1YPEG - Backtested?

http://discussion.fool.com/a-new-statistic-8211-the-1ypeg-317149…

I may be wrong but I believe this was the genesis of the 1YPEG?

There seems to be a great deal,of interest and effort in calculating this number and getting it just right.

But has anyone Backtested this number to see if it correlated with stock price movement, lower betas, etc?

Seems like a great deal of preoccupation with this number if there isn’t some evidence that it is of value rather than subject to confirmation bias.

Since the denominator is growth rate, it may naturally select for stocks that are smaller caps and high flyers?

But more importantly, does it have any positive or negative predictive value?

8 Likes

Since the denominator is growth rate, it may naturally select for stocks that are smaller caps and high flyers?

Hmmm. Whats the largest cap stock of them all? And what’s it 1YRPEG?

The answer is AAPL. And the 1YRPEG is extremely low (0.45 as per Kevin’s Google Sheet). Sure AAPL’s had one hell of a run in growth over last year, but the point is that large cap companies can have low PEGs too. Look at SWKS, for example: $20B in market cap and a PEG below 0.3.

Small cap high fliers often have very high P/Es which would drive the PEG up unless they also have very high earnings per share growth. One problem with high flier small caps is that they often have negative earnings so using the 1YRPEG isn’t very relevant.

Chris

4 Likes

Great Idea.

I bet the GTR1 backtester used on MI board could do it, but I am not facile enough to do it without spending the rest of my free time this month (which isn’t much)

Maybe someone here has the skillz to do it!

Gator

And Chris,
No disrespect, but cherry picking a couple examples from thousands of stocks after a massive 7 year bull run is not really a very convincing argument. Logic would say that stocks doing well would have good growth rates vs. price. That is basically the first chapter of any investing book. Of course, the devil is in the details, and that’s the rub. This game is not as easy as some posts make it out to be, and I bet Saul would have agreed with that when his port lost 65% of it’s value in 2008. That type of event would have many here wimpering under their beds and wishing they had never logged on to the Fool.

Duma poses a very good question, as 1YPEG has become probably the prime metric followed here. It could be great! But not unreasonable to do the math and know for sure if the grail is really the grail. Trolling out AAPL and SWKS really has no bearing on his question, unless anecdotes have become the scientific method.

Gator

16 Likes

Hey Gator:

A quick place to start is looking at the PEG for it’s predictive value…been somewhat disappointing. Maybe that is because the forward growth prediction is super flawed as Saul has suggested or maybe these indicators just confirm our own bias.

While not referencing individual stocks, this article describes the lack of predictive value for PEG as regards the market as a whole:

https://www.advancedeventsystems.com/Public/EventTeamList.as…

But again…maybe the PEG is flawed by the guesswork on future growth rather than what JUST occured this trailing 12 months and may therefore continue for another.

Just food for thought as the number crunchers crunch :slight_smile:

3 Likes

Gator,

You made some good points. I try to use common sense. I don’t just invest in a stock because it has a low PEG. I look at PEG maybe as a first pass but then I look more closely at some other details because making an investment.

For instance, today I looked at Kevin’s 1YRPEG Google Sheet to see if I could find some more candidate for investment. For the list of companies there that I don’t already own, I looked at the following:

AAPL
AFSI
CASY
INFN
JPM
LUK
LUV
NXPI
PAYC
TARO
ZOES

Of these companies, I was able to rule out the following companies for the the following reasons:

AAPL too big; see no change to grow 10x

JPM own BOFI and INBK; don’t want to own a big bank

LUK probably not but need to check where revenue comes from; don’t want to invest if their revenue is from their investing activities as I don’t think this is necessarily sustainable

LUV don’t want to own an airline

NXPI own SWKS so I don’t need to own a SWKS competitor

PAYC not convinced their EPS growth will continue; high P/E so very dependent on EPS growth to continue. May only get 7c EPS next Q

ZOES negative earnings in several quarters really skew results; not confident in future growth

The list, however, did leave me with several companies to investigate further. They were AFSI, CASY, INFN, and TARO.

So I think that using the historical 1YRPEG is a good initial screening tool for new ideas to investigate further.

Chris

20 Likes

as 1YPEG has become probably the prime metric followed here.

While it has certainly attracted attention by being new and there are a flurry of posts from people trying to make sure they have figured out how to compute it … a reasonable response, considering that it isn’t a standard thing one can just look up … I would say that there has also been a significant amount of balance in terms of both that it is only one number and one should pay attention to the rest of the company and in terms of special conditions in which the number itself is suspect.

Myself, I would give it 6-12 months. If we turn into a bunch of slavering idiots pursuing low 1YPEG above all else, then there is reason to think that we have all lost our minds. But, if the newness fades and it becomes simply one more tool for sanity testing potential investments and/or screening, then I don’t think we have much to worry about.

