If You Want To Win At The Races - You Need To Know Your Odds

18th Nov '15 by Edward Croft

35 comments 22622 reads

I kicked up a bit of a storm on Monday with an article titled “Why you think there are more high StockRank profit warnings than there really are”. Yes, this year has been tough for many investors in the stock market. There have lots of profit warnings, a few accounting scandals, Greek executives doing a runner, Chinese companies vaporising and seemingly most of the speculative & blue sky AIM stocks being ground to dust. As I pointed out in my article, while the prevalence and ferocity of the ‘bad news days’ tends to cloud our perceptions, it actually hasn’t really been that bad a year for sensible investors. There have been solid gains for the kinds of stocks we favour at Stockopedia - good, cheap, improving companies have (on average) done very well.

Understanding averages and probability is where many of us slip up. Ultimately the StockRanks are a stock selection system based on probability, and an understanding of probability is very useful to using them. So I thought I'd indulge in an investigation into 'hit rates' and basic probability.

A trip to the races ?

Most investors seem to treat the stock market a bit like going to the races. They find a horse (or a couple) they like the look of and put all their bets on them. Then they grip the side of their seats and hope their horses comes home. That’s really not the way to play the stock market - owning one or two stocks leads to binary outcomes - you’ll either win, or you’ll lose. The sensible way to play the stock market is to put the odds in your favour by using broad diversification while investing in higher probability stocks.

Here are 3 charts which show the percentage of stocks in each of of 10 StockRank deciles that either outperform (blue) or underperform (red) the FTSE All Share. For my own ease I’ve used annual periods to create these charts since the UK StockRank inception - from April 2013 to end March 2014, April 2014 to end March 2015, and April 2015 to today. (For benchmarking - the FTSE All Share was up 8% in the 2013 period, up 3% in the 2014 period, and has been down 9% in the latest 2015 period.)

Alternatively we can look…

Unlock with Email

Unlock with Google

Already have an account?
Login here

dmjram

18th Nov '15

I suspect Ed is referring to the likes of what happened at the turn of the millenium. Highly priced junk sold for ever increasing prices (low stock rank) and you could pick up quality companies at bargain prices (high stock ranks). It went on for ages.

Have a read of the some of the press articles at the time - Warren Buffett was a washed up investor stuck in boring stocks (brick manufacturers, soft drink makers, credit card providers etc), a relic who'd seen his best days who didnt understand the new world order. Enron, Boo.com etc were where the smart money in the market was and specific unit trusts were launched to feed the frenzy.

Hence at some point Ed thinks the market will again start chasing expensive junk and ignore quality, good value stocks. It has "previous".

tomomo

Joe

Your question to Ed reminded me of a passage from Jack Schwager's interview of Joel Greenblatt in "Hedge Fund Market Wizards".

If you asked Joel Greenblatt the same question about his Quality/Value screen, set out in "The Little Book that Beats the Market", this is what he would answer you:

"If I wrote a book about a strategy that worked every month, or even every year, everyone would start using it, and it would stop working. Value investing doesn't always work. The market doesn't always agree with you. Over time, value is roughly the way the market prices stocks, but over the short term, which sometimes can be as long as two or three years, there are periods when it doesn't work. And that is a very good thing. The fact that our value approach doesn't work over periods of time is precisely the reason why it continues to work over the long term. Our formula forces you to buy out-of-favor companies, stocks that no one who reads a newspaper would think of buying, and hold a portfolio consisting of these stocks that, at times, may underperform the market for as long as two or three years. Most people can't stick with a strategy like that. After one or two years of underperformance, and usually less, they will abandon the strategy, probably switching to a strategy that has done well in recent years."

Schwager summarises at the end of the interview:
"Greenblatt provides three critical lessons about value investing:
"1. Value investing works.
"2. Value investing doesn't work all the time.
"3. Item 2 is one of the reasons why Item 1 is true."

cig

Analyst estimates are also a convenient way to get access to management estimates (and changes thereof) -- analysts will usually start with the management estimate as their input and modify it a little bit depending on their own opinion (indeed for small caps with one lazy house broker it can be a simple pass through). So they'd have value even if the analysts' tweak was mostly noise, as raw management estimates are not usually available in databases I believe.

pka

19th Nov '15

Ed,

I would like to support the request by another poster for you to repeat your performance analyses using data for AIM stocks separated from data for stocks on the main UK market. I suspect that excluding all AIM stocks from the portfolios would significantly improve performance, even for portfolios made up of stocks in the top-decile of the StockRanks.

Edward Croft

Regarding the requests for AIM and Full list... we do have the data but just not in the same database (nor even the same database type)... so it's quite a bit of work for me to do this. I can't promise it at the moment - I've rather over-allocated my week to these couple of blogs already - but I do hear the request and will consider how to achieve it.

