The Magic Formula Experiment

[Totoro]

Thanks!

I indeed meant with survivorship bias if he was the only guy there because others folded.

The drop in 2008 would be interesting to find out if he went down more than the S&P (-50% roughly from highs). If not that'll be a big plus (less volatility, better results).

 

[sengsational]

The system on motley fool is a competition, and so doesn't allow for cheating.

Everything I said about the "trackjgoldblatt" player on motley fool caps must be seriously discounted.

Although I don't think it's worthless, I did find an inconsistency that calls the whole thing into question.

In the "notes" section of each stock pick there was a paste from the randomized list of stocks from the magic formula site (see image). For each transaction day (6 transaction days per year), it picked the top 5 or 6 in the list (dropping ones that appeared before, but basically this much looks legit). The problem is that the buys that day did not always equal what was randomly picked. So it seems that the guy behind this tracking was putting his own spin on things.

I think the lists he posted were really picks from the magic formula site, but he didn't act only on the proper stocks.

Since companies like the ones getting picked go through buy-outs and go out of business, getting historical pricing on ALL of the picks is very difficult. If 100% of the picks had historical pricing, I could re-create a clean test, but the first one I tried "NOOF", and many more I tried, didn't have historical pricing.

I'm disappointed I can't test this thing myself using historical information, but alas, I must throw in the towel. If I get going on 'the real thing', I'll start a "newsletter".

 

[petershk]

Thanks!

I indeed meant with survivorship bias if he was the only guy there because others folded.

The drop in 2008 would be interesting to find out if he went down more than the S&P (-50% roughly from highs). If not that'll be a big plus (less volatility, better results).

I can tell you that with a real money account using his site and method for two years the volatility was higher and the returns were lower than the S&P. Might have been unusual. The s&p had good years back then .

From what I remember that's one of the reasons he says it will continue to work. People like me can't stomach how badly these companies seem to be doing.

Examples: Apollo education and ITT were picks when for profit education was tanking. Game spot when video games were moving to digital from retail. Coach when Michael Cors was killing them the buckle when retail was losing to Amazon.

Those companies all had excellent returns on investment over many prior years but at the time experienced large share drops, which is more or less what he does (good companies cheap). By buying 20-30 you get protection from a couple failures because many are temporary setbacks.

Still... itt and Apollo both lost over 70% in one year AFTER appearing on the screen. Gamestop was up 30-40%, then down if I recall.

It was a very wild ride and it was hard psychologically holding companies that the news is incredibly negative about. I suspect if I did it with fake money it would have been easier.

I couldn't do it... I really thought I could and this is from someone that has been buying the oil crash for the last year . I suspect it does work so wish you luck!

Sent from my HTC One_M8 using Early Retirement Forum mobile app

 

[sengsational]

I couldn't do it... I really thought I could and this is from someone that has been buying the oil crash for the last year . I suspect it does work so wish you luck!

In running through the situation outlined in motley fool caps, I saw some jaw-dropping price crashes. I think it's just a waiting game for the occasional huge payoff.

By the way, I went back and removed the trackjgoldblatt trades that didn't appear to be supported by the random bi-monthly picks and it still came out to be 9.1% IRR. I kind of threw the baby out with the bath water when I saw the extra trades, but there were only a few of them. Some disturbance that I thought was bad, but wasn't was that some of the stocks were "re-picked", so there was no sale/purchase. I'm not sure if that's legit or not. Maybe once you gave a stock a chance, you should take it out of the running.

Here are some big winners:

                            price            #shares  $in/out
12/15/2009        QCOR.DL 1    4.25    BUY    700    -$2,975
12/15/2010        QCOR.DL 1    15.26    SELL    700    $10,682

4/15/2013        GTATQ 2        3.43    BUY    944    -$3,238
4/25/2014        GTATQ 2        16.52    SELL    944    $15,595

12/26/2013        STRA         36.14    BUY    230    -$8,312
12/23/2014        STRA         72.91    SELL    230    $16,769

8/16/2013        IQNT         7.26    BUY    802    -$5,823
8/18/2015        IQNT         19.12    SELL    802    $15,334

 

[petershk]

Hehe.
Yes, when I was doing my own experiment (2012 ish), I had STRA, QCOR and IQNT and they were big winners, but offset by the pain in mostly for profit education and retail. It's REALLY hard not to play the "but I KNOW that company X is actually bad" game which Greenbladt warns against. For example I'm "sure" that gamestop is dead because digital downloads is totally replacing retail game sales, so it's really hard to buy that stock in 2012... but if you look what happened it languished at 19-25 and then surged up to 50 because a new generation of consoles drove massive retail sales.