Thus, I question the value of much backtesting. Oh, it might be interesting, but backtesting in general is so vulnerable to bias from the period selected that it is suspect in the best of circumstances. And, if what one is testing is how well 1YPEG would do if it were the only number one paid attention to, then how is that really relevant to a strategy in which one is always paying attention to the company as a whole and 1YPEG is just one more number that one uses for screening and checking?

12 Likes

http://discussion.fool.com/10-year-backtests-to-2015-feb-3169702…

you can find back testing of various stock selection methods . While Saul’s variety of PEG has not been tested, varieties of regular PEG have been. With disappointing results , they are only a tad better than SP 500.

Some of these methods have been back tested to the 1920’s and tested post discovery too. I am not using any at this time because the bull is old and I know that despite the historical numbers I will not be able to stick to them through big loses. I’m just not enough of a spread sheet guy. And the stocks they pick often seem terrible, that’s why they are cheap.

. No doubt somebody on that board can backtest almost any configuration of PEG. The tools are very powerful but hard to master.

2 Likes

Thus, I question the value of much backtesting. Oh, it might be interesting, but backtesting in general is so vulner

Seems you know more than the rest of the seasoned and enlightened investment world:

http://www.investopedia.com/articles/trading/05/030205.asp

Backtesting is one of the most important aspects of developing a trading system. If created and interpreted properly, it can help traders optimize and improve their strategies, find any technical or theoretical flaws, as well as gain confidence in their strategy before applying it to the real world markets

As Mauser has stated, other measures have been well studied both retrospective and prospective…there has been very little correlation with stock price.

That holds as a screeing tool or otherwise.

If I understand Saul’s investing premise correctly, he has stressed the difference between RB and his process largely on the basis of two themes:

  1. he is willing to sell a stock
  2. his landscape of stocks have lower 1ypeg than the RB

If that is true, then your circuitous post would seem merely argument for argument sake because many people on this board are using the 1yoeg as an important determinant either justifying their present holdings or why they don’t hold a particular stock.

1 Like

If that is true, then your circuitous post would seem merely argument for argument sake because many people on this board are using the 1yoeg as an important determinant either justifying their present holdings or why they don’t hold a particular stock.

Duma, I think you are the one being argumentative. You have put up a premise that backtesting is the superior way to test the validity of any trading system but haven’t performed backtesting on the 1YPEG. So since you think this is so important I am asking you to perform the backttest. If you can’t perform the test than how can you understand it? If you can’t understand it how can you say it is superior?

Andy
Who doesn’t believe that the 1YPEG is a trading system only 1 tool by which to pick stocks.

6 Likes

Andy:

The thread starts with a simple question…have you backtested the 1ypeg…pretty simple question really and since many (perhaps most) posters here seem interested in using it as a selection tool, it seems odd that You would view it argumentative to ask what validity it may have.

Why wouldn’t you want to know if the 1ypeg afforded you better returns or less risk?? Why wouldn’t you want to know if the tabulations and tracking of this indicator is of any value especially since it is a Saul inspired indicator that is not an “industry standard”.

And yes, back testing is a superior means to test a system particularly compared to blind faith and prospective chance. At the same time, a previous tested indicator may not guarantee future performance so nothing is fool proof.

Spend some time at the mechanical investment board. You can see efforts at trying to identify useful screens/tools to improve results.

I don’t use the 1ypeg…never have, but that doesn’t mean I don’t see the “argument” in favor of it over the PEG as Saul has explained…simply stated, it retrospectively states what actually is the valuation based on growth rate of trailing 12 month… I get the concept as compared to a guess about future growth. But since the PEG has not turned out to be useful, why should one expect the 1ypeg?

I don’t know the answer mind you, that was the question.

Since you use it, I am sure you will want to explore its usefulness. As you may know, there are quant programs such as these to drill down to specific indicators:

http://quantpedia.com/Links/Backtesters

1 Like

Duma,

The thread starts with a simple question…have you backtested the 1ypeg…pretty simple question really and since many (perhaps most) posters here seem interested in using it as a selection tool, it seems odd that You would view it argumentative to ask what validity it may have.

I do not find the question argumentative, what I find argumentative is saying that backtesting is a superior method to test a system without being able to do the backtesting. If someone claims to me that any method is superior at least they should be able to perform the test. Otherwise how can you have a discussion. I am completely open to any thesis you throw out Duma but at least you should be able to explain how you came to your thesis. If you come on the boards and state that backtesting is a superior method but can’t explain the method, and then tell me to go and figure it out myself, well I tend to ignore the person. Only because I have enough on my plate and can’t go out researching everyone else’s hypothesis. That is what makes this board important. People bringing ideas and companies to the board to share with everyone.

So what I would ask you Duma is go out and make yourself an expert on backtesting and then bring your findings to the board. That would be more useful.

Thank you,
Andy

3 Likes

So what I would ask you Duma is go out and make yourself an expert on backtesting and then bring your findings to the board. That would be more useful.

Now that was argumentative Andy.

I assume you are not asking this as a rhetorical question and that you do in fact know how to back test but are “testing” me?