Timarr - I totally agree with you. I've been mulling over how to do a social score. It would be fascinating to quantitatively prove that highly social stocks that are highly ranked tend to underperform. I'm going to keep thinking... there must be a way without too much work.

Joe, just to follow up on the Greenblatt response to your question I posted yesterday. I have come across this from a recent interview of Cliff Asness

https://medium.com/conversations-with-tyler/cliff-asness-986ba8e92b4f#.wr9obaiqx

Greenblatt is propounding a quality+value system. Asness is talking about a momentum approach. But the point they make is really the same, regarding the effectiveness -- and periodic ineffectiveness -- of investment strategies based on these different factors. (And I think it's the appropriate response to your query in relation to the StockRank quality+value+momentum approach.)

Asness says:

'One thing I should really be careful about. I throw out the word “works.” I say “This strategy works.” I mean “in the cowardly statistician fashion.” It works two out of three years for a hundred years. We get small p-values, large t-statistics, if anyone likes those kind of numbers out there. We’re reasonably sure the average return is positive. It has horrible streaks within that of not working.

'If your car worked like this, you’d fire your mechanic, if it worked like I use that word. I think it is harder than you might guess, even if something works long term, to have it go away because a lot of investors can’t live through the bad periods. They decide why it’s never going to work again at the wrong time.'

John

20th Nov '15

Hello dmjram and tomomo.
Thanks for your answers and proposed explanations of Ed’s statement. However I remain unconvinced

Bear in mind that what Ed said was:
“As a group, high ranked stocks (i.e. quality, value, momentum stocks) have had a tendency to out perform low ranked stocks in the past over most 2 or 3 year cycles.” but we must be prepared for them to underperform in future. Let it be clear - I am expecting the 90+ StockRank set to dramatically underperform at some point - probably for up to a couple of years. When this happens I have no idea - but I can guarantee that it will.”

I take this to mean that at some point the top ranked stocks will underperform the bottom ranked stocks for a period of a couple of years. In other words the top decile of stocks will outperform the bottom decile of stocks.
I can visualise the Stock Ranks underperforming the market as a whole or a group of dotcom stocks outperforming the top decile for a period as during the bubble or a group of oil stocks outperforming the top decile as recently, but this is not how the Stock Ranks work. The whole principle of using the Stock Ranks is based on diversification of a sufficiently large group, not stock picking. I still can’t imaging the top decile of Stock Rank stocks underperforming the bottom decile of Stock Rank stocks. Both of these groups properly diversified would never IMO act in this way; It would have needed an awful lot of dotcom stocks or over-inflated oil stock to be occupying the bottom decile.

“Warren Buffett was a washed up investor stuck in boring stocks (brick manufacturers, soft drink makers, credit card providers etc), a relic who'd seen his best days who didnt understand the new world order. Enron, Boo.com etc were where the smart money in the market was and specific unit trusts were launched to feed the frenzy.”
Buffet is hardly the best example.
I seem to recall that Buffet survived the dotcom bubble remarkably well. He may not have performed as well as dotcom or oil shares for a period but at the end of the day he is still the winner!

As to Greenblatt’s system:
“Our formula forces you to buy out-of-favor companies, stocks that no one who reads a newspaper would think of buying, and hold a portfolio consisting of these stocks that, at times, may underperform the market for as long as two or three years. “ I don’t think this is in any way comparable to the Stock Rank system which involves researching a much wider range of factors. Granted some individual shares chosen on his system will outperform some of the time, but over the long haul a well diversified group of Stock Rank shares will always win.
At the end of the day the Stock Rank system is about picking a portfolio and not some individual shares.

Regards

Joe

Joe - dmjram is right. Expensive Junk could dominate again. The dotcom bubble was a really wild time. All value investors got crushed. The StockRank definitely wouldn't work during those periods.

If expensive junk comes to dominate, then we'd be looking at the "momentum trap" profile dominating the market. This is precisely what happened in 1998-2000 when the market defied logic, dotcoms with no business model and sky high valuations soared.

564f3d4d2163dStockRank_Taxonomy_-_Market

There's an old saying:

"If someone comes to you selling an unbeatable stock market system, then ask yourself 'why on earth is he selling it?'

I don't want to pretend that the StockRanks are an unbeatable stock market system. They've had a good recent run ... maybe it's prudent for us to expect some payback. Andrew Ang - who wrote 'A systematic approach to factor investing' - explained that the reason Value, Momentum and other factors 'payoff' is that they are a "premium earned for the risk of enduring bad times".