So what I "knew for sure" was completely wrong... and that is JG's dominant theory... that "historically well run businesses that fall out of favor and become cheap are, in aggregate, better investments than ones that aren't." It makes total intuitive sense... but wow... it's pain.

 

[petershk]

just to elaborate think that APOL was around 40 and showed up on the magic formula screen. cut in half over the next 2 years... and then cut in half again. ESI (ITT Tech) was around 30-40 and is now at 3. Weight Watchers was around 40-50 and was as low as 3...

those 70%-90% drops in value are far more painful than the 100-200% gains are happy .

 

[sengsational]

Does Greenblatt's technique say to exclude from consideration issues currently owned? That would result in holding for longer than one year, so I think the answer there is "no". What about issues that were previously owned? Those probably would be fair game.

 

[nun]

It's great that we can do these long term experiments and see what really happens. I'm doing a very simple one to see if a lump sum deposit in the Vanguard Balanced Index fund will beat an annuity. It will be interesting to compare it to this stock picking approach too.

http://www.early-retirement.org/forums/f28/results-of-my-experiment-after-one-year-80526.html

 

[sengsational]

This is probably the most interesting graph I was able to come-up with out of the trackjgoldblatt data. It's the cash balance after selling. The buying and selling happened roughly six times per year. Each "bucket" started with $16,667 (the six starting dots, buying about 6 stocks two months apart for a year) . From that same starting point, look at how radically each bucket balance diverged.

 

[sengsational]

Market Cap

I thought I might look at this a little closer, now that I've got some spare time, and I found something kind of surprising. I might need to check-out that book again and see what Goldblatt has to say about this...

I pulled a random sample of 50 stocks from the magic formula site (specifying over 100M market cap) and they averaged 18B market cap. One was even over 500B. I found this surprising because I had been presuming that in order to get the financials to be in that set, those companies would have to be "too small for the mutual funds to bother with". Obviously not.

Next I pulled a random sample of 50 stocks contained in Vanguard Small-Cap Value ETF (VBR) and it averaged 2.2B with a maximum of 7.6B.

Are the days of the institutions ignoring small companies gone? Is this any different than it was in the past (like when the book was written and later updated)?

 

[Totoro]

Greenblatt didn't specify size as a cutoff criterion. He does claim that the average outperformance is higher if one focuses on smaller companies, and then goes one by saying that it makes sense since those are covered less well.

Regarding your second point: institutions ignore small companies just as before because they are too small. I don't see that changing.

A 300B fund has no use for a 50M market cap company. Too little liquidity and too expensive to research. Same thing with HFT, they focus on deep and liquid markets.

Factor funds (like DFA) and robotraders do use small caps and mid caps though. That might have changed the landscape some.

In general from my own limited experience foreign small companies still hardly get covered. While big US companies have plenty of peering eyes.

 

[sengsational]

It looks like the example I was looking at, VBR, is 6.8B in net assets, so they can dive for the smaller market caps, I suppose. Also, I wonder how much research they need to do if they're just trying to match an index (CRSP US Small Cap Value Index). Of course just matching an index, they are not really "playing against me", since they're not supposed to be making bets for or against anything. Goldblatt also says about indexes "And a market-cap weighted index, if a stock is overpriced, it buys too much of it automatically. And if it’s underpriced, it buys too little of it automatically, because it’s basing it on market-cap, which is essentially price." Based on that, I'm not too concerned about institutional ownership by indexes. And I'm also not at all concerned about institutional ownership by managed funds because they're not going to have the guts to wait a long time for the hopefully eventual payout. They'll be the ones bidding up the price when the weather looks good, at which point I might be selling after the one-year holding period.

In this article, the author comments on how some big players (Apple, Microsoft, Cisco, Lorillard) are fixtures in the list, but some of those have also beaten the S&P by a healthy margin over the past five years. The author says some of those large caps that are in the magic formula screen are also in his dividend portfolio, which kind of makes sense if they are undervalued.

