Use access key #2 to skip to page content.

"risk" and playing the odds

Recs

24

June 27, 2009 – Comments (35)

I have often said that my strategy since coming to the market to buy into this big crash has been based around playing the odds.  I'll explain

-Dreman's book describes the fact that buying stocks with low p/e, low p/b, low p/s etc valuations gives much better historical returns than buying expensive stocks even with no other considerations (bankruptcies, etc.).  His stats ended in 1996 and began in 1971 i believe. 

-Historically coming out of bear market crashes, small caps perform the best

-Historically coming out of bear market crashes the stocks that went downthe most come up the most

-Historically small caps beat big caps

-Historically buying into big crashes when markets have fallen dramatically has yielded good returns

-Historically money is made primarily by going against the trend.  Everybody is bearish, be bullish.  If Casino's are really in favor (2006 or 2007) they aren't likely to yield good returns over time, if they are really out of favor they are much more likely to.  etc.  Buying cyclics when they are on cycle has historically cost you money, buying them off cycle has made money.

One of my first blogs was titled something like "the boom or bust portfolio" and described how I thought the best return and lowest risk for the current market was to buy stocks that were so battered in share price that if they simply lived they'd at least double or triple.  And diversify that across many stocks so that if even 1/2 of the companies you pick just live yo'd make money.  

Along those lines, and along with the theme of playing the odds, lets take a look at a few scenarios and analyze them for risk and probable reward (remember, its all about playing the odds).

-A.  your portfolio consists of 1000 stocks that are guaranteed by a higher power to go up 5%

-B.  your portfolio consists of 1000 stocks that will have random returns from 0-10%

-C.  your portfolio consists of 1000 stocks that will give random returns from -10% to +30%

-D.  your portfolio consists of 1000 stocks that will give random returns of either -100% or +200%.  boom or bust.  

assuming that in the cases where there is a spread in the odds that the actual outcome is in the middle.  So 1/2 of the companies hit the high mark and 1/2 hit the low markt.

the odds are that A. returns 5%

the odds are that B. returns 5%

the odds are that C returns 10%

the odds are that D returns 50%

The odds favor the most dramatically battered stocks in this example, and plain and simple I think the odds favor the most dramatically battered stocks in real life as well.  

But nowhere that you read does anybody condone buying battered stocks, except maybe Dreman.  No fund managers talk of buying them, you'd probably get a lecture if you posted on these blogs about buying them, and all of the stock picking guidelines you wille ver read will probably tell you to pick companies with criteria like

1.  good growth

2.  low debt

3.  good profit margins

4.  favorable fundamentals in their industry or business

Yet without a doubt, rarely, if ever, are stocks like those ever really cheap.  Stocks like that may have been cheap eralier this year, but thats a rare event.  And stocks not meeting those criteria were REALLY REALLY REALLY cheap earlier this year. 

I don't submit this post as a guideline or recomendation, I submit it as a thought,one that I think would probably prove to give superior returns historically if you ran the calculations... 

In essence I offer this thought:  the market so badly misprices actual material risk, especially in bearish times, that the best odds are quite probably with the most risky stocks and not with the safest. 

Or, alternately, if the market does mis-price risk to the high side in bullish times, I submit that big very safe stocks are the best things to own in bull markets when nobody wants them, and then transfer to battered beaten high risk stocks when things turn profoundly bearish. 

Diversification or "being right" is clearly a key here... 

I am of the opinion that a machine could invest better than 99% of minds, and I simply assume that my mind is one of the 99%, and as such I sould try to align myself not with hunches or thoughts or forsight, but with odds.  And I submit that moving up the risk ladder in bearish times is quite possibly something thathas better odds than moving down it.  The opposite may well be true in bullish times.

35 Comments – Post Your Own

#1) On June 27, 2009 at 8:38 PM, anchak (99.85) wrote:

You get the rec! This is contrarian investing to the hilt - however the success/failure probabilities are not statistically researched ( except the metric which Dreman had - I actually came across a recent one - and actually in the 2000s - the middle bands - made money - while the extremes got hammered - ie both low and high were bad investments - the low one- due to survival probabilities shifting I would guess) and completely hypothetical.

