Skip to main content
No. of Recommendations: 2
Comparing Mungo’s old VGHorse slightly modified to the elann’s version of ROE_Cash screen.
Elann’s ROE_Cash screen is a hold 30 where VGHorse is usually run as a 20 day hold 5 or 10 but
even if you Intend to only hold a small number if the screen works for a larger hold there is more
assurance that the parameters are less likely to be data mined. VGGorse uses SIP fields only available
back to 1997. I modified VGHorse into a top 30 hold till drop 45 to try and better match the ROE_Cash
low turnover numbers. For comparisons both screens are run from 1997, 42 day period
and hold 30.

         ROE_Cash  Horse
CAGR 14.6 24.9
TR 2142 16415
SAWR 8.6 11.4
GSD 21.7 27.8
DIGSD 26.4 30.1
LDD 11.6 15.2
LDDD3 11.7 10.5
MDD -49.6 -47.4
UI(126) 14.0 12.0
Sharpe 0.7 1.0
Beta 1.1 0.8
TI 12.7 27.7
AT 1.1 5.7


And their yearly returns:
          ROE_Cash  Horse
19971231 5.3 4.5
19981231 63.0 31.6
19991231 58.6 110.6
20001229 -6.9 46.7
20011231 2.1 58.3
20021231 -17.2 13.7
20031231 37.7 133.7
20041231 17.1 50.0
20051230 7.8 28.3
20061229 13.9 48.4
20071231 12.6 36.7
20081231 -36.5 -33.7
20091231 43.3 37.7
20101231 22.1 25.2
20111230 2.5 -5.7
20121231 11.5 14.5
20131231 36.8 64.0
20141231 16.5 -0.5
20151231 -2.5 -6.4
20161230 15.6 22.9
20171229 32.5 14.2
20181231 0.3 -11.2
20191231 32.6 11.2
20200710 17.0 -9.6

http://gtr1.net/2013/?~ROE_Cash_noFinance_with_crap_filters:...

http://gtr1.net/2013/?~VG-Horse-T30HTD::styp.a:ne14!15!44!45...


Horse wins the CAGR and SAWR by a neck over this shorter 20 year period but
with a significantly higher GSD and turnover.

RAM
Print the post Back To Top
No. of Recommendations: 0
http://gtr1.net/2013/?~VG-Horse-T30HTD::styp.a:ne14!15!44!45......


Horse wins the CAGR and SAWR by a neck over this shorter 20 year period but
with a significantly higher GSD and turnover.


Would you care to remind us what VGHorse screens for? The screen definition is such spaghetti that I can't easily figure it out.

Elan
Print the post Back To Top
No. of Recommendations: 4
Unfortunately my faulty memory attributed VG-Horse to Mungo when it actually came from tpoto’s
observation of a screen where he states “I did not develop this one and have seen other slight
variations, but this one as is seems to do the trick with the above variations.” Which I believe he
is stating that he saw a similar screen run on Keelix’s and is presenting his modified version.

back in 2013 were he presented a new VG-SCRN 1/25/2013 post 241235.
https://boards.fool.com/new-screen-votes-needed-30505987.asp...
In a 95 thread different observations and modifications are suggested resulting in VG-Zebra and
VG-Horse. Can’t depend on my memory any more, I even suggested a couple of change in the original
thread and forgot the true discoverer.

Elan: I don’t have all the screen factors memorized I typically download the GTR *.s field definitions and then use my Stock Investor Field definition Help screen. Like: EPSQ5=Earnings Per Share 5 quarters ago,Q1 is last quarter Sales is self-evident,PFCF = Price Free Cash Flow, . . .

RAM
Print the post Back To Top
No. of Recommendations: 2
Horse wins the CAGR and SAWR by a neck over this shorter 20 year period but with a significantly higher GSD and turnover.

But in the CAGR comparison, VGHorse loses to ROE_Cash each and every year since 2009. That'd be hard to stick with.

(I clicked on RAM's links, but don't know how to get GTR1 to restrict the testing period to start on 1/1/09 so I could see what the stats are for ROE_Cash's recent (and consistent) out-performance.)

Tails
Print the post Back To Top
No. of Recommendations: 4
Would you care to remind us what VGHorse screens for?