If I execute on this, I need to come to terms with including large cap stocks or finding a defined way of keeping them out. I'm loath to keep them out because then I'd be doing my own thing, not following the magic formula. I suppose I could justify leaving them in if, I also have a rule: "once purchased, once sold", which I'm not sure is how it's supposed to work. In other words, say you own 36 stocks and your random screen comes up with 6. What if one of those 6 was owned before and later sold? Do you re-buy it? What if another one of those 6 are currently in your portfolio? Do you keep it for another year, rather than take the next random pick? These are things I didn't consider while reading the book, so I've got it on hold at the library again.

 

[sengsational]

The book doesn't seem to answer the question of whether to keep a stock longer than a year if it comes up on the screen again. Everything I've read just says "rebalance". I did learn, though, that increasing the holding period is beneficial ((from this academic paper)), so I'm going to let positions go longer if they show up again. Also, this site that seems to be trying to replicate the methodology allows longer holding periods if still on the screen (magicdiligence).

It's perfectly acceptable to hold MFI stocks 2, 3, even 5 years or more, but only if they remain on the screen!

The book does address the large vs small cap question. Small cap has the edge, but not by a huge margin 12.1% vs 11.9% (bottom and top deciles).

So no "twists" approach leaves me feeling good about it. If I can manage to keep this 'mechanical' and stay with it, I might do OK. I just hope I don't turn it into a J*b!

EDIT:
I can't believe I missed this. Or maybe I saw it and still had the question. It looks like Joel doesn't want to tell you to hold-em or cast them off:

Q: How long should I own these stocks?

A: The Magic Formula system was designed to hold stocks for approximately one year in order to maximize your after-tax return. After the one year period you should screen for new stocks and establish a new portfolio based on the most current financial statements and stock prices. Sometimes a previously owned stock will remain on the list and then you must decide if you want to continue to hold this stock.

 

[Sarat0ga]

I use Greenblatt's book to choose my pick since fall 2015. Until now I've been a bad follower picking up only the stock I thought was good, not the same amount on any pick and trying to time the market, with as always, good and bad result.

Regarding Totoro test, I think picking up every stock at the same time of the year is what make the result looks bad, as well as many test made on the Magic formula. The market condition change a lot along the year and Joel says to get a few pick "each month" which means it spread the risk along the year and the up and down of the market. I think the best way to apply the method blindly is to take 2 pick each and every month then keep it for a year.
After a year on it means you'll have 24 stock which match his 20-30 stock.

 

[keyboarder]

Hey guys, it's nice to see so many other people trying the same method. Not so nice to see everyone not performing well, like me.

I started using the formula in 2014 based on the first book.

Since then:

• I have used the stock screener to pick stocks
• I pick 2 stocks per month and sell old ones if no longer in screener
• I spent roughly the same amount on each stock (450$ with low commissions)
• My portfolio currently holds 24 stocks that rotate during the year
• Each stock is held for around a year, but longer if it is still in the screener after one year

Since 2014, S&P 500 appreciated around 33%
Since 2014, my portfolio appreciated around 0%

Since January 2015 I write down every day which stocks are in the screener, their mkt value and I write down all my transactions in an Excel sheet.

Conclusion, I am in the 3rd year, but still waiting for the 3-5 year long-haul magic.

 

[travelover]

.........Conclusion, I am in the 3rd year, but still waiting for the 3-5 year long-haul magic.

Nice first post.

 

[Blue Collar Guy]

Love seeing these update threads. I dont even know how to post a picture, much less Greenblatts strategy. But I am curious if The OP is winning on his method.

 

[keyboarder]

Nice first post.

Why the facepalm? Was it due to the wonderful 0% growth or do you see a clear sign that I am doing something wrong in following the guidance steps of the book?

 

[travelover]

Why the facepalm? Was it due to the wonderful 0% growth or do you see a clear sign that I am doing something wrong in following the guidance steps of the book?

Some people come here, introduce themselves and become part of the forum community. Other barge in hoping to find some marks for annuities or get rich quick investment schemes. The latter group generally don't last long.

 

[target2019]

Why the facepalm? Was it due to the wonderful 0% growth or do you see a clear sign that I am doing something wrong in following the guidance steps of the book?

Welcome!