Also I do not know whether you read the comments or not - please read thru the well accepted form of literature and analytics that exist in the works of Joseph Piotroski. That is the basic definition of "Metric driven Value" investing

Report this comment
#2) On June 27, 2009 at 8:47 PM, LongTermBull (93.66) wrote:

I definitely think we just encountered a time in market history that most of us will probably never see again.  Stocks were so cheap it was almost impossible not to make money.  And I think your assumptions were definitely correct, if the most battered companies could just survive they would make you lots of money.

Or, alternately, if the market does mis-price risk to the high side in bullish times, I submit that big very safe stocks are the best things to own in bull markets when nobody wants them, and then transfer to battered beaten high risk stocks when things turn profoundly bearish. 

I agree with this 100%.  Contrarian investing at it's finest.

 

Report this comment
#3) On June 27, 2009 at 8:51 PM, anchak (99.85) wrote:

Also - on a historical "Secular Bear" low market P/B is between 1-1.2 ....The S&P I think had something close to 1.7 in Mar 2009.

As you did - I also thought it was great to buy at that level ( if you are interested in Piotroski P/B stocks - you may look at my other profile ac360 : the 5Y Greens are mostly those) - however -whether it is THE low - no one knows. And if it goes to 1.2  - tha'ts like a 30% hit in valuation from there

I know you are trying your best to put conviction to your "own" actions - and I laud them. I think you should hang on - there's a very good chance of 666 being the low. In which case - most of your stocks will possibly not even retest....

However when you advocate buying here - either for yourself or propound that to others - be very sure of your metric.

And incidentally, look thru Dreman's numbers - take out the 1980s' and see what the strategies did in circa 1963-1979 - the bear growled for 15 years then ...more or less -especially buy & hold types. I think in real terms - they yielded -ve returns - although they still beat the market.

Wishing you success in all your ventures

 

 

 

Report this comment
#4) On June 27, 2009 at 9:00 PM, portefeuille (99.60) wrote:

But nowhere that you read does anybody condone buying battered stocks, except maybe Dreman.  No fund managers talk of buying them, you'd probably get a lecture if you posted on these blogs about buying them, ...

------------------------------

wait a second.

------------------------------

here are two pieces of advice you may not want to follow:

i) "never catch a falling knife"

ii) "never throw good money after bad"

I have given "outperform" ratings almost exclusively to stocks that would be considered "falling knifes" at the time and ... it worked! 

...

In 18 cases I have ended an "outperform" call that was "in the minus" and "reinstalled" that "outperform" call the same day or the next day. I consider this the "caps" way of "throwing good money after bad" or "doubling down". You increase your risk because now you can incur losses of more than 100%, a pleasure that is otherwise a privilege of "underperform" calls in the "caps" game (if you ignore the benchmark thing ...). Your upside is in "lowering the basis". This has worked out in 17/18 cases (make that 18/18 if EK has risen to more than $4.51 by the time you read this - it closed at $4.37 yesterday).

I guess the more important point is the first: DO CATCH FALLING KNIFES!

Some more knife catching is done by my "portefeuille unleashed" player (see here).

...

What I am trying to say is, I guess, that my players have done Alright listening to neither i) nor ii).

------------------------------

(from this post of mine)

But you are right, I did get some feedback of the "how can you advise ..." kind.

 

 

Report this comment
#5) On June 27, 2009 at 9:01 PM, portefeuille (99.60) wrote:

(and yes, I know that I am not a fund managers)

Report this comment
#6) On June 27, 2009 at 9:02 PM, portefeuille (99.60) wrote:

managers

manager

Report this comment
#7) On June 27, 2009 at 9:12 PM, portefeuille (99.60) wrote:

(preliminary results of the strategy described in comment #4 above can be seen in comment #29 here)

Report this comment
#8) On June 27, 2009 at 9:12 PM, anticitrade (99.61) wrote:

I totally agree with this style of investing (that's where all my money is).  Additionally, I agree that a machine is better than 99% of investors... I am almost certainly in the 99%, thats why I built a machine.

 

 

Report this comment
#9) On June 27, 2009 at 9:19 PM, TMFBabo (100.00) wrote:

I like the post as a whole.  Dreman's book was definitely one of my favorites.  I also especially like the machine part.  I employ a mechanical investing approach for most of my money, so my subjective analysis can't get in the way.

