAPS needs to work towards. Streetflames comment on Dogs of the Dow is bang on. That strategy was something which the Gardner brothers eulogized in their book and it faded. However, one thing though - if the ratings marginal utility goes down - then I guess we are saying that CAPS ratings will see decay. I also personally, was thinking whether by writing this blog - I may introduce some anticipatory bias into the segmentation I tested.

The main thing I WANT is to understand the veracity and the willingness/appetite within this community to  pursue a goal as this one.

As I said, personally segmentation works - especially if it has the ability to adapt. Thus if it is viable in a  Low Cost Index like Fund/ETF

 

Report this comment
#13) On January 11, 2010 at 11:10 AM, binve (< 20) wrote:

AC, Hey brother! Man, this is one awesome post!! I love the ideas, and I love the backtesting of your thesis. This is really some great work!!

First, thanks for the props, I really appreciate that man :)

I really have to second all of Tasty comments above. I was going to go through and comment on your ideas one at a time, but then I read Tasty's comments and he said everything I was going to, LOL! 

One random thought I had while looking at this, which I admit is not really a comment on your basis thesis, but was sort of inspired by some of your observations: Investors are optimistic and generally have a bullish bias. And so the transition from 3->4 star status marks a nice turning point.

I was wondering on the bearish side, say all of the short/ultrashort funds which are currently at 1 star (almost all of them). I wonder if a similar timing criteria for exiting positions (since you post is very good about telling you when to get into equities but doesn't really address the exiting side) would be when a basket of short bias funds peform an average transition from 1->2 stars.

Just a thought. 

Thanks again for a great post man!!

Report this comment
#14) On January 11, 2010 at 12:02 PM, TMFJake (76.05) wrote:

Great post Anchak.  And nice work in roughing out a compelling quant screen/strategy.

It's interesting, I believe I hear on the CAPS Blogs a frequent request for some form of a "Best Ideas" portfolio from best investors on CAPS--and AllSTarPortfolio and TigerPackFund are two versions of delivering on this request.  Less freqently, I hear request for someone to create a quantitative strategy.  Two projects along these lines that immediately come to mind are two Magic Formula-influenced projects:

Hanshauge's experiment on Marketocracy:   http://www.marketocracy.com/cgi-bin/WebObjects/Portfolio.woa/ps/FundPublicPage/source=OcNfDnJbEiCdIpFhMaKiAbDd/maxDays=10000

and DCSTrade's analysis on Google:  http://sites.google.com/site/capsmfitracker/

Personally, I've love to see someone well versed in the Haugen and Baker's multi-factor models attempt to extract a CAPS amplification signal.

What I'd love for CAPS would be to have a challenge in which would match, head-to-head, a CAPS-influenced quant strategy against a Best Ideas portfolio from a team of CAPS all stars, either using one of the existing All Star portfolio experiments are a separate experiment.  I'd love to host one night each week a CoverItLive chat in which the All Star portfolio team made their buy/sell decisions, and in which we compared performance of the actively managed versus passive CAPS quant strategy. If there's interest in this, I'll be happy to put in the work to coordinate everyone and make this happen.

Report this comment
#15) On January 11, 2010 at 12:23 PM, SolarisKing (< 20) wrote:

 TMFJake (95.72) wrote:

using one of the existing All Star portfolio experiments are a separate experiment.  I'd love to host one night each week a CoverItLive chat in which the All Star portfolio team made their buy/sell decisions, and in which we compared performance of the actively managed versus passive CAPS quant strategy. If there's interest in this, I'll be happy to put in the work to coordinate everyone and make this happen.

As you likely know Jake, there is a second portfolio on Wall Street Survivor. I would be happy to provide access to it for comparison of active to passive.
   I would also be happy to attend the CIL and try to sync the admin of the basic overview of AllStarPortfolio with the comparison experiment.
   If you decided to use ASP i think it would have to stay yes man, etc, and if you decided to use WSS they don't do .PK.