The last time I used VG-Horse was in 2016, but the steps are roughly:

1. Industry not equal to "Financial"
2. YoY EPS must be positive
3. YoY sales must be positive
4. Price/Free Cash Flow per share <= 20
5. Institutional Ownership <= 70%
6. Market cap >= $100M
8. Price >= $8
9. Positive 6-month return
10. Avg daily volume >= 30K
11. Calculate VR = ((6-month return)/(Avg monthly volume))*(Avg daily volume*21)/(Avg monthly volume)
12. Sort descending by VR, keeping the top 10

You can get the Radiscript for VG-Horse and other MI screens from lohill's gtr1 data helper program.

VG-Horse
Keep :[SI Sector]<>"07-Financial"
Keep :[SI EPS Q1]>[SI EPS Q5]
Keep :[SI Sales Q1]>[SI Sales Q5]
Keep :[SI Price/FCFPS]<=20
Keep :[SI Float]<20
Keep :[SI Institutional Ownership %]<=70
Keep :[SI Market Cap Q1]>=100
Keep :[SI Price]>=8
Keep :[SI Price Change 26 week]>0
Keep :[SI Volume--Average Daily 10d]>=30
Create [VR] :21*[SI Volume--Average Daily 10d]/[SI Volume--Average Monthly 3m]
Create [SORT] :([SI Price Change 26 week]/[SI Volume--Average Monthly 3m])*[VR]
Sort Descending [SORT]
Top :10
Print the post Back To Top
No. of Recommendations: 3
Note VG-Horse and ROE_Cash from the VG post discovery date performance (seven and a half years
from 1/25/2013):

VG-Horse CAGR 7.5% GSD 16.9 compared to ^GSPC’s 10.5% GSD 16.5. . . Poor to say the least.

ROE_Cash CAGR 17.5 GSD 11.5 Beat the S&P by 7%, looks great like all the other screens when they
were initially discovered but how many post discovery winners do we still have.

RAM :<(
Print the post Back To Top
No. of Recommendations: 0
Has anyone tested a final sort to run the ROE_Cash as a more concentrated screen (using our usual top 4,5, or 10)?

Just curious if there is some performance improvement (at the cost of higher volatility) to be had by running a final RRS sort or some other such thing.

--G
Print the post Back To Top
No. of Recommendations: 1
Has anyone tested a final sort to run the ROE_Cash as a more concentrated screen (using our usual top 4,5, or 10)?

Just curious if there is some performance improvement (at the cost of higher volatility) to be had by running a final RRS sort or some other such thing.


See message 278153
Print the post Back To Top
No. of Recommendations: 0
I've used VGH off/on since '13 or '15 as 1/3 of a blend... in the first year or so it was a top screen but then some factor(s) fell out of favor so it has been tepid since.
Print the post Back To Top
No. of Recommendations: 5
Horse wins the CAGR and SAWR by a neck over this shorter 20 year period but with a significantly higher GSD and turnover.

But in the CAGR comparison, VGHorse loses to ROE_Cash each and every year since 2009. That'd be hard to stick with.

(I clicked on RAM's links, but don't know how to get GTR1 to restrict the testing period to start on 1/1/09 so I could see what the stats are for ROE_Cash's recent (and consistent) out-performance.)


This caused me to question, what if the performance of ROE_Cash since 2009 was only due to its mega-cap tilt?

I ran the ROE_Cash backtest from 2009 to present, and compared it to the S&P index, and to a screen that selects the top 30 stocks by market cap, HTD 45 (with the same crap filters as ROE_Cash).

I got the following CAGRs -
S&P 11.18
MCap 15.35
ROE_Cash 18.41

So more than half the advantage of ROE_Cash over the S&P index came from simply selecting the very largest stocks by market cap. Will the outperformance of mega caps that we've seen recently continue into the future? If not, the ROE_Cash screen may not perform so well.

Elan
Print the post Back To Top
No. of Recommendations: 3
I've used VGH off/on since '13 or '15 as 1/3 of a blend... in the first year or so it was a top screen but then some factor(s) fell out of favor so it has been tepid since.

Yes, the entire category of value stocks fell out of favor, and the entire category of small cap stocks fell out of favor.
Print the post Back To Top