Report this comment
#10) On June 27, 2009 at 11:16 PM, checklist34 (99.72) wrote:

anchak, thanks for the replies.  Just a couple of comments.

First, I stated above that I wasn't recommending buying anything, and second the price/book at the march lows was 1.2, not 1.7.  Its 1.7 today with the S&P at 920.  Next quarter may see some not-insignificant mark-ups in book value also. 

And as one more comment, enver before in history, ever, has the dividend yield on the S&P or earnings on the S&P or cash flow on the S&P or book value on the S&P been so valuable relative to current yields on treasuries and bank CDs.  The very high interest rates on "low risk" investments in the late 70's and early 80's simply deleted demand for stocks in a way that may not (or, i suppose, may) happen today.  

The "fed model" suggests that the earnings % on the S&P and treasury rates are about equivalent.  This is historically true unthil this year when the earnings return on stocks became dramatically higher than the yield on "safe" investment.

Report this comment
#11) On June 27, 2009 at 11:21 PM, checklist34 (99.72) wrote:

LTB, i agree that we aren't likely to ever get another chance to "shoot fish in a barrel" like we could earlier in late 2008 or early 2009. 

 

Porte:  thats a pretty  nifty way of upping the ante on CAPS...  end a pick, start it again, come out ahead...  nice.  Simliar to GMXs recommendation of shorting and re-shorting the same stocks to increase accuracy and amount of points.  

 

Report this comment
#12) On June 27, 2009 at 11:23 PM, checklist34 (99.72) wrote:

anticitrade, i think i may have seen a website of yours once???  you use a machinistic formula for trading?  is it also applied here in caps?

 

And bulla, thats interesting that you like a machine approach as well...  I wonder if we should start a thread to discuss no-human-intellect-required techniques...  ?

Report this comment
#13) On June 27, 2009 at 11:24 PM, portefeuille (99.60) wrote:

similar but different.

Report this comment
#14) On June 27, 2009 at 11:43 PM, checklist34 (99.72) wrote:

porte, i...

 

imagine this situation, which would apply to basically every pick i have on caps.  i picked ash at like 12 bucks, xl at like 4.50, bz at like 2.20...

and those have or will beat the S&P.  but say i cancel my BZ pick now and start it again, and it goes ont o beat the S&P from my original pick price...  

now i'd have the same amount of points but lower accuracy.

so couldn't that backfire on a guy in a case where you wind up underwater but would ultimately come out ahead?

Report this comment
#15) On June 28, 2009 at 12:14 AM, portefeuille (99.60) wrote:

now i'd have the same amount of points but lower accuracy.

You would almost always not have the same amount of points because you would then have a different entry point for both the stock and the benchmark. I don't really care about the "caps" game "accuracy".

Report this comment
#16) On June 28, 2009 at 12:29 AM, anticitrade (99.61) wrote:

Check,

Every buy I make on CAPS is based solely on the ouput of my system, Additionally, I sell when the system says to sell.  The only filtering of my picks is when MF qualifies them as "Not a pickable ticker".  

For my real life investments I do an additional 5 minutes of research to make sure I am not missing something an automated system can not detect.  

It would be interesting to see how many users here have a mechanical selection method.  I would imagine that anyone with 200 picks and a good score is leveraging some sort of technology.  Although, I really don't understand how to automate underperform picks......  It seems like those have an absence of reasons to buy, rather than reasons to short...

Report this comment
#17) On June 28, 2009 at 12:46 AM, portefeuille (99.60) wrote:

I would imagine that anyone with 200 picks and a good score is leveraging some sort of technology. 

And even more so those with 0 "picks". Have a look at some of those guys that post here. There are quite a few "decent" posts there (see comment #2,4,5 here).

Report this comment
#18) On June 28, 2009 at 1:10 AM, anticitrade (99.61) wrote:

Thanks for the link port, I am continually surprised by how little I know!  (My wife, however, is never surprsied by how little I know.)

Report this comment
#19) On June 28, 2009 at 2:23 AM, checklist34 (99.72) wrote:

short selling is a difficult game.  I have dug up the figures on a variety of short funds this evening including the grizzly short fund, some comstock funds that i don't think have been long on a stock in 15 years, and more.  They all underperform the S&P fairly dramatically in reverse.  As in they don't go up as much as the S&P goes down over time in downtrends.  Dividends work against you there, I suppose, borrowing fees, all that.

i think anticitrades comment of finding "absence of reasons to buy" vs "reasons to short" is a nice one.

porte you're a fountain of knowledge that only slowly exposes itself. 

