Post New

Thought this article would be of interest to the MI board.
I couldnt help wonder if Jim or one of the other prolific screen creators over the years had proposed and backtested something like the strategy discussed in this article
I also found the discussion about how these factors tend to work for a while, and then they dont so much, illuminating.
I recall a whats working now analysis of screens - which this article would lend support to.
Enjoy
https://pracap.com/whats-driving-stocks

I couldn't help wondering if Jim or one of the other prolific screen creators over the years had proposed and backtested something like the strategy discussed in this article

There is a screen called Up 5% that is based on earnings revisions, but I don't know how well it has done over the years.

DB2

One of the RevUp screens on GTR1 is Up5X3:
https://gtr1.net/2013/?!!QlpoMTFBWSZTWZb2hSQAAi1fg...
From 1997 till present CAGR is 18.8% with a SAWR of 10.3%, sounds nice but screens like this are very
Volitle.

Year	          Avg
20151231	-5.1
20161230	38.0
20171229	15.6
20181231	-7.2
20191231	9.9
20201231	1.9
20211231	52.5
20221230	-10.6
20231229	25.8
20241119	1.4

A similar Screen I used briefly several years ago over on P123 has similar results from 2002
CAGR 19.3%
Annual Turnover 483%

Year	EstRev	SPEW	Excess
2015	9.17	-2.67	11.84
2016	15.54	14.51	1.04
2017	58.58	18.52	40.06
2018	-30.13	-7.84	-22.29
2019	42.25	28.91	13.35
2020	14.86	12.66	2.2
2021	36.32	29.41	6.91
2022	-26.44	-11.62	-14.82
2023	6.45	13.7	-7.25
2024**	33.57	16.06	17.52

A rough ride in 2018 and 2022.

A recent post being discussed on another form with some interesting observations about Earning Revisions and short term overreaction compared to long term fundamentals.
What's Driving Stocks - Praetorian Capital
https://pracap.com/whats-driving-stocks/
A Quote from the article:
“Over the past two decades, quant shops have attracted a tremendous amount of assets — especially the pod shops that use many of these factors as the basis of their trading. There can be no denying that today there are way less fundamental-type managers. Heck, some traditional long-only funds even have quant screening applied to their list of securities they are able to purchase. These factors are increasingly driving portfolio construction.”

A bit late to the thread, but thanks for posting that link. Quite a good read.

A screen strategy springs to mind: don't look for positive earnings revisions per se. Look instead for the transition from a trend of negative earnings revisions to positive ones, with a layer for good longer term prospects.

The idea would be to try to capture the short-term-hiccup-long-term-great firms just as their stretch of being out of fashion is ending and turning around. Then you are holding during a stretch that BOTH factors are going for you: good prospects and probable undervaluation from being oversold, and current earnings revision momentum. If you held till old fashioned price momentum fell off (while earnings revision momentum still good), you could ride it all the way from underloved to overvalued, then bail.

As a bonus, this would probably be a longer average hold than a pure earnings revision screen.

Jim

A bit late to the thread, but thanks for posting that link. Quite a good read.

A screen strategy springs to mind: don't look for positive earnings revisions per se. Look instead for the transition from a trend of negative earnings revisions to positive ones, with a layer for good longer term prospects.

The idea would be to try to capture the short-term-hiccup-long-term-great firms just as their stretch of being out of fashion is ending and turning around. Then you are holding during a stretch that BOTH factors are going for you: good prospects and probable undervaluation from being oversold, and current earnings revision momentum. If you held till old fashioned price momentum fell off (while earnings revision momentum still good), you could ride it all the way from underloved to overvalued, then bail.

As a bonus, this would probably be a longer average hold than a pure earnings revision screen.

I'd say this is a very good stock picking investing strategy, but I doubt that it's practical as a mechanical screening strategy. We're looking here at something that requires a dive into a company's business cycle, which may happen on a different time scale for different companies. Why was it down? Have its business prospects turned a corner? Does it have potential for continuing to surprise to the upside? Those are all questions that fundamental company analysts try all the time to address. Many people do this successfully, I believe. But it involves hard work as opposed to mechanical screening.

Elan

I looked at buying on a surprise after the prior report being negative, and it was not as good as when both were positive.

Here is a new monthly screen using the last two eps surprise periods that trades monthly and substantially outperforms the S&P since 12/17/2008 (the date gtr data exists. It outperforms on:
CAGR, by 10.7% (25.88% to 15.1% for the S&P) ,
sharpe (1 to .81 for the S&P),
beta (1 to 1.1 for the S&P),
LDD (10.9 to 12.7 for the S&P),
Treynor (23.5 to 13.3 for the S&P)
but not
GSD (29 to 20 for the S&P)
or max drawdown (-45.3% to -39.5% for the S&P)

