Hi, Shrewd!        Login  
Shrewd'm.com 
A merry & shrewd investing community
Best Of MI | Best Of | Favourites & Replies | All Boards | Post of the Week!
Search MI
Shrewd'm.com Merry shrewd investors
Best Of MI | Best Of | Favourites & Replies | All Boards | Post of the Week!
Search MI


Investment Strategies / Mechanical Investing
Unthreaded | Threaded | Whole Thread (24) |
Post New
Author: mi4me   😊 😞
Number: of 3959 
Subject: RunRadisJM
Date: 06/13/2023 5:18 PM
Post Reply | Report Post | Recommend It!
No. of Recommendations: 8
I've created a Java version of RunRadis that uses memory-mapped files and, as a result, runs significantly faster than RunRadis. Source code can be downloaded from github:

https://github.com/jrh3/RunRadisJM/tree/1.0

-Jim
Print the post


Author: TGMark 🐝  😊 😞
Number: of 3959 
Subject: Re: RunRadisJM
Date: 06/13/2023 10:55 PM
Post Reply | Report Post | Recommend It!
No. of Recommendations: 0
Is RunRadis the screening tool of choice these days? Or is Radiscreen still widely used?
I'm partial to spreadsheets, but RunRadis sounds good.
Print the post


Author: RAMc   😊 😞
Number: of 3959 
Subject: Re: RunRadisJM
Date: 06/14/2023 1:16 AM
Post Reply | Report Post | Recommend It!
No. of Recommendations: 5
Is RunRadis the screening tool of choice these days? Or is Radiscreen still widely used?

I have used both for many years, both are good but. . . To the best of my knowledge Radiscreen
Was conceived and written by Phil . .The Great Radish .. in 2003

He defined the language format, wrote Excel macros import selected VL IBD and later SIP data into excel.
Radiscreen to run screens. It was a great step forward and it still works but in my limited experience
Runradis is a stand alone program written by mi4me to run screens using the Radiscript screen format that Radish defined. It runs a screen or screens from a list very quickly directly from your SIP download.

I see above that Jim mi4me has written a new version. Without question that will be the one to use. I browsed the source code of the original written in Ocaml. The coding was a class above what I wish I was capable of.

RAMc
Print the post


Author: TGMark 🐝  😊 😞
Number: of 3959 
Subject: Re: RunRadisJM
Date: 06/14/2023 10:28 AM
Post Reply | Report Post | Recommend It!
No. of Recommendations: 0
Thank you. I've been using GTR1 for a long time but since that's not available at the moment, have to do something else.
Amazingly, all the old Radiscreen spreadsheets on my computer still work. I'll figure out how to get RunRadis going too.
Print the post


Author: mi4me   😊 😞
Number: of 3959 
Subject: Re: RunRadisJM
Date: 06/14/2023 10:36 AM
Post Reply | Report Post | Recommend It!
No. of Recommendations: 4
Thanks for that, RAMc!

RadisScreen is used to identify stock selections and includes all of the functionality you get with Excel. RunRadis can print stock picks straight from your SI Pro installation, but only supports a limited set of Excel functions.

Nevertheless, RadisScreen cannot perform back-tests (AFAIK), while that's the primary purpose of RunRadis - it performs back-tests via a DB built from multiple SI Pro installations. RunRadisJM is also intended for running back-tests, and in fact, it currently has no support for listing stock picks from a screen (unless you add "Print" statements), nor does it work directly on an SI Pro installation.

-Jim
Print the post


Author: TGMark 🐝  😊 😞
Number: of 3959 
Subject: Re: RunRadisJM
Date: 06/14/2023 11:18 AM
Post Reply | Report Post | Recommend It!
No. of Recommendations: 1
Nevertheless, RadisScreen cannot perform back-tests (AFAIK), while that's the primary purpose of RunRadis - it performs back-tests via a DB built from multiple SI Pro installations. RunRadisJM is also intended for running back-tests, and in fact, it currently has no support for listing stock picks from a screen (unless you add "Print" statements), nor does it work directly on an SI Pro installation.