 

Report this comment
#20) On June 28, 2009 at 4:06 AM, JakilaTheHun (99.93) wrote:

I totally agree.  My strategies since November have been remarkably similiar to this.  It's my belief that the market prices stocks on an individual basis rather than a collective basis and for this reason, huge bargains are to be had.  While some of these stocks may have very high risks, if you diversify your portfolio properly, the risks of the collective are much lower than the risks of the stocks individually and you can take advantage of this big time. 

Report this comment
#21) On June 28, 2009 at 4:11 AM, JakilaTheHun (99.93) wrote:

It would be interesting to see how many users here have a mechanical selection method.  I would imagine that anyone with 200 picks and a good score is leveraging some sort of technology.  Although, I really don't understand how to automate underperform picks......  It seems like those have an absence of reasons to buy, rather than reasons to short..

 

Not I.  

If I had some sort of technology, I imagine it would make things easier, though.  I actually just research a bunch of companies every week, green thumb the ones that look very undervalued, and occasionally red thumb the ones that look very overvalued. 

I have a personally system for quick valuation that allows me to get a ballpack valuation on a company in less than 3 minutes.  I normally do some research after that for about 15 minutes and then re-do the valuation and look at some different scenarios and then, if I come up with strong results one way or another, I act.  

However, I probably only act on maybe 1 in 8 stocks I come across.  Maybe even less than that.  

Report this comment
#22) On June 28, 2009 at 4:34 PM, checklist34 (99.72) wrote:

jakila, i'm interested in what your 5-15 minute strategy is.

I act on probably about 1 in 8 stocks also, but perhaps 1 in 3 or 4 of the severely traumatized ones...

Report this comment
#23) On June 28, 2009 at 4:41 PM, checklist34 (99.72) wrote:

ok, lets take a look at porte's cancel/re-strat strategy, not to critique porte, but because i'm curious.

I'm not likely to implement any strategy here in CAPs, I've been fairly well remiss at maintaining or paying attention to my CAPs portfolio, but I like strategies...

if we rec a stock, say ASH, with the S&P at 830 and ASH at 12.  Say in real life you bought a few shares there.  Then a few more when it hit 9, then a few more at 8, then several more at 7, several at 6, and tripled down in the 5's.  you get a cost average well under your original purchase, with the S&P well lower.

So lets say ASH goes from 12 to 6 (-50%) and the S&P goes from 930 to 744 (-20%), your score on ASH is -30.

If you then rec ASH again at 744 with ASH at 6, and the S&P goes to 930 (+25%) and ASH goes to 30 (+400%) for a score of 375.  Your total score on ASH is 345.

If you'd just let it ride, your score on ASH would be 250 (250% gain on ASH, 0% gain on the S&P).

Porte is right, this will enhance your score... but potentially reduce your % of correct calls.  It could also enchance your correct calls if the original pick would never have made it positive.

I once again submit that I am not skilled in the art of the CAPs game, lol.  oh well, life goes on.

Report this comment
#24) On June 28, 2009 at 5:05 PM, JakilaTheHun (99.93) wrote:

Checklist, it differs company to company and industry to industry, but there are certain things that tend to hold true.  I could automate part of my analysis, but not all of it. 

In general, though, I open up the most recent 10-Q for a company first and look at the balance sheet.  I start finding ways that assets might be written up or written down (so to speak) since GAAP oftentimes understates accounts and it can't predict the future either. 

I universally write goodwill down to $0.  As for intangibles, I tend to treat IAs differently for different companies.   If the IAs are "circular" (i.e. their value is completely dependent on future cash flows), I normally discount them some (or completely).  If they are not, I try to come up with a reasonable discount if applicable.  After all this, I come up with an "Adjusted Book Value" for the stock.

After that, I open up the most recent 10-K and glance over historical cash flows.  I try to figure out what drives them and how they differ over the long-term from Net Income.  I then tend to make conservative projections moving forward (if going long) and aggressive projections moving forward (if going short).  

Depending on the company, I will normally do a little more research after that.  For example, with a REIT, I might take a look at their depreciation schedule to find their year of acquisition and location of all their properties.  