CAGR 25.88
GSD 28.97
SHARPE 0.99
BETA 1

price > 2
ave dollar volume adv (1,63)>$500,000
[[SI Quarterly Surprise-Est; lag=1 days]/[SI Qtrly Surprise-Prior Qtr-Est; lag=1 days]] > 1.2 ratio(qseps.s,sq2eps.s) > 1.2
[SI Qtrly Surprise-Prior Qtr-Est; lag=1 days] > 1 sq2eps.s>1
top 30 by price change last 200 days tr(1,200) top 30
5 day simple moving average>1,1 x 200 day sma ratio(aprc(1),sma(1,200))>1.1
5 day return bottom 5 tr(1,5) bottom 5

It performs even better with price>5 and adv>1,000,000 if you want even more liquid stocks

https://gtr1.net/2013/?aprc%281%29gt5:adv%281,63%2...

cagr 26.6

v.

https://gtr1.net/2013/?aprc%281%29gt2:adv%281,63%2...

cagr 25.8

Musselmant

The screen certainly has a good return, but somewhere in my distant memory, I remember Robbie saying you shouldn't mix aprc with sma. aprc is actual price and sma is based on gprc.

5 day simple moving average>1,1 x 200 day sma ratio(aprc(1),sma(1,200))>1.1

On the screener ZIM is number 2 (I just happen to be familiar with this stock as it is on my YEY screen)

The screener says aprc is 24.23 which is the close on Nov 21. The sma(1,200) is 3.86. However, the lowest adjusted close is about 8.9. So the sma is not in adjusted close units.

Looking at the 3rd stock NMM, aprc is 51.38 and sma(1,200) is 0.62

So to eliminate the problem, I eliminated the step.

https://gtr1.net/2013/?aprc%281%29gt5:adv%281,63%2...

CAGR 23%
MDD -44%
SAWR 14%

Aussi

thanks; cagr 69% this year, and 20% last with correction

And with a little bit of "harmless" data mining.

Hold 15 HTD 25 for 15 days without the short term tr(1,5) step. Also revised tr(1,200) to include tr(1,175) and tr(1,225)

CAGR 23%
SAWR 16%
MDD 37%

2023 22%
2024 72%

https://gtr1.net/2013/?s20081208h15::aprc%281%29gt...

Aussi

My apologies for my novice question.

Can the picks from this screen be obtained directly from SIPRO, without using GTR1?

Thank you very much.

Juan Carlos

“Can the picks from this screen be obtained directly from SIPRO, without using GTR1?”

The short answer is No but you do need a paid subscription to SIPRO to obtain a password to
actually, get the picks to actually invest in the screen.

A longer answer: From SIP you can screen directly for price (in GTR aprc(1) > 5, use the
SIP Volume-Dollar Daily Avg 3m for the GTR adv(1,63) > 63
SIP also has the eps data (all SIP fields are indicated by the *.s format)
But GTR1 also has a data base of daily price for all the stocks which it uses to calculate
linear(1,tr(1,175),1,tr(1,200),1,tr(1,225)) top 25

You could of course use SIP for steps 1 thru 3 which typically results in around 65 picks and
then download the daily price data for those 65 stocks over the last 225 days into a spread sheet
and then calculate total returns for step 4 and step 5.

Or you could be thankful that Robbie has done all the work to reduce your effort by an order of
magnitude for free! And just use GTR1.

Thank you very much for your detailed explanations.

I am immensely grateful to Robbie for his extraordinary work. In that, there is no doubt.

On the other hand, perhaps this screen has sufficient merit to be included in the weekly "SIPRO Rankings".

Thanks again.

Juan Carlos

There are a few reasons to wait before adding any screen to a permanent list. Maybe wait until at least 1 year post discovery results are in. (What is the range of expected excess returns over the next 1 year?) Also, there might be alternates proposed, and having multiple similar screens makes the permanent list cluttered. And the SIP data has just been changed to a new vendor, which is another reason to wait a year.

I looked at buying on a surprise after the prior report being negative, and it was not as good as when both were positive. Here is a new monthly screen using the last two eps surprise periods...

The screen uses qseps.s which is one of three Quarterly Surprise (a.k.a. announcement day) fields that could be used:

epslr.s [SI Quarterly Surprise-EPS]
qseps.s [SI Quarterly Surprise-Est]
qsdiff.s [SI Quarterly Surprise-Difference]

qseps.s is the quarterly earnings estimate for the latest reported fiscal quarter (i.e. the EPS estimate on announcement day). The screen ignores actual EPS, and instead uses past estimates; looking for qseps.s 20% higher than the previous quarter.

A 20% increase in epslr.s is 20% earnings growth.
A 20% increase in qseps.s is 20% estimates growth.
A 20% increase in qsdiff.s is 20% growth in earnings surprise.

==== SIP field descriptions ====
Quarterly Surprise-EPS
Data Table Name: EPS_LR
The latest reported quarterly earnings. The quarterly period matches the Quarterly Surprise-Est period.

Quarterly Surprise-Est
Data Table Name: QS_EPS
The quarterly earnings estimate for the latest reported fiscal quarter.

Quarterly Surprise-Difference
Data Table Name: QS_DIFF
The difference between the actual reported earnings and the consensus earnings estimate for the most recently completed fiscal quarter. It is calculated in the following manner: Actual (announced) EPS – EPS Estimate

Post New

US Policy FAQ

---

Contact Shrewd'm
Contact the developer of these message boards.