OK, so for screening, sounds like RunRadis is preferred.
Print the post


Author: lizgdal   😊 😞
Number: of 3959 
Subject: Re: RunRadisJM
Date: 06/14/2023 12:53 PM
Post Reply | Report Post | Recommend It!
No. of Recommendations: 3
RunRadis sounds amazing. I wish AAII would coordinate some of this work, so the 2 to 3 hour task of moving individual files into a DB could be streamlined. SIP has one of the lowest prices for historical PIT fundamental data, with prices similar to those 10 years ago. SIP data is PIT, with full snapshots of the data on each date. This was a fortunate accident produced by the simple approach of duplicating all the data every day. AAII is not advertising SIP, and so I doubt AAII knows the value of what it has recorded.

Point-in-time data is difficult to reproduce, and so most backtesters (e.g. P123, Equities Lab) use something like "filing date plus 4 days". That will often be good enough, but sometimes will fail. Some screens are more affected by non-PIT backtests, especially complicated tuned microcap screens. Of course, PIT data vendors have white papers saying PIT data is preferred. I might sign up for P123, and run some benchmarks. Or maybe I will just use the free Stock Rover screener.

current prices:
SIP 3-Year Subscription is $599
P123 Screener is $299/year (5Y backtest)
SimFin Basic is $420/year (15Y backtest)
P123 Pro is around $1000/year (20Y backtest)
Equities Lab Basic is $1200/year (28Y backtest)

--- links ---
prices from about 10 years ago:
"Portfolio123 is now 995/year for 5 years historical,
1995/year for a little less than 11 years historical data.
Stock Investor Pro $247 (non AAII member)"
http://www.datahelper.com/mi/search.phtml?nofool=y...

Equities Lab data:
"Our data changes daily, and covers each trading day from January 1, 1995 to present. Morningstar takes a few days to clean and process the data they get from the SEC, while they update the price data after the close of the market... Equities Lab uses the SEC release date, adds four days to account for processing delays, and makes the data available then."
https://www.equitieslab.com/features/comprehensive...

P123 data:
"If you want a more conservative backtest approach you can choose to 'Exclude Prelims' which will expose the data on the File Date + one day. NOTE: this may still suffer from a minor look-ahead bias since Factset may have taken a few days , or even weeks , to process the filing. We hope to revisit this later when we learn more about typical delays."
https://community.portfolio123.com/t/factset-beta-...

SimFin data:
"SimFin extracts fundamental data from original quarterly and annual reports. Our innovative AI algorithms have been improved and trained over years to achieve the current accuracy."
https://www.simfin.com/en/fundamental-data-downloa...

PIT whitepaper:
"Our research finds that PIT backtests produce significantly different results than lagged Non-PIT data using common factors. In addition, when single factors are combined into multifactor tests these differences in results may become magnified."
https://www.spglobal.com/marketintelligence/en/new...

PIT whitepaper:
"If data is not recorded point-in-time, we could be using 'future' data in our backtest, This will pollute the analysis with hindsight bias, which can artificially make results look better."
https://www.refinitiv.com/perspectives/future-of-i...

15 Best Stock Screeners of 2023 (Free & Paid)
Standalone stock screeners (the primary tool is the screener):
FINVIZ
ChartMill
Trade Ideas
StockFetcher
Integrated within platforms for investors:
Stock Rover
Seeking Alpha
Simply Wall St
Zacks
Integrated within technical analysis platforms:
TrendSpider
TradingView
Integrated with trading platforms:
TC2000
StocksToTrade
MarketGear
Integrated within platforms suitable for both traders and investors:
Atom Finance
Yahoo Finance
https://finmasters.com/best-stock-screeners/#gref
Print the post


Author: RAMc   😊 😞
Number: of 3959 
Subject: Re: RunRadisJM
Date: 06/14/2023 4:17 PM
Post Reply | Report Post | Recommend It!
No. of Recommendations: 3
Lizgdal writes:
I wish AAII would coordinate some of this work, so the 2 to 3 hour
task of moving individual files into a DB could be streamlined.