Then I plug all my numbers into a spreadsheet I call the "Quick DCF".  It's a bare-bones DCF.  IMO, the P/E ratio is just a simplistic DCF analysis,but it doesn't work because it ignores asset valuation and fluctuations in cash flows for volatile firms.  The Quick DCF compensates for this.  I run the anlysis and take a look at a few different scenarios and then make a decision (or non-decision). 

 

At least, that's the quick story of my typical analysis :)

Report this comment
#25) On June 28, 2009 at 5:08 PM, JakilaTheHun (99.93) wrote:

Really, my biggest problem is merely trying to filter companies out.  Unlike people who have built programs, I have no way to filter for companies that might be cheap or expensive under my standards.  I know nothing about programming at all.  But I normally look at a variety of sources for ideas.

I wish I could start an investment firm with someone who was good at programming so we could set up an automated filter system and then I could look over interesting results with my own analysis. 

Report this comment
#26) On June 28, 2009 at 6:02 PM, anticitrade (99.61) wrote:

Jakila,

I think if you took a class in VBA programming you could pretty much do everything I did with my automated filter system.  I essentially do the same thing you do, but automatically.  Personally, I am trying to improve my ability to predict how the soft issues of a company will affect their stock price. 

I am also trying to decide if I believe in technical analysis.  

Report this comment
#27) On June 28, 2009 at 6:49 PM, checklist34 (99.72) wrote:

cool jakila, thanks for posting it up.

my system in the last 6 months isn't one that's going to work going forward, because it calls for incredibly pounded down stocks.  It was also set up with the basic goal of going 2-3x my money in 2-5 years.  So I picked mostly all stocks that were down at least 75% from their highs (in my first month or so I dabbled in some stocks that didn't meet that criteria like BA), with deep valuations.  I think at the march bottoms my average pick was about 87% off its previous high and the price/book in my portfolio was about 0.25.

Beyond that I just dug through as much of the details as I could from 10-ks and 10-qs to best assess the risk of having to realize a loss due to bankruptcy.  In a couple of cases I bet a small amount (CEM) on a stock I thought was probably going BK on the chances that it didn't, and in one case I bought a decent stake in one I basically knew was going BK (GGP) because BK was the whole investment thesis of pershing square so I thought i'd give it a shot.  I think its a 5 bagger so far.

So far so good as I'm clear of 2x my money from early jan when the S&P was a bit higher than it is today...

but this strategy probably won't work again until there is another epic crash and severely bearish period.  Andit might be outright dangerous in bullish times when risk is probably mispriced high instead of radically mispriced low.

So i'm going to need a new outlook and a new bag of tricks and a new stratgy.  Lately i've been leaning towards some market neutral options strategies... but an open mind is alwasy the key too success.  

Report this comment
#28) On June 28, 2009 at 9:39 PM, anchak (99.85) wrote:

The MI board is a killer one - I strongly advocate - and Hans ...it doesn't hurt to acknowledge my friend.....

And all the best to you guys - you have a collective way of thinking - which is good.

Incidentally....where did you get the 1.2 Checklist - I got the 1.7 from the MI board - the guy who made that statement - his data is solid ( CSRP).....

Report this comment
#29) On June 28, 2009 at 9:46 PM, anchak (99.85) wrote:

I guess S&P disagrees .....and possibly agrees with the 1.7 conjecture ( remember Q1 2009 is March end)

S&P 500 Index Level Fundamentals

Report this comment
#30) On June 28, 2009 at 10:09 PM, checklist34 (99.72) wrote:

1.  Ned Davis Research, as presented by perma-bears Comstock Funds

2.  Barras is reasonably close to NDRs estimates

3. I calculated 1.8 from Zacks and Yahoo data... that was current as of last when or whenever i posted it.  The march lows were substantially lower so by my calculation we'd have come up with about 1.3 at the march lows or so.

etc, etc, etc,

seek truth, anchak, not justification for what you want to be right, and you'll probably live long and prosper

Report this comment
#31) On June 28, 2009 at 11:12 PM, anticitrade (99.61) wrote:

Incidentally as of today, I get 1.48 for the MEDIAN value of the current p/book from my yahoo data and 1.45 from my calculated value.  The average values are much higher as a consequence of having a denomenator that approaches zero.  Also, as long as we are sharing interesting data...  I compared the total sales of all companies I analyze between 2007 and for the last 12 months.  