   I'm eager to follow this blog thread

Report this comment
#16) On January 11, 2010 at 12:40 PM, TMFJake (76.05) wrote:

SolarisKing, yes, I think the WSS portfolio could be a good candidate for what I'm proposing *if* the portfolio mgt team were willing to sign up to participate in the weekly chats.  :)

Report this comment
#17) On January 11, 2010 at 1:51 PM, Option1307 (29.90) wrote:

Excellent ideas, we need more of this around the Fool.

Report this comment
#18) On January 11, 2010 at 9:29 PM, anchak (99.86) wrote:

streetflame:

The point about the 5 stars being the "expected outcome" - may or may not be true. I actually was thinking about it - when I devised my filters. The natural expectation would be for Stocks to open either at 5 or 1 star...However CAPS has a gestation period for the pick to gain critical mass.

Also something I SHOULD have mentioned - I asked 4+ Stars to be returned - so any Postive Gradient is welcome. And if you think about it - most scienitific ascent methods are about gradient or momentum as they say. I really did not hunt thru the data - ie to FIT. Its a hypothesis which I logically formed hoping to land on or nearabouts the Efficient Frontier ie Optimality.

Also one big difference about Dogs of the Dow or Foolish Four/Five - are price data derived strategies with percentile or tail-search filters

The CAPS rating is very different though - because it introduces an additional layer -in terms of judgement - and the Law of Large Numbers. The rating is driven by the interpretation of the price action -rather than the action itself. This doesn't mean it'll work - but the premise is slightly different.

 

 

 

Report this comment
#19) On January 11, 2010 at 9:33 PM, anchak (99.86) wrote:

SolarisKing :

WSS says ist 19% lifetime - is that excess over S&P?

Your concept is also unique - in the sense its a quant- high tail filter on investor performance - while bringing in the element of judgment thru the 2 picks. That is selection bias - however the premise being that's desired bias  - to add to alpha.

Congrats on the success!

 

Report this comment
#20) On January 11, 2010 at 9:54 PM, anchak (99.86) wrote:

My dear friend Tasty: (Tastylunch)

" Some of us you know actually want to talk about things that can make us money. :)"

You know why I like you.......

"Speaking of that CAPS does a particulary poor job at showcasing good momentum stocks. " VERY TRUE.

The accuracy thing is becoming a hindrance - yet it has its appeal . Some more thought - open debate needs to happen. I am wondering whether the Fool maybe open to adding a couple more measures?

 

On the subject of GV: He emailed me and few others - he wants to pass on the baton of the Community - now rechristened the Circle - transitioning from his persona. He wants to pursue a more focused initiative - more trading and information sharing driven ( he opened an alternative CIL for this). The old CIL will still be there - but basically more chat room.

Will keep posted. Hans has some jabs back and forth with GV  - but GV always considered him -as we all do - in very high regard - even before he became active in CIL. He's one of the most active posters there also.

 Hand Data Entry: Don't ask! My daughter practically rebeled - saying I like the computer more than her or something. I kinda kept at it to test out the premise. Unless they do something with the screener - I can't really take this much further!

The 2 long portfolio tests: 9/19 to 3/5 and 3/6 to 1/8. There  in lies the conundrum and the question that I am trying to investigate. Tasty, look at the numbers , not the performance - and you'll maybe pick up the direction! :)

The HOW: Why not - isn't this Community based intelligence? Make the data open source - and let people be at it. Heck, if you bombard me with such stuff - I may be compelled to write some unique test and search routines myself - you never know :)

Testing from 2007: Yes I would like to test that .....except it involves another ton of Data entry! Based on my time on CAPS - I almost know the answer a-priori  - it would be full of Commodity stocks!

So I am taking a shorter route on testing the alternate metric - more high level ...from Jan 2007 - month by month.I'll try to share - in any case I would have posted it at TSF or KP

Tasty....a lot of this testing depends on complete freedom to run any type of querying I want to run on the MF database. The screener is not adequate for this.