I took me months to put SIP data from '97 into a database. It's not hard to automate all the individual SIP monthly then weekly data from AAII. The problem for me was automating the individual install programs, the required interaction 'select / skip update' Then for windows 'do you want .. . This interface changed over time, many data fields were dropped, added, definitions changed, company id's changed, sector and industry codes changed and even a few supposedly unique company id being reused. The fact that Robbie kept up with all this in a first class way, only occasionally asking for small amounts of help to partially pay for his expenses but never rewarded for his labor is amazing.

However, I believe the advantages simple screens based on comparison to a linear combination of fundamentals is being arbitraged away by the big boys using much more sophisticated data analysis tools (ML).

If I could add one feature to radiscript it would be the capability of doing sum of factor rank. An example of what this can do for screens was posted by borisnand back in 2018. His example using GTR1:
where [rAll] is sum of the ranks of three fields:
[rAll]: [-1*[rA] + 1*[rB] - 1*[rC]]
where:
[rC]: [Rank by [SI PE-Forward EPS Est Y1] (Asc) at step6]
[rB]: [Rank by [Total Return % over 252 days] (Asc) at step6]
[rA]: [Rank by [Average dollar-volume over 63 days] (Asc) at step6]

http://www.datahelper.com/mi/search.phtml?nofool=y...
Simple screen from'97 to 3/23 CAGR = 26.95%

RAMc
Print the post


Author: lizgdal   😊 😞
Number: of 3959 
Subject: Re: RunRadisJM
Date: 06/15/2023 2:30 PM
Post Reply | Report Post | Recommend It!
No. of Recommendations: 3
Radiscreen can RANK.

Define {rankPefDvMo2018}
Uses [Price Change 26 week] [DolVol] [Country] [ADR/ADS Stock] [Price as % of 52 Week High] [PE-Forward EPS Est Y1]
Deblank [Price Change 26 week] [DolVol] [Country] [ADR/ADS Stock] [Price as % of 52 Week High] [PE-Forward EPS Est Y1]
Keep :[Price as % of 52 Week High]>88
Keep :[PE-Forward EPS Est Y1]>3
Keep :[Country]="USA"
Keep :[ADR/ADS Stock]="F"
Create [r1] :RANK([DolVol],[[DolVol]])
Create [r2] :RANK([Price Change 26 week],[[Price Change 26 week]])
Create [r3] :RANK([PE-Forward EPS Est Y1],[[PE-Forward EPS Est Y1]])
Create [rAll] :[r1]-[r2]+[r3]
FinalSort Descending [rAll]
End

"Since Radiscreen runs off excel spreadsheets, you can use a range of in-built excel functions, PERCENTRANK is one, and RANK is another. NB the RANK function reverses the parameters, so you have lines that look like:
Create [RkEE] :Rank([EBITDAtoEV],[[EBITDAtoEV]])
and then you can sort, filter etc on these custom columns."
http://www.datahelper.com/mi/search.phtml?nofool=y...
Print the post


Author: RAMc   😊 😞
Number: of 3959 
Subject: Re: RunRadisJM
Date: 06/15/2023 5:40 PM
Post Reply | Report Post | Recommend It!
No. of Recommendations: 3
Radiscreen can RANK.

Oops! The post you link from tetranomad back in 2012 was replying to my post back in 2012 in which
I stated 'I don't know how to approach rank-order of the results of a custom factor made up from an equation of factors.
Keelix did once mention that he had implemented a PERCENTRANK function.'

But as tetranomad noted:
' I think a Radiscript formulation of this screen is perfectly do-able, the questions are whether it would be parsed by RunRadis'
And although the Rank function runs in the Excel macro version which tetranomad pointed out to me more than 10 years ago
(I'm a slow learner or getting a little senile). However it doesn't work in Runradis at least for the Ocaml version.

