Jims ROE and 5 yr sales growth rate screener

I ran Jims ROE and 5 yr sales growth rate screener on value line, and was hoping to see if I got this correct.

POST number - 283472

https://discussion.fool.com/i-tried-the-strategy-and-probably-be…

  1. For some reason, I am never able to get all the 1700 stocks in the universe. I tried to run the classic stock screener on its own (with the 3 auto selected labels{International -ALL, Industry - ALL, and Valule line preset screens ALL] - However, that doesnt return anything!

  2. So, I chose all the 5 timeliness, and that produced about 1398 stocks

  3. Then chose the 1200 close to 52 week high.

  4. Then only chose those with the ROE and 5 year sales growth available (many dont have one or the other, despite being highly ranked on one of the criteria { for example Godaddy etc}

  5. And chose the 50 with top ROE; 50 with top 5 year sales growth rate; And then added the sum of ranks for these. From those, I chose the top 20.

Cheniere Energy Inc. LNG
Medifast, Inc. MED
Fortinet Inc. FTNT
Louisiana-Pacific LPX
Golden Entertainment GDEN
Lam Research LRCX
Gartner Inc. IT
Credit Acceptance CACC
AbbVie Inc. ABBV
MasterCard Inc. MA
Apple Inc. AAPL
Dick’s Sporting Goods DKS
Blackstone Inc. BX
Regeneron Pharmac. REGN
IDEXX Labs. IDXX
QuidelOrtho Corp. QDEL
CDW Corp. CDW
NVIDIA Corp. NVDA
lululemon athletica LULU
Sherwin-Williams SHW

Would greatly appreciate if Jim or anyone else with this screener can check if this is correct?

Thanks so much

Charlie

8 Likes

2. So, I chose all the 5 timeliness, and that produced about 1398 stocks

The count at that step sounds about right.
I got 1397, but that’s possibly because one of us is using last week’s data and the other is using next week’s data.
Close enough for government work.

Why so few? Not all of the stocks in the standard “1700 stock” edition have Timeliness ranks.
The average is 1517 over time; it’s a little lower than usual right now.
Stocks are excluded from that ranking if they have too few years of history, or of they are in the midst of a major M&A event, or bankruptcy.

And for that matter, the “1700” edition doesn’t have 1700 stocks.
It’s only 1545 at the moment; I’m not sure why.
It was 1710 last June and has been gradually falling, pretty much steadily. 1663 at the start of the year.
My guess is that they’re short of staff and haven’t yet gotten around to adding replacements for the ones that have disappeared through M&A or delisting.

Jim

3 Likes

I gave it a whirl.

The obvious difference between your method and mine in this case is requiring that there be a Timeliness Rank.
I can backtest that as well…it backtests a hair worse.
Slightly lower long run return, slightly lower chance of any one pick being a winner, etc.
But it’s not a night and day difference so I wouldn’t lose sleep.

It does mean different picks.
Here are the top 25 picks I get at the moment, NOT requiring a valid Timeliness rank.
CQP LNG FTNT RH LPX LRCX IT ABBV MA CACC KLAC ULTA
AAPL AMAT BX DKS QDEL REGN IDXX ALSN CDW BLDR NVDA TSCO CNR
Done by hand, so an error is always possible.

Here are the top 25 picks if I require a valid Timeliness rank as a first step.
CQP LNG FTNT RH LPX LRCX IT ABBV MA CACC KLAC ULTA
AAPL AMAT BX DKS QDEL REGN IDXX ALSN CDW NVDA BLDR TSCO LULU
Again, done by hand, so an error is always possible.

Requiring Timeliness or not makes only a couple of differences at the end: ranks 1-22 are identical.

So, 6 differences is still quite a few.
Then we have to start speculating why.

As mentioned before, it’s possible we are using a different week of data.
Try yours again in a couple of days to see.

But that doesn’t seem to be enough to create 6 differences.

One difference may be your step 4?
I generate the ROE ranks from all those with ROE, whether they have sales growth figures or not.
I generate the sales growth ranks from all those with sales growth, whether they have ROE figures or not.
Then I sum the ranks.
But the number of stocks with one and not the other is very low. All of the top 1000, give or take.
So…that’s probably not it.

Jim

9 Likes

Great, thanks so much.

You said " I generate the ROE ranks from all those with ROE, whether they have sales growth figures or not. I generate the sales growth ranks from all those with sales growth, whether they have ROE figures or not. Then I sum the ranks."

So, if stock A had a ROE rank of 1, but no sales growth figure, what score do you give for the sales growth? Do you just give it a score of “0” for the sales growth? And so that total Sum of ranks is just 1 +0 = 1 ?

Thanks a lot

Charlie

So, if stock A had a ROE rank of 1, but no sales growth figure, what score do you give for the sales growth?

Infinitely negative replacement value, or infinitely positive rank number.
Assuming you’ve sorted them so that rank 1 = biggest best number, you want the ones with the lowest sum of ranks.
This will make anything with missing data sort to the end.

Can it make a difference whether you rank them separately, or require both fields to be populated before sorting?
I’m not sure, but I know such questions can be a bit non intuitive.
For example, consider the requirement for Timeliness.
It doesn’t change the list much this week, no change in the top 21 picks.
But, surprisingly, it does change the sort order of two stocks which both have Timeliness and are in both lists.
(BLDR and NVDA).

Jim