1 Like

Now that was argumentative Andy.

I have no clue how to back test and have never tried to understand it Duma. That is why I am asking you to explain it to us and backtest the 1YPEG. Take us through the process.

Thank you
Andy

1 Like

Mauser:

I know that you developed some proprietary indicators and had back tested this to the market. I noticed that you now appear a bit more bearish on the market than you have been in the past couple years. Are your proprietary indicators flashing a downturn in the market?

One backtests because it is the only form of testing available, other than making predictions and patiently waiting for them to come true or not. But, backtesting is highly vulnerable to period selection and secular change. Pick a start date in 2008 versus one in 1999 and one is likely to get a very different result testing anything. Moreover, there are secular shifts in the market so that what might be a valid pattern in one decade may no longer be a valid pattern in the next.

So, yeah, backtesting is better than mere speculation, but at the very least, one should run the test over multiple periods to see what variations in pattern one gets.

I might also note that Saul’s metric is different than some related metric because it is confined to historical data and thus does not have the vulnerability of relying on estimates. But, by the same token, it is a metric about the past and what we want to know about as investors is the future. So, it would be small f foolish to rely on it alone.

7 Likes

But since the PEG has not turned out to be useful, why should one expect the 1ypeg?

Isn’t one of the big problems with a typical PEG that it is based on an estimate of future earnings, which is necessarily speculative? This is not the case with the 1YPEG.

3 Likes

many of the screens on MI have been back tested for decades, then tested for years post discovery so the likelihood of the results being random noise is very small.

The best screens seem to vary from year to year with no pattern .At least no pattern I have been able to find. So investing in one is hit or miss if your attention span is only a few years. It requires iron discipline , something most of us do not have…

When Saul saw his stock picks fall 60+%, it must have been hard not to sell. And since they were US stocks they recovered. If he had been Japanese investing in Japanese stocks in 2000, 15 years later he might still be losing money .

And it’s possible some things might “be different” this time- for instance globalization might adversely effected screens relying on differences between US and EU stocks.

Watching a few dice throws, by luck picking most of them right ,will not give the valuable information that backtesting of a few hundred throws will give you.
Back Testing is the only even semi reliable way to determine the probabilities of any system. But you can’t back test a system that has variable rules, that depends on changing human judgement, so you can’t back test the totality of what Saul does. Or what most investors do. You can only back test parts of it.

Check out SI Pro at the AAII site- there are a huge number of parameters listed. Digitalization of required SEC filings gives us a database that would have cost several million dollars only a few years ago. I can’t over stress that these mechanical systems do not always work, but they do increase the odds of success in some measurable way, and risk can be calculated right along with return.
I am a neophyte on MI ,and mostly used insights there to pick general market timing tools, transforming them from spread sheet based to chart based. Lots of work but mostly nothing very original.

One of my systems is " disruptive innovation investing", investing in companies involved in things like robotic surgery, electric cars, electron beam 3DP etc. Payoff can be huge but it takes years to know. This can not be automated or backtested because there are no numbers early on that tell you what is true disruption. Or who will wind up as the successful disrupters.

In the case of two I have now, TSLA and AMAVF, they are way ahead of everybody else. But could still fail. Payoff may be years away. but I think I will have the guts to hold these during a bear market . I believe in the story, the narrative, and that is not something you can put numbers on.

And for me an important part of this is that learning about these companies and their industries is more fun than it is work.
The older I get , the more quality time outranks money as a goal.

14 Likes

duma- all of my long term indicators are still bullish.

OTOH none of them pick the very top., they all lag to some extent. On average they will get me out after about 1/4 to 1/3 of the way from the bull top. And they work only in ordinary markets. They will fail with something like a big meteor hit on NYC or the crash of 1987.

Some of these systems are very simple. Bring up a monthly (not daily) chart of Std and Poor 500, Overlay with a simple moving average of 10 months. Sell when SMA is below the stock index ,buy when SMA is above. Test using other SMA.8, 9,11, 12, 13,14 months. Pick the one that you like. the longer the SMA the fewer whipsaws but profit is less too. I use both e 10 month and the 13 month SMA in a slightly modified form
(see post 243063 Mechanical Investing forum on motleyfool.com)

Because of the smoothing effect this works better than the widely known 200 day SMA of SP500 on a daily chart.
There is massive academic confirmation that trend following works with broad averages(AKA momentum)

This has been a long bull market without significant corrections so it makes me nervous.
But being nervous isn’t actionable.

The above chart is actionable -hedge by using it on only part of your portfolio.

Most of my methods do not work with individual stocks,they are designed for indices, but the bigger the number of individual stocks in a portfolio the more likely it is to work for them too. But the only way it can be backtested is for indices like SP500. Most of the market pros in mutual funds etc know of these methods. But because of institution restraints can’t use them. Good thing, because their machines are better than mine…

2 Likes

Mauser,
Could you give a link to the board?

Andy