 

[sengsational]

Hey guys, it's nice to see so many other people trying the same method. Not so nice to see everyone not performing well, like me.

I'm publishing random picks in another thread on this board and reporting the results, along with some benchmarks.

The initial post was just over a year ago, and I just reported how the first year went for the first batch: Appreciated 30% vs S&P at about 20%.

I really don't expect that to be sustained, but it's nice that the first batch came in above the S&P 500.

• I have used the stock screener to pick stocks
• I pick 2 stocks per month and sell old ones if no longer in screener
• I spent roughly the same amount on each stock (450$ with low commissions)
• My portfolio currently holds 24 stocks that rotate during the year
• Each stock is held for around a year, but longer if it is still in the screener after one year

That seems like you are executing approximately the procedure outlined in the book. But I've never been clear on what happens in all cases if the random selection contains a stock already in the portfolio. Say you have a set of stocks that you've held a year, and so need "replacing". The process I follow in the other thread is to pull the screen, remove any stock currently held, including the ones to be replaced, then randomly select from the remaining stocks. This guarantees that everything gets sold after one year, nothing gets held longer. This would not guarantee that during the next cycle, previously held stocks would not get re-bought, but there would be a gap in owning that stock.

I haven't looked at the contents of the screened list as much as you have...I literally spend just an hour or so on it, every couple of months. But what I've noticed is that there seem to be "permanent members" in the screened set of stocks. So year after year, they never get bid up to a high enough price to get out of the screen. It would seem like those might be better to avoid. But of course you can't start doing that or you'd not be following the procedure in the book.

It sounds like you didn't build-up to 24 stocks over 12 months, rather you started with 24? In that other thread, you can see that every two months, I've got 5 random picks, so target a few more stocks than you: 30.

Since 2014, S&P 500 appreciated around 33%
Since 2014, my portfolio appreciated around 0%

I wonder if you have been "unlucky", or if the method isn't going to pan-out for you, or both. Do you REALLY pick randomly (i.e. use the random function in the spreadsheet)?

 

[keyboarder]

I'm publishing random picks in another thread on this board and reporting the results, along with some benchmarks.

The initial post was just over a year ago, and I just reported how the first year went for the first batch: Appreciated 30% vs S&P at about 20%.

I really don't expect that to be sustained, but it's nice that the first batch came in above the S&P 500.
That seems like you are executing approximately the procedure outlined in the book. But I've never been clear on what happens in all cases if the random selection contains a stock already in the portfolio. Say you have a set of stocks that you've held a year, and so need "replacing". The process I follow in the other thread is to pull the screen, remove any stock currently held, including the ones to be replaced, then randomly select from the remaining stocks. This guarantees that everything gets sold after one year, nothing gets held longer. This would not guarantee that during the next cycle, previously held stocks would not get re-bought, but there would be a gap in owning that stock.

I haven't looked at the contents of the screened list as much as you have...I literally spend just an hour or so on it, every couple of months. But what I've noticed is that there seem to be "permanent members" in the screened set of stocks. So year after year, they never get bid up to a high enough price to get out of the screen. It would seem like those might be better to avoid. But of course you can't start doing that or you'd not be following the procedure in the book.

It sounds like you didn't build-up to 24 stocks over 12 months, rather you started with 24? In that other thread, you can see that every two months, I've got 5 random picks, so target a few more stocks than you: 30.

I wonder if you have been "unlucky", or if the method isn't going to pan-out for you, or both. Do you REALLY pick randomly (i.e. use the random function in the spreadsheet)?

Thank you for your insight, Sengsational. I built up to 24 stocks throughout the first year, but by now I guess it wouldn't matter so much.

It seems to me that some permanent in the screener are dying businesses, but some are just slow growing companies, like Argan (AGX). They build huge energy related facilities internationally, for example. That single stock appreciated 100% for me, but many others have failed.

I certainly glad to hear that you had success with your take on the strategy!
That lifts my spirits and gives me strength to keep going.

And yes, every day I take a look at the screener using the "Get Stocks" button. It just pops out a new fresh list. Some new stocks show up, most of the time.

I am actually starting to look deeper into the stocks in the screener to rule out possible big losers and I read "5 rules for successful stock investing" by Pat Dorsey to help me doing that. It's not that it will save me from disaster, but maybe I can avoid some obvious ones.