Magic Formula Tracker

[sengsational]

For what it's worth, I'm going to attempt to document how the Joel Goldblatt magic formula does against some benchmarks. The first result will be computed in December of 2019 (that's how long he says you need to wait for it to work).

The idea will be for me to pick six dates, about 2 months apart where I "execute trades". So there will be six groups. The first six of these dates will be buys only, where 1/6th of the eventual seed money will be invested evenly over that month's picks. For the rest of these dates, I will document sells from the prior year and document the new buys. The dates will only be approximately 2 months apart.

If a randomly selected stock is already owned (in any except the current group), it will be skipped, and the next randomly selected stock will be substituted. If a randomly selected stock is already owned and in the current group (to be sold), it will not be sold (be kept for another year).

On each of the buy/sell days, I will document what my random selection of that day's magic formula stocks came up to be, and their prices.

I'll also try to keep track of dividends, splits, buyouts, etc. Not sure how good I'll be at documenting that, but I'll give it a try.

Ticker|Executed price |Execution date |
CRSPSCVT|1653.38|06/23/2016 14:15|Benchmark
VBR|106.68|06/23/2016 14:15|Benchmark
BKE|25.76|06/23/2016 14:15|June 2016 Pick
HSII|17.66|06/23/2016 14:15|June 2016 Pick
INSY|12.92|06/23/2016 14:15|June 2016 Pick
KFY|22.57|06/23/2016 14:15|June 2016 Pick
MSGN|16.81|06/23/2016 14:15|June 2016 Pick
STRA|48.22|06/23/2016 14:15|June 2016 Pick

See you in a few months!

 

[Totoro]

Keep us posted!

 

[Texas Proud]

How are you randomly selecting the stocks

If not with a computer doing random, then I would think there would be bias in the selection some way or another...

 

[sengsational]

How are you randomly selecting the stocks

If not with a computer doing random, then I would think there would be bias in the selection some way or another...

The screen on magicformulainvesting.com gave me a list of 50 stocks that ranked the highest on Goldblatt's valuation technique. I pasted those into a spreadsheet, typed =Rand() in the adjacent column and sorted the list. So, yes, using a computer and applying a mechanical, repeatable process.

 

[sengsational]

Ticker|Executed price |Execution date |
CRSPSCVT|1705.99|08/17/2016 13:20|Benchmark
VBR|110.02|08/17/2016 13:20|Benchmark
CALM |42.7|08/17/2016 13:20|August 2016 Pick
DHX |7.15|08/17/2016 13:20|August 2016 Pick
IQNT |15.35|08/17/2016 13:20|August 2016 Pick
SYNT |45.2|08/17/2016 13:20|August 2016 Pick
TGNA |21.27|08/17/2016 13:20|August 2016 Pick

 

[Totoro]

Are you recording the prices of your already bought stocks too at this new date?

Might make it easier to compare later on with the benchmark.

 

[sengsational]

No, I didn't look-up the prices on the earlier batch. Yesterday was kind of an arbitrary day in the life of the earlier batch anyway. I can generate a "brokerage statement" to mark to market on month end, I suppose.

In my spreadsheet for tracking, I'm "buying" the benchmarks with the same amount of cash I used to initiate the position. So I "own" three sets of share counts. I probably will need to use a historical lookup once a year to "reinvest dividends", adding to share counts.

 

[sengsational]

Typed in the prices today, just for fun. It would be nice if this kept happening!

Ticker|XIRR|Gain/Loss
Magic Formula Picks|43.3%|$1283
US Small Cap Value Index (CRSPSCVT)|21.4%|$682
Vanguard Small-Cap ETF (VBR)|20.9%|$667

The gain/loss dollar amount assumes that in the June and August buys, $15K was put into VBR, $15K was put into CRSPSCVT and $15K was split evenly among the magic formula picks.

 

[sengsational]

Ticker|Executed price |Execution date |
CRSPSCVT|1717.36|10/06/2016 13:45|Benchmark
VBR|110.12|10/06/2016 13:45|Benchmark
CSGS |40.93|10/06/2016 13:45|October 2016 Pick
KORS |47.44|10/06/2016 13:45|October 2016 Pick
MPAA |27.90|10/06/2016 13:45|October 2016 Pick
RGR |56.28|10/06/2016 13:45|October 2016 Pick
SWHC |25.189|10/06/2016 13:45|October 2016 Pick

A little early this time since I'm not going to be near the computer when closer to the 2 month interval.