 

 

Report this comment
#21) On January 11, 2010 at 9:56 PM, anchak (99.86) wrote:

Dobbes: Thanks!

 caidencollett07: Smart observation. I was actually thinking if I could test something on the Precious Metals and the Banking tags ! Hope Chris/Jake you are hearing

 

Report this comment
#22) On January 11, 2010 at 10:05 PM, TMFJake (76.05) wrote:

Hand Data Entry: Don't ask! My daughter practically rebeled - saying I like the computer more than her or something. I kinda kept at it to test out the premise. Unless they do something with the screener - I can't really take this much further!

Anchak:  Did you use the Excel export feature from the Screener???

Yes, I hear you on the Screener and tag support.  It's on our list.

Report this comment
#23) On January 11, 2010 at 10:09 PM, anchak (99.86) wrote:

Dear Tiger ( TigerPack) .....

Good idea - but as I find - rarely is something original in CAPS - there are sooo many smart people here.

Are you aware of this guy : LongShortIndex ? You'll like him.

Although I will say this - I am not very comfortable with individual selection methodology. To me - the statistical probability of LLN ( Law of Large Numbers) trumps out.
However, Hans had this idea of using Tag leaders, testing it in the Test of Sample Proportions - prove that you are consistently beating Random ( ie adding alpha) and then use weighting for picked stocks.

Something of that sort might work in this concept also. 

 VALUELINE: Very very pertinent comment. Although I stuck to the CAPS rating to make this more relevant for CAPS - the truth is this is still in its infancy.

The Valueline system has been "around" - you know. The MI ( Mechanical Investing ) board screeners - I think both Gritton and Keelix - run a myriad of MI Screeners based on Valueline. It is actually an option - much more suited to individual investors - you have to put in more effort - maybe take a little more volatility - but its DIY.

Its just that a concept of a broader based Indexing based on a Smart Selection strategy sounds appealing to me - and of course I can then DCA based on my own Entry/Exit -algortithms :)

Report this comment
#24) On January 11, 2010 at 10:11 PM, anchak (99.86) wrote:

Yo Binve...thanks man - I have 2 blogs to read from you still :(

Recd though!

Report this comment
#25) On January 11, 2010 at 10:13 PM, anchak (99.86) wrote:

 TMFJake (95.67) #22:

Yes I did - unfortunately most other screeners need a price as of date feature to accept automated loads....and MF screeners doesn't have it

Report this comment
#26) On January 11, 2010 at 10:19 PM, anchak (99.86) wrote:

Binve...Just a though - you kidding? right?

"I was wondering on the bearish side, say all of the short/ultrashort funds which are currently at 1 star (almost all of them). I wonder if a similar timing criteria for exiting positions (since you post is very good about telling you when to get into equities but doesn't really address the exiting side) would be when a basket of short bias funds peform an average transition from 1->2 stars."

Boy! Do I want to test this? Except there are a few dependencies

(a) Ultra/Leveraged Tag filtering
(b) Additional criteria of just restricting to 2 star . I do not think any Ultra ever went beyond the 3 star - excpet like SKF and SRS in those days!

 Also one should take the Ultras out for most of the stock selection criteria. So like if you want to short the market with 1 stars - and end up shorting SRS and SKF - you get my drift!

And yes - I was wondering who would hit this one - and I was sure someone would - gives me great delight to see its Binve - my friend......timing using exit criteria. I'll have the CAPS Buy/Sell index ready hopefully in a week  - the overall one.Whether it'll work - well I do not know!

 

Report this comment
#27) On January 11, 2010 at 10:26 PM, anchak (99.86) wrote:

TMFJake and Solaris....great thoughts! I wish I could take up the challenge - love to contribute - but time is a luxury at this juncture.

Jake.....yes one could definitely take a Multi-Factor oriented approach  - a tweak on Time Varying Multinomial Regression. But you know what - you need to confer in the Fool HQ and talk about some infrastructure - is CAPS non-profit? or for-profit?

In any case get in touch with SAS Institue and talk about it. If required I can give some pointers and possible contacts. They may just collaborate for the foray  with Fool.com. You need SAS tools to do something of this magnitude - honestly.

