3/28/2018

Visual Discovery 03/28/18

Visualizing over 5,300 stocks using the Magic Formula within a scatter plot for companies less than 2 billion.


The X-Axis represents EV/EBIT as a measure of “cheap companies”. The Y-axis attempts to rank “good companies” using the average ROIC over the past 4 years. If your search is only cheap companies. Just focus on the X (EV/EBIT) and ignore or filter out the Y (ROIC) measure. 

The upper left quadrant clusters good companies (high ROIC) and cheap companies (low EV/EBIT) and with the same color. The lower left is cheap companies but not as “good” (high ROIC) as the upper left quad.

Plug and play to find companies reducing shares outstanding from 2015 to the MRQ (% Chg Shrs Outs 2015 vs MRQ) or total liabilities(Ttl LIAB MRQ % Chg vs 2015). Further, you can modify Industry, financial strength, revenue growth, exchange or market size.


Note the default filter is currently set where both EV/EBIT (Cheapness) and ROIC (Goodness)  greater than zero. Bubble size is determined by the % off its 52 week high. Updated (03/30/18) the tool tooltip too contain additional decision-making data. See the new measures by hovering over a bubble. Additional improvements to be added.  

  




9 comments:

Ben's Jamin said...

This is awesome. Btw the irony of magic formula, as Carlyle found, is that removal of ROIC filter actually led to higher returns than without.

Chris Emiot said...

Can you change the EV/EBIT from a slider ? It's a little hard to fine tune :)

Anonymous said...

How did you even do this?

So cool.

What's your calculation of ROIC?

Lots of people use different inputs....

ShadowStock said...

Thanks for taking the time to look!

I agree the ROIC is highly mean reverting. So I took the ROIC over the past 4 years and averaged. But just look at the X axis (EV/EBIT) to the far left to review low EV/EBIT (cheap companies) without reviewing ROIC.

Ben's Jamin said...

Yeah that is truly awesome.

Hang on though, where does the data pull from? Like what universe? (p.s. If you'd rather not reveal, that's okay).

ShadowStock said...

Blogger Chris Emiot said...
Can you change the EV/EBIT from a slider ? It's a little hard to fine tune :)

Hi Chris,

You can just key in the range of values. For example key in 1 then 8 as the upper limit of the EV/EBIT. Let me know.

ShadowStock said...

Anonymous Anonymous said...
How did you even do this?
So cool.
What's your calculation of ROIC?
Lots of people use different inputs....

Hello Anon,
Actually, I’m adding today the option for “Good Companies” ROA, return on tangible assets. GP/Assets, so there will be “drop down” to change the Y or X axis metric.
Most of my data scrapped using a script. That scrapped data is cleaned, then improved by adding more calculated measures such as a change in shares outstanding, total liabilities from 2015 as an example. This is useful if you want to eliminate companies diluting shareholders with equity or debt. Or create a field in the table where I parsed out the F score. For example, Yes or No YOY shares outstanding reduced or reduced leverage YOY.
ROIC uses EBIT / Book Value – Cash + Long-term debt. But I wouldn’t get too caught up in whether you use ROIC, ROA or GP/Assets as a measure of a good company. Personally, I ignore ROIC and focus more on the cheapness and price performance as the change in enterprise value over time. I’m more deep value so I focus on cheapness and other mean reverting metrics to find stocks.

ShadowStock said...

Yeah that is truly awesome.
Hang on though, where does the data pull from? Like what universe? (p.s. If you'd rather not reveal, that's okay).





Ben, I run automated scripts that scrap various websites with financial data, etc. Then all these saved text files are cleaned, improved with new metrics and consolidated before developing something with Tableau public(free). My universe was for companies below 2 billion, price >.10 (I think), and excluded banks, REITs, and Utilities

Chumpmenudo said...

Wow, this is really great work, and useful for screening. Will you be updating the data to the latest quarter, is was this a one time thing?

Best,

Chump