Ticker|XIRR|Gain/Loss 10/06/2016
Magic Formula Picks|5.3%|$326
US Small Cap Value Index (CRSPSCVT)|11.1%|$675
Vanguard Small-Cap ETF (VBR)|8.0%|$492

 

[Totoro]

That's a big swing

 

[sengsational]

Yeah. The indexes, with all of their diversity, buffered it well, but my small list of stocks didn't fare as well. I never expected a smooth ride, though. It'll probably see far worse days. Hopefully I've got a lot of years to play with it.

 

[sengsational]

Ticker|Executed price |Execution date |
CRSPSCVT|1905.79|12/15/2016 16:00|Benchmark
VBR|122.17|12/15/2016 16:00|Benchmark
FPRX |50.42|12/15/2016 16:00|December 2016 Pick
HPQ |15.49|12/15/2016 16:00|December 2016 Pick
ICON |9.19|12/15/2016 16:00|December 2016 Pick
PBI |15.54|12/15/2016 16:00|December 2016 Pick
SCMP |15.45|12/15/2016 16:00|December 2016 Pick

Ticker|XIRR|Gain/Loss 12/15/2016
Magic Formula Picks|24.7%|$3430
US Small Cap Value Index (CRSPSCVT)|43.0%|$5796
Vanguard Small-Cap ETF (VBR)|42.9%|$5757

Wow. Am I doing this right? Bought VBR on 6/23, 8/17, and 10/6, 139, 136, and 147 shares, respectively, $106.68, $110.02, and $110.12, respectively. Closing today is $122.17.

 

[23window]

Are you annualizing the rate of return instead of calculating the YTD rate of return? The YTD return for VBR is closer to 21%.

 

[sengsational]

Are you annualizing the rate of return instead of calculating the YTD rate of return? The YTD return for VBR is closer to 21%.

Yes, XIRR would be annualized. Still some big numbers. I wish I had every penny in that. Of course 20 20 hindsight. And of course I'm not talking about having all invested on January 1 until today. I'm talking about investing on those three days.

 

[sengsational]

These are internal rates of return for buying in June, August, October and December. The earlier bigger XIRR's got watered-down a bit.

Ticker|XIRR|Gain/Loss 12/30/2016
Magic Formula Picks|15.6%|$2588
US Small Cap Value Index (CRSPSCVT)|34.5%|$5557
Vanguard Small-Cap ETF (VBR)|34.4%|$5521

 

[sengsational]

Here are the random picks for February.

Ticker|Executed price |Execution date |
CRSPSCVT|1957.15|02/23/2017 13:40|Benchmark
VBR|124.67|02/23/2017 13:40|Benchmark
AKRX |21.63|02/23/2017 13:40|February 2017 Pick
DIN |59.475|02/23/2017 13:40|February 2017 Pick
GME |25.82|02/23/2017 13:40|February 2017 Pick
MDCA |8.895|02/23/2017 13:40|February 2017 Pick
SPOK |19.09|02/23/2017 13:40|February 2017 Pick

XIRR is over 243 days with a uniform amount added every two months since June. It looks like the magic formula is not keeping up with the indexes.

Ticker|XIRR|Gain/Loss 02/23/2017
Magic Formula Picks|2.4%|$622
US Small Cap Value Index (CRSPSCVT)|30.5%|$7669
Vanguard Small-Cap ETF (VBR)|30.5%|$7617

Oh, and ticker SWHC became AOBC. And ticker IQNT got sold to a private equity firm. So since that one was chosen in August and the proceeds were used just now in the February picks, we no longer have a uniform amount in each of the groupings.

 

[sengsational]

This shows pricing of the June, August and October picks. There's a lot of diversity in the way these prices move. The max down has been 25% and the max up has been over 50%.