 

Report this comment
#28) On January 11, 2010 at 10:27 PM, anchak (99.86) wrote:

Option1307 : Thanks!

Report this comment
#29) On January 11, 2010 at 10:53 PM, Momentum21 (96.33) wrote:

wow....great post and contribution.

Report this comment
#30) On January 12, 2010 at 12:27 PM, SolarisKing (< 20) wrote:

anchak (99.60) wrote:

SolarisKing :

WSS says ist 19% lifetime - is that excess over S&P?

It is not. The SP has risen approximately 17% since portfolio inception on WSS. On the other hand, it has been fairly passive; we didn't set stop loss, and perhaps should have.

Your concept is also unique - in the sense its a quant- high tail filter on investor performance - while bringing in the element of judgment thru the 2 picks. That is selection bias - however the premise being that's desired bias  - to add to alpha.

Congrats on the success!

 Thanks, though we're not out of the woods yet. It would be nice to optimize the exits on our losers. WSS is not CAPS.
   IN FACT, only two of the twelve 1-2-3 start picks in AllStarPortfolio are in the green, shattering both the accuracy and points of the portfolio with a 122 point loss.
   A stop of -20 points, and only picking 4-5 star stocks would have performed better i think.

(Tastylunch)

The accuracy thing is becoming a hindrance - yet it has its appeal . Some more thought - open debate needs to happen. I am wondering whether the Fool maybe open to adding a couple more measures?

 TSIF works his accuracy; closing out of money picks in a ratio to raise his accuracy while keeping his currency flowing. I just try to see how the caps model reflects real investing.
   SO; the one week hold thing represents the need for DD/limits trading. The accuracy represents whether i'm exiting efficiently (if your accuracy is high it should only be a matter of time before you start to make money).
   The game certainly could use some improvements.

 

The HOW: Why not - isn't this Community based intelligence? Make the data open source - and let people be at it. Heck, if you bombard me with such stuff - I may be compelled to write some unique test and search routines myself - you never know :)

    I wonder, because i don't know, how hard it is to harvest data from caps. It it possible to feed large portions to a data sifter of our own? It would not backdate well (possibly) but might be rigged to front test.
   I could supply hosting for webspace if fools wanted to combine our web building and programing expertise.
   We would set up access to different folks and each 'project' would have it's access, and then fools could control the projects authors and start feeding data directly into spreadsheets. Two years from now we would have lots of fun numbers.
   This idea is like your OS idea, but without us getting their code (but as i was saying, i don't know how much automated date mining we could do here on caps). Or perhaps TMF could be talked into sharing just certain info feeds.. . like newly minted 5 star stocks.

blah, etc.

 

Report this comment
#31) On January 13, 2010 at 12:35 AM, streetflame (30.54) wrote:

I've tried to grab CAPS data before, downloaded a couple of free web data scrapers but they did not seem robust enough. If anyone is good at scripting I'm sure it's possible, however it's probably against the terms of service and possibly not much better than using the screener and exporting to excel.  If you want access to TMF rating algorithms your best bet is applying to work there :)

What is the best resource anyone has found for grabbing financial statements?  Do you think anticitrade has a pay data subscription from Reuters or something like that?

 

 

Creating a mutual fund or ETF requires a lot of work and money for lawyers, auditors, etc.  I know it's just an idea, but it could add some realism to call it an "open-source quant-based investment club" or something like that.

I think we should build an open-source model/website kind of like anticitrade, which combines some fundamental data with CAPS sentiment. On top of that there could be an interface for partners to pick highly ranked stocks, write a brief thesis on it, and send it out to the group to vote.  The nominating partner would then be authorized to buy and sell it.

If this ever gets off the ground we should use Interactive Brokers. It is by far the best and cheapest brokerage, as low as $1 a trade for regular trades, but with even greater advantages in shorting, thinly traded securities, options, foreign markets, etc. I believe they have linked accounts purposely for investment clubs.
 