Value investing stopped working for years in the 1990s, and had a viciously bad time in 2008 (ask David Dreman how much he enjoyed buying bank shares that year). We've had a good run in value stocks in the UK for the last few years. At some point those bad times will come back and put everyone off value investing again. The StockRanks could underperform for an extended period.

I am going to blog more deeply on this.

snickers

Advfn is a completely horrible site aesthetically. I wonder if that's not accidental. The chaotic, uncared-for, obsolescent appearance might be tacit encouragement for the chaotic, trolling content, and allow the site's owners to do the sort of social data analysis you're thinking about.
How well-off is Clem Chambers?

Thanks Ed - I look forward to your blog!
Joe

21st Nov '15

Morning Jo,

The point with Buffett is he endured years of underperformance (and much criticism) as the market chased junk and ingnored value/quality companies. Which ties in with Ed's comments re the highly ranked StockRanks shares underperforming for a a prolonged period.

Hi dmjram
Yes I understand (and lived through) the Buffett situation.What I am querying is Ed's comments re:

"highly ranked StockRanks shares underperforming for a a prolonged period. Underperforming what?
i.e.
would the highly ranked shares have underperformed the market, (the market being driven by the dotcom bubble) which I think he meant, or they would have underperformed the lowly ranked shares in the StockRank list which is what he said.

Or are they necessarily the same?

Best, Joe.

Joe - the highest ranked decile could underperform lower ranking deciles and/or underperform the market.

Beyond the momentum trap scenario outlined above - the lowest ranked decile is also full of a lot of hard asset firms - E&P explorers, miners - and blue sky pre-revenue tech and biotech. If these sectors see a turn then of course the low ranked set may well outperform.

"The Market" is generally thought of as one of the benchmark indices - e.g. FTSE All Share, S&P 500 etc. These are market cap weighted indices and their returns can at times be dominated by a small set of mega cap stocks. There are plenty scenarios where you could see a run in mega-caps and outperformance by the indices. Ken Fisher has been writing recently that he expects large caps to outperform small caps in the late cycles of bull markets. To be honest - in mainland Europe this is what we're seeing ... the FTSEurofirst 300 has almost kept pace with the 90+ ranked set - it's outperformed all the other deciles - so it's not an unusual scenario.

25th Nov '15

Ed on this theme have just come across this article, published recently

http://papers.ssrn.com/sol3/papers.cfm?abstract_id=2676583

Use of Google Search Volume data to identify stocks which have abnormally low investor attention. 10% p.a. alpha after controlling for size, value, and momentum.

bargainvalue

30th Nov '15

It always good to narrow the group, in which is higher probability of return or even higher than average return. StockRank take into consideration many features and is quite complicated. I think the F-score is simillar, but easier ranking system. In my last article, I have shown, that Piotrosky strategy do not work well on LSE. On the other hand, F-score can be used as simple method to establish a group of stocks with high probability of success (yes, it won't be only the cheapest one). However, as you can see on the chart below, it has worked:

565c4c495c0928-9;_whole_market;_zmienne.

This folio has been created from stocks with 8 and 9 F-score, volume above 8mln pens and minimal price 20 pens.

DWit199

I have looked at using the Piotroski F Score on it's own but there is very little correlation between score and return. The same applies to the Altman Z score. However, combining the two into a Badness(tm) Rank does seem to predict the terrible performers The chart below is derived from some F and Z score data from an alternative provider. I took fully listed shares on the LSE, excluded financials and the IOB and generated a rank for Z and F score based upon 1 yr ago figures. The ranks were summed and then the ranksum was ranked to give the final Badness(tm) rank. The shares were divided into quintiles based on this rank and the performance of each quintile was assessed over the following 12 months for an equally weighted portfolio, not accounting for dividends or costs. The results are as below.

The results are even more striking when arranged by decile where it is quite clear that a portfolio constructed from the lowest ranked shares would have performed extremely poorly.

The sample size is small and the time period short but I do think that this ranking has some predictive power for below average returns and therefore provides me with a list of shares to avoid.

Societe Generale use the F-Score and the Merton Distance to Default (bankruptcy risk) as their 'quality' score in the Quality Income strategy. Works v. well.

jjis

Hi Ed,
Any progress on making the Merton Distance to default model / data that you have in your database available in the Screener?

Regards
Jamie

iwright7

6th Dec '15

Ed,

Just watched the webinar and I enjoyed the content. The 2013-2015 Stockrank 90+ graph shows a 3Y average 70% chance of "winning", which is most impressive. What about say 95+ SR - Do the odds improve further? Ian

If you want to win at the races - you need to know your odds

A trip to the races ?

Unlock the rest of this article with a 14 day trial

About the Author

Edward Croft

35 comments

Unlock the metrics for all these stocks and more...

Unlock the metrics for all these stocks and more...