 

[sengsational]

Here are the random picks for April:

Ticker|Executed price |Execution date |
CRSPSCVT|1894.44|04/13/2017 13:36|Benchmark
VBR|120.23|04/13/2017 13:36|Benchmark
BBSI |54.29|04/13/2017 13:36|April 2017 Pick
MCFT |15.81|04/13/2017 13:36|April 2017 Pick
PDLI |2.07|04/13/2017 13:36|April 2017 Pick
TVTY |28.05|04/13/2017 13:36|April 2017 Pick
UIS |11.3|04/13/2017 13:36|April 2017 Pick

That makes six sets of buys, two months apart, just like the book prescribes. From now on, no more additional influx of funds. Every two months holdings that are one year old get sold and a new set gets bought.

XIRR for 294 days:

Ticker|XIRR|Gain/Loss 04/13/2017
Magic Formula Picks|1.1%|$406
US Small Cap Value Index (CRSPSCVT)|13.0%|$4768
Vanguard Small-Cap ETF (VBR)|12.9%|$4710

The benchmarks gave back a bunch. The random choices are not doing nearly as well as the benchmarks at this point.

 

[OldShooter]

I have always found it a bit paradoxical when I see people hawking books and newsletters that claim to disclose the secrets of getting rich in the markets. Seems like a very unlikely gig for anyone who actually knows how to do it.

If there are actually people who know the secret, I think they are lounging on tropical islands being hand-fed peeled grapes and drinking from glasses that have little paper umbrellas sticking out. They are not selling advice. To anyone.

 

[Totoro]

This one is a bit different: Greenblatt actually can afford the proverbial hand-fed grapes lifestyle, and then some. His net worth I believe is around 500 million USD.

The magic formula thing is what he sees as a way to give back and teach (he's a professor too). The book sales won't make a difference, and he has no consumer offering. Also: his website is absolutely free.

So I wouldn't be so quick to dismiss him like that.

 

[OldShooter]

This one is a bit different: Greenblatt actually can afford the proverbial hand-fed grapes lifestyle, and then some. His net worth I believe is around 500 million USD.

The magic formula thing is what he sees as a way to give back and teach (he's a professor too). The book sales won't make a difference, and he has no consumer offering. Also: his website is absolutely free.

So I wouldn't be so quick to dismiss him like that.

I admit that I hadn't researched this guy so, prompted by your post, I did.

He looks like a huckster to me. "Professor" yes, but don't forget the adjective "Adjunct." In other words he is paid an hourly rate to teach one course. Typically across the country, adjuncts earn $15/hour. Maybe Columbia pays a little more.

He was a principal in a hedge fund called Capstay Partners that made some big plays 20 years ago but the fund appears to have dried up and blown away.

For about the last 5 years he has been running a mutual fund company called Gotham Funds, which offers three funds that are clear losers. GARIX, GENIX, and GONIX (Portfolios 1, 2, and 3 below). These funds are distinguished by stunning expenses (3.5% range) and eye-watering turnover (250-300%). There is no more sure recipe for a losing fund. The only surprise is that they are not bigger losers. Here ya go:

Sorry about the quality. This forum apparently limits pictures to a very small size. PM me with an email address if you what to see the whole gory mess. Since inception:

Vanguard 500 Index Fund: 9.99% CAGR
GARIX: 2.51%
GENIX: 6.69%
GONIX: 0.34%

This is not a leader that I would consider following.

I also did a little looking at the origins of the "Magic" formula. It appears to be completely based on backtesting. In other words, no dollars were risked to come up with this theory. Relying on backtesting is very dangerous. I recommend Nate Silver's book the signal and the noise and, particularly, his discussion of overfitting. The OP is very wise to approach this as a forward-looking experiment, as that is the only real way to test a strategy. Having only looked at the guy and not much at the "magic" formula, the results to date do not surprise me. YMMV, of course.
​

 

[sengsational]

I also did a little looking at the origins of the "Magic" formula. It appears to be completely based on backtesting. In other words, no dollars were risked to come up with this theory. Relying on backtesting is very dangerous. I recommend Nate Silver's book the signal and the noise and, particularly, his discussion of overfitting. The OP is very wise to approach this as a forward-looking experiment, as that is the only real way to test a strategy. Having only looked at the guy and not much at the "magic" formula, the results to date do not surprise me. YMMV, of course.