 

TigerPack: Just look at my CAPS score over the last couple months to see that market-neutral doesn't work in a lot of markets :) OK the current rally is an extreme case, but I think neutrality is not easy to achieve and furthermore not that profitable compared to some long-bias over the long term. And the vast majority of hedge funds agree.

BTW your find on ValueLine equal weight returns is excellent.  And it shows that CAPS has a TON of room for improvement in performance.  Portefueille and others have posted about improving performance by: getting rid of or de-emphasizing accuracy, expanding the price and market cap limits, putting stocks on a 100 point scale vs. a 5 point scale, open-sourcing some of the algorithms so we can help improve them, adding non-US exchanges, and so on.  I wish we could hear from TMFJake what the next big thing is on their plate, and how we could help (beyond just making picks).

anchak #18: I know you didn't hunt through the data but TMF did when they discovered that 3 -> 4 star stocks had higher returns than 5 star stocks.  It makes some sense post facto, but why 3+ -> 4- vs. 3- -> 3+, 4- -> 4+, 4+ -> 5-, etc.  I think if you asked the original architects of CAPS they would tell you that their ideal was to mark the absolute best opportunities as 5 stars. Now 2 star stocks just had a 6+ month run of monster outperformance.  The results are shifting significantly over time and you don't necessarily want to lock in on a tall, narrow performance peak, you may want to lock in on the most massive peak. Or use both.

SolarisKing #30: Speaking of which, have you seen my WSS portfolio lately? Up 58% over the same time period as ASP.  And that is despite fairly big short positions. More similar to my real life returns over the last few months than my CAPS score.

Report this comment
#32) On January 13, 2010 at 12:57 AM, portefeuille (99.56) wrote:

#31 "caps" game blog post comment of the year (2010)!

Report this comment
#33) On January 13, 2010 at 12:58 AM, portefeuille (99.56) wrote:

... and it would still be if it didn't mention me, hehe ...

Report this comment
#34) On January 13, 2010 at 1:14 AM, streetflame (30.54) wrote:

Ha, thanks port.

I'll get the spelling of your name right one of these days...

Good night (or good morning to you, right?)

Report this comment
#35) On January 13, 2010 at 1:40 AM, AllStarPortfolio (38.77) wrote:

#31)  streetflame (99.70) this one?

The objective of this portfolio is to beat AllStarPortfolio by at least 30% per year. So far I'm on target :)

Yes i have seen it. Very impressive returns. I will move into your house and be your gardener if you teach me how to do that. I'm actually rather good at math.

Very impressive returns, but. . .

Total Lifetime Trades  387

AllStarPortfolio, lifetime trades 

Total Lifetime Trades  57

I am fairly certain that you could beat ASP consistantly. But remember that AllStarPortfolio was not built to beat Streetflame. It was built for folks that aren't you to not have to worry, or work, but still make money, safely. It really does do almost all of that (so far). :)
   There are a lot of ways to work hard and risk money. 

But i really do 'envy' your prowess.  :)

I think if we had a rule in ASP to only pick 3->4 would we do even better? I am excited to see what this new idea of anchak's comes to.

-solaris

 

 

Report this comment
#36) On January 13, 2010 at 7:36 PM, streetflame (30.54) wrote:

AllStarPortfolio: Yep, that is a lot of trades.  (I believe about 3x as many as I made in real life.)  Too many trades cost more, give the potential for error, and require more work.  I think the CAPS scoring system has led me a bit astray in the direction of quantity over quality.  And of course, I spent significantly less time per trade than each person in ASP does per pick.

I think my investment approach needs more structure, diversification (sounds strange considering the number of positions, but true) and deeper focus.  That's why I like the idea of being part of a structured group instead of going it alone.  But I am not content squeaking out small edges, and I do not accept the idea that the only way to beat the market is to take on more risk.

AllStarPortfolio and TigerPackFund are good proofs of concept but I think we need a more revolutionary improvement.  The average allstar pick doesn't outperform by enough, and if left to their own devices, even allstars will occasionally make stupid mistakes (including me).  That's why I think we need some kind of system with more structure built into it.  Right now "community intelligence" might be 2x stronger than any single person, but if we can focus it more effectively it could easily be 3x or 4x.