C:\runradis>pick rankPefDvMo2018 -max 10
Fatal error: exception Lexer.Syntax_error("rankPefDvMo2018.txt:8: 'RANK' not a keyword")

No standard SIP screens currently use the rank function. An typical output from a screening using runradis.
C:\runradis>pick yldearnyear -max 10
** warning: no basic filter **
yldearnyear_sip SMFG ASX FRO GGB CEIX VLO PSX NOG PAA ET

RAM / FLARAM


Print the post


Author: TGMark 🐝  😊 😞
Number: of 3959 
Subject: Re: RunRadisJM
Date: 06/15/2023 11:32 PM
Post Reply | Report Post | Recommend It!
No. of Recommendations: 0
The RunRadis build notes look like they are last updated 2011:
REQUIRED COMPONENTS ON WINDOWS
- caml from http://caml.inria.fr/ (developed on v3.11)
- flexdll from http://alain.frisch.fr/flexdll.html
- bzip2 source code from http://www.bzip.org/ (if building the BinaryDB flavor)
- MS VC++ 2008 from http://www.microsoft.com/express/vc/ (2005 works, too)
- MS Platform SDK 2003 R2 from http://msdn.microsoft.com/en-us/windows/bb980924.a...
(use whatever SDK version is appropriate for your machine)


Is this all still valid for Win10 or Win11 machines? It mentions caml, but looks like that link above takes you to ocaml, with the most recent version being 5.0.0. Is that the one I want to get RunRadis working?
Print the post


Author: RAMc   😊 😞
Number: of 3959 
Subject: Re: RunRadisJM
Date: 06/16/2023 11:29 PM
Post Reply | Report Post | Recommend It!
No. of Recommendations: 1
The RunRadis build notes look like they are last updated 2011:

You somehow got a link to the old ocml version!

In the first post in this thread person who wrote the ocml version of runradis, mi4me announced
that he has created a new Java version of Run Radis.

The link to the source code of the new version released only 4 days ago is:
https://github.com/jrh3/RunRadisJM/tree/1.0

RAMc

Print the post


Author: TGMark 🐝  😊 😞
Number: of 3959 
Subject: Re: RunRadisJM
Date: 06/17/2023 10:47 AM
Post Reply | Report Post | Recommend It!
No. of Recommendations: 0
You somehow got a link to the old ocml version!

Right, but earlier in the thread I think that mi4me indicated that the new Java version was not set up to do screening yet, which is what I wanted to do.
Maybe I misunderstood. I guess he said that picks could be listed by adding some print statements which should be trivial.

Mark
Print the post


Author: mi4me   😊 😞
Number: of 3959 
Subject: Re: RunRadisJM
Date: 06/21/2023 7:22 PM
Post Reply | Report Post | Recommend It!
No. of Recommendations: 1
Any time you "sort", RunRadisJM populates three variables, [Rank], [Tied Rank], and [Tied Rank %]. Thus:

Sort Descending [VR]
Create [r1]:[Rank]

Sort Ascending [SI Price/Sales]
Create [r2]:[Rank]

Create [v]:2*[r1] + [r2]
Sort Ascending [v]

These variables were added to RunRadis, too, but not sure I ever uploaded that version.

-Jim
Print the post


Author: mi4me   😊 😞
Number: of 3959 
Subject: Re: RunRadisJM
Date: 06/21/2023 7:29 PM
Post Reply | Report Post | Recommend It!
No. of Recommendations: 4
I'm thinking about adding a "pick" program to generate the currently passing tickers. However, I'm a little hesitant, because if there are any bugs, then it will generate erroneous output - not something I want to be on the hook for, disclaimers not withstanding.

-Jim
Print the post


Author: TGMark 🐝  😊 😞
Number: of 3959 
Subject: Re: RunRadisJM
Date: 06/21/2023 9:04 PM
Post Reply | Report Post | Recommend It!
No. of Recommendations: 0
I'm thinking about adding a "pick" program to generate the currently passing tickers. However, I'm a little hesitant, because if there are any bugs, then it will generate erroneous output - not something I want to be on the hook for, disclaimers not withstanding.

Couldn't output be compared against RunRadis as a check?
I'm interested in generating picks, one way or the other.

Mark
Print the post


Author: RAMc   😊 😞
Number: of 3959 
Subject: Re: RunRadisJM
Date: 06/22/2023 10:56 AM
Post Reply | Report Post | Recommend It!
No. of Recommendations: 2
Jim:
I'm thinking about adding a "pick" program to generate the currently passing tickers. However, I'm a little hesitant, because if there are any bugs, then it will generate erroneous output

Mark:
Couldn't output be compared against RunRadis as a check?