​

Without reading the book, I'd say a skeptical reaction would indeed be expected. I like that "The Little Book..." uses the words "Magic Formula", because I think it means he's poking fun at so many attempts at formulaic methods to beat the market, most of which are from flim-flam men trying to sell something. Actually reading the book, I realized it's not quite the traditional con-man theme, so worthy of my time to experiment with.

With respect to the data being over fit, the book was written and then another book was published five or more years later. The first title was "The little book that beats the market", and the next title was "The little book that STILL beats the market". The results from those various time spans was consistent, indicating that the noise was not modeled.

Most of the book is about the way a company is measured, financially, and how, if one is patient, buying good companies that are selling for less than they "should be selling for" have, in general, a higher upside potential. In other words, most of the book is about the concepts behind why measuring companies the way he suggests is logical.

For me, this is entertainment...just something to play with. I'm not suggesting that anyone else should play. But I wanted to maintain this post to track one path through the process outlined in the book.

 

[OldShooter]

... For me, this is entertainment...just something to play with. I'm not suggesting that anyone else should play. But I wanted to maintain this post to track one path through the process outlined in the book.

Oh, I think that's fine. Very sensible. I am presently running three test portfolios, each started with $100K, and I am using them as benchmarks to evaluate other portfolios. One is a simple 65/35 index fund, one is with Schwab's robot "Intelligent Portfolios" and one is with an advisor who is authorized to sell DFA funds. I really agree with your approach; it takes real money and time to generate credible numbers.

 

[Totoro]

To each his own of course, and surely not following him blindly. Given my own results with his method even less so, I just thought (and think) the approach makes sense. I ran my own version of what sengsational is doing. Am now employing my own approach with much better results (so far), based on the same principles.

Nevertheless: he is an adjunct professor because he isn't full time - he is a hedge fund manager after all. Columbia Business School isn't some little in-state university either, this is one of the six ivy league schools and top ranked in business schools. This isn't a $15 an hour gig.

Regarding recent performance I am a bit puzzled:
https://www.gothamfunds.com/performance.aspx

Any clue how to reconcile this with your graph? Curious.

Also cannot find much if anything about Capstay, his hedge fund was and is Gotham Capital (since 1985 or so), started with $7M. Hard to find what his actual performance was, but given his net worth there had to be some success beyond market returns. He only started writing books after his fortune was already made.

 

[OldShooter]

... Regarding recent performance I am a bit puzzled: https://www.gothamfunds.com/performance.aspxAny clue how to reconcile this with your graph? ...

Good question. It looks like a difference in time periods. The backtesting tool that I use cut the history off before sometime in 2014, even though I asked it for everything ITD. I am sure it uses the CRSP database, so probably the funds were not included in the database from their opening date. I have seen this before, but for the time period the numbers are valid.

I also looked at Morningstar data and Schwab data, but I think Schwab uses the Morningstar database. The claims on the Gotham web site are roughly the same as what these resources report.

I looked at the most successful fund, GENIX, and ITD it is approximately equal to the S&P500 total return, but I have no way to know whether it was an incubator fund prior to general availability. (Incubated Fund) With only 4 1/2 years of history that could make a big difference and make recent performance far more important.

My conclusion is the same in any case. I have long since switched to passive funds, but I wouldn't touch funds with super high expenses and super high turnover. I am not a scholar of this type of fund but the Gotham funds are the most extreme I have ever seen. I'll stick to my "huckster" assessment but of course YMMV.

Re net worth, we have no way to know. I'll bet Bernie Madoff advertised some pretty impressive numbers too. Not that this guy is Madoff, but the point is the same. Nobody is passing out audited financials when these claims are made.

The point of my original post, though, was not to go after this particular guy. In the big picture all these schemes fail on average. Even if someone has a good scheme, the market quickly recognizes and destroys it. And I continue to believe that if someone has the magic touch, they are not going to be giving it away in a book or working as a fund manager using their good idea to benefit others. In this particular case, I find it interesting that he is not offering a mutual fund based on this "magic" formula. If it is so successful, why not?

The other thing to remember is the four most expensive words in investing: "This time it's different."