In 2007 I get 15,558,971 MILLION

TTM I get 17,045,707 MILLION

That's about a 10% increase.   That seems high, but may be because I do not factor in any finance companies... (I checked to make sure that I did not have stupid excel error this time.) 

Report this comment
#32) On June 28, 2009 at 11:35 PM, anchak (99.85) wrote:

Interesting...I simply put in what S&P thinks of its own index...however since you are so sure ..I did check a different resource - and I trust them much more on quality of data - Morningstar - and they do seem to agree on the S&P valuation :

http://quicktake.morningstar.com/ETFNet/Portfolio.aspx?Country=USA&Symbol=SPY

Report this comment
#33) On July 01, 2009 at 12:07 AM, checklist34 (99.72) wrote:

hey anticitrade, thanks for the input, that is really fascinating (the sales figure) and my instinct is to be skeptical and wonder if maybe you have alot of growth companies in there?

I did calculate this:

if the p/e of the S&P (TTM earnings) was 15, SIMPLY REMOVING AIG WOULD MAKE IT ABOUT 12.5.  if it was 20, simply removing AIG would amke it 16.5 or so.  If it was 100, simply removing AIG would drop it to 45. 

AIG lost $100B via the miracle of GAAP accounting.  

To put that into perspective, consider these:

-mkt cap of the S&P 500 right now is about 8.4 trillion.  The Bears love to post a p/e of 100 graph, which would imply about 84B of profits.  AIG had 100B of losses, which alone would move the S&P to 184B of earnings and vastly lower the multiple.

Throw in NON-REALIZED losses due to mark to market at insurance companies alone and the automakers, and you'd probably have darn near twice that non-considerable, one-time, oddity, loss.  That'd make for a fairly dramatic change.

AIG=100B of losses, F=15B, GM=31B.  Thats $150B... throw in the mark-downs on the balance sheets of ACAS ($3B), HIG, etc, etc, etc, and I bet we approach $200B of losses that are either anamolies that the market as a whole shouldn't be judged by or aren't actually losses but mark downs on investments here and there that may well, at least in part, be marked up one day.

Even just taking $150B for AIG + Autos... if the S&P had a TTM P/E of 100, it would be 35 without those.

My point is we have a REALLY BIG skewing of p/e's caused by things not likely to repeat or things which are temporary.  Throw in the question of how many companies were forced to write down goodwill due to the very bad economic conditions OR just chose to take charges out of the way and the figure of non-considerable losses gets even higher.

ALSO AND THIS IS REALLY INTERESTING AND I DIDN'T KNOW THIS:

insurance companies MARK DOWN REVENUE for investment portfolio mark-downs, so the P/S figures quoted above are acutally lower as well.  Witness HIG, AIG, and others had negative revenue in certain quarters.  

Some of those M2M markdowns will be marked up one day, plain and simple.

This is more complex than I originally anticipated, but it does seem that the preponderance of the complexities point towards cheaper, not towards more expensive.

I have not blogged my last blog on this topic. 

Report this comment
#34) On July 01, 2009 at 12:10 AM, checklist34 (99.72) wrote:

anchak, I appreciate the links, i'll be sure to include them in a future commentary. 

however, none of my posts are intended to skew reality, just to assess it.

I calculated the figures I got (p/b = 1.9, p/s=0.9) exactly as stated in the other thread.

I offer Ned Davis research #'s as presented by some hyper-perma-bears (comstockfunds.com, you'd love 'em if you're bearish, check them out).

 

Report this comment
#35) On July 01, 2009 at 12:13 AM, checklist34 (99.72) wrote:

FWIW, anchak, they list

p/e, prospective = 14.1 (i assume thats either forward (fiscal year 2010) or this year) not far from what I calculated

price/book = 1.8, again more bullish than what I calculated (that corresponds to about 1.3 at the lows)

price/cash flow = 5.7 (cheap, and better than I calculated)

price/sales = 0.8 (again, cheaper than I calculated)

so that link has basically all #'s showing the market cheaper than I calculated.

historical price/book (30 year)=2.4 (50 year) =~2.6

 

Report this comment

Featured Broker Partners


Advertisement