Report this comment
#37) On January 13, 2010 at 9:45 PM, anchak (99.86) wrote:

Solaris....The 3 to 4 transition has high prob of success during times of market ascent.

My contrived test - definitely lacks one big aspect - it needs to be tested from mid 2007 ( maybe market top in Oct) - for the Commodity Play - as was the norm in CAPS those days...ie how much of the gains would CAPS have let you keep - if you used the same Star migration as exit criteria.

As I said - its a thought

Report this comment
#38) On January 13, 2010 at 9:51 PM, anchak (99.86) wrote:

streetflame

"The average allstar pick doesn't outperform by enough, and if left to their own devices, even allstars will occasionally make stupid mistakes (including me).  "

Truer words have not been spoken :)

Yes - hence I floated the idea - bring in some mechanical regimen - while leveraging the intelligence from CAPS.

I personally am much more comfortable with a diversified holding strategy - which is not prone to selection bias ( unless I somehow know that the individuals are great at selection) - so MI based with Human Intelligence is very appealing!

Report this comment
#39) On January 13, 2010 at 10:14 PM, TMFJake (76.05) wrote:

@streetflame:  I did propose an idea that I'd love to get volunteers to help make happen:

What I'd love for CAPS would be to have a challenge in which would match, head-to-head, a CAPS-influenced quant strategy against a Best Ideas portfolio from a team of CAPS all stars, either using one of the existing All Star portfolio experiments or a separate experiment.  I'd love to host one night each week a CoverItLive chat in which the All Star portfolio team made their buy/sell decisions, and in which we compared performance of the actively managed versus passive CAPS quant strategy. If there's interest in this, I'll be happy to put in the work to coordinate everyone and make this happen.

I want volunteers to run the quant portfolio as well as to be the actively managed portfolio management team.

Report this comment
#40) On January 14, 2010 at 10:50 AM, TigerPackFund (< 20) wrote:

I will take 10 smart people over a computer ANY day of the week.  Computers only spit out what you put into them.  They lack creative thinking and visualization skills. Always will!

I don't think computers put man on the moon.  MAN put man on the moon, with almost no computing power.  Didn't a bunch of smart people "invent" and "create" computers to begin with?

-TigerPack

Report this comment
#41) On January 15, 2010 at 1:12 AM, streetflame (30.54) wrote:

Jake, I would like to try quant stuff but unlike you guys I don't have access to Capital IQ and the CAPS database.  Of course, I could download excel files from the CAPS screener but that is pretty limited.  (By the way the screener functionality is still impaired from where it was before its facelift. Short interest is completely broken and it doesn't allow the correct range for several other variables.  I sent you guys a note.)  And then there is backtesting. Not sure where to get started really.  As far as I can tell, to get into interesting and useful quant stuff you have to lay out a lot of cash.

Report this comment
#42) On January 15, 2010 at 9:36 PM, Tastylunch (29.54) wrote:

Just wanted to bump the thread to keep it active. I don't wnat this discussion to ide. This is the direction CAPS needs to go if it's truly going to be useful.

streetflame

I agree with porte, streeflame your commnet 31 was excellent

and yes I also agree the VLAI is a superior metric than the S&P500

TigerPackFund 

I do agree to achieve maximum perfromance you have to have human operators. MI only takes you so far. In the end You need a Jim Simmons or Warren Buffet to surpass that. A lot of the best invetsors I've ever seen just plain outsmart others (many by just trying to frontrun proven mechanical methods), but most of their techniques suffer from scalability issues.

anchak

The 2 long portfolio tests: 9/19 to 3/5 and 3/6 to 1/8. There  in lies the conundrum and the question that I am trying to investigate. Tasty, look at the numbers , not the performance - and you'll maybe pick up the direction! :)

Man I am stumped I've looked at it for days and I;m drawing a blank. I fear I haven't been sleeping enough lately. You'll have to take pity on me  with this one man .

The HOW: Why not - isn't this Community based intelligence? Make the data open source - and let people be at it.

Yes it is, but I'm not sure the community based intelligence is menat for the free community. Kinda like those saver cards at many retailers.

Certainly I'd prefer it if it was open source like you suggest. Well hopefully they will gie you what you need!

As for young one and time spent

and hey man maybe your daughter will like this stuff! I'm lucky my grandfather taught me some things before he passed away. I'll always remember reading smart monye and value line with him as a kid. Plus I think for him it served two purposes at once :)

Tasty....a lot of this testing depends on complete freedom to run any type of querying I want to run on the MF database. The screener is not adequate for this.

yeah it's one of many great ideas that has not reached it's potential yet.

 just wanted to sya again fantatisc post AC!

 

Report this comment
#43) On January 17, 2010 at 12:04 AM, anchak (99.86) wrote:

Tasty...Its really simple man....

27 stocks in Sep 2008 while about 103 in Mar 2009.....

The delta seems to have been moving. Binve - alluded to it - saying if there was a way to indicate a downturn signal.

I constructed a CAPS Buy/Sell Index from Jan 2007. Thought it would be straight forward - however ran into some survivor/entry issues with the data - need to figure out a way to fix it.

Report this comment
#44) On January 17, 2010 at 1:03 PM, Tastylunch (29.54) wrote:

thanks , I've been working so many hours lately doing inventory/taxes I'm not thinking very straight. 3 hours of sleep nightly is bad for cognition, it's amazing how stupid it makes you :(

Report this comment
#45) On January 17, 2010 at 5:09 PM, SolarisKing (< 20) wrote:

I have this conversation on my "following" list. I wanted ya'll to know that; just because i don't reply, doesn't mean i'm not here.

 I would love to be at any chat that happens on this subject. 

One of the problems i see is the amount of energy that folks have to put into any certain subject. In AllStarPortfolio there was (as i predicted in writing) a lot of energy at first, but quickly tapered off. When we first spoke of the WSS portfolio, and it's blog requirements, there was a lot of 'yes, i'll help', but in the end folks are just all very busy.
  Currently we plan on the possiblility to use the WSS portfolio as an experiment between active and passive, but we are in hazard of losing our status there because we (me) don't blog there hardly ever.

   Still, success breeds interest, and if easy money was being made fools might pay more attention.

Report this comment
#46) On January 21, 2010 at 8:33 AM, TigerPackFund (< 20) wrote:

anchak-

Did you email me that VMW pick back in September?  It would have added 20 points or so to the group effort.

I noticed a month ago that some of the first emails sent by new members to TigerPackFund were automatically moved to my spam folder at the Yahoo! address.  I apologize if this was the case with your email, and I missed your call.

We could definitely use your smarts at TigerPackFund.  Send other great picks by email to get them added, up to 5 "open" at a time to tigerpackfund@yahoo.com.  Some red thumb picks of a few individual high-flyers or dead beat companies right now could help us hedge the stock market decline that is upon us.

You are one of the real assets here on CAPS!

-TigerPack

Report this comment
#47) On January 21, 2010 at 3:28 PM, anchak (99.86) wrote:

Yes I did Tiger...... Hans was pretty insistent on me doing an analysis on it - and the Technicals looked awesome at that time - infact I was pretty certain it would see around 45......

So I had put in a limit open on some VMW 35 calls - my last trading action of the rest of the year ( other than regular DCA etc) - never got filled.

I rarely get any opportunity these days to devote any time to trading...but still play CAPs and do CIL - due to the community/friendship etc.

TIger....if you have PM ( Precious Metals) shorts - keep an eye on them - I think they just entered a downturn channel. :)

All the best to you!

Report this comment
#48) On January 21, 2010 at 10:28 PM, TigerPack1 (99.02) wrote:

I am so sorry.  My bad!!!  I wonder if others slipped into my spam folder and got deleted before I could read them.

If you want to make new picks, post on the TigerPackFund blog that you emailed me, so I will make sure to find it.

-TigerPack

Report this comment

Featured Broker Partners


Advertisement