I think a pick program would be the icing on the cake for your backtester!
And I agree with Mark that users could use the Excel based RunRadis program to compare results. Any differences could be analyzed for errors.
No one could fault your open source program freely given to them with any malice. One of the advantages of open source programs is the users have the ability to analyze and fix errors in the programs.

Thanks for all the effort you have put in to this.
Again wish my hacker backtest code looked as well structured as yours.

RAMc
Print the post


Author: mi4me   😊 😞
Number: of 3959 
Subject: Re: RunRadisJM
Date: 06/22/2023 9:25 PM
Post Reply | Report Post | Recommend It!
No. of Recommendations: 5
Added the "pick" program to RunRadisJM - version 1.1.

-Jim
Print the post


Author: RAMc   😊 😞
Number: of 3959 
Subject: Re: RunRadisJM
Date: 08/23/2023 11:07 AM
Post Reply | Report Post | Recommend It!
No. of Recommendations: 9
RunRadis by mi4me or Jim H already a very capable screener and back tester is being actively improved.
On 6/22 v1.1 added pick to allow screening on the original backtester.
7/17 v1.1.2 added mod() function and instructions for compiling
8/1 v1.1.3 fixed bugs in the DB loader.

I probably spent 100's of hours and 2 years building my own version of a similar Python and MySQL based screener and back tester.
Jim/mi4me has done all the programming plus documenting the results and giving the community an open source free downloadable program.

The current version still has a few limitations like monthly only but is definitely usable.

All you need is a AAII SIP subscription $299/yr plus a days effort to download and install his program and then download and install as many years of historical data from AAII you want to backtest over.

https://github.com/jrh3/RunRadisJM/blob/v1.1.3/REA...

Of course it's a lot more trouble than just subscribing to P123 ultimate currently at $2,400 / year where I get 20 years and a few more capabilities.

Seriously I believe this could be an open source gold mine. If a few more people start using this and sharing results, then start adding a few more features (again it's open source).

RAMc

Print the post


Author: mungofitch 🐝🐝🐝🐝 SILVER
SHREWD
  😊 😞

Number: of 3959 
Subject: Re: RunRadisJM
Date: 08/23/2023 5:30 PM
Post Reply | Report Post | Recommend It!
No. of Recommendations: 3
Forgive my ignorance, but can someone summarize which tools are able to produce [current] picks from [current] tab delimited data files?
(i.e., able to use any data source including VL, SIP, etc)

The original Radiscreen spreadsheet can, but I haven't ever used any other tool.

Jim
Print the post


Author: RAMc   😊 😞
Number: of 3959 
Subject: Re: RunRadisJM
Date: 08/23/2023 6:24 PM
Post Reply | Report Post | Recommend It!
No. of Recommendations: 2
Jim:
can someone summarize which tools are able to produce [current] picks from [current] tab delimited data files?
(i.e., able to use any data source including VL, SIP, etc)


You are correct, only the original Radiscreen spreadsheet can.

But the Radiscreen spreadsheet requires running either your VL or SIP program and then exporting the all the specific fields your screens use to Excel. When I was using Radiscreen my export field list included almost 200 fields.
RunRadis is a different program that runs only SIP screens written in Radiscript. It runs directly from the SIP programs downloaded and installed db files. You can use any of the ~2000 factors from SIP.
The additional advantage is that it also does backtesting if you install the older historical monthly or weekly SIP programs from the AAII website and load them into a db. Easier and better than Radiscreen for SIP users, no help for a VL user.

RAMc


Print the post


Author: mungofitch 🐝🐝🐝🐝 SILVER
SHREWD
  😊 😞

Number: of 3959 
Subject: Re: RunRadisJM
Date: 08/23/2023 9:20 PM
Post Reply | Report Post | Recommend It!
No. of Recommendations: 0
You are correct, only the original Radiscreen spreadsheet can.

Thanks for the summary.
I have programs and data feeds that calculate some custom fields, (and I use VL for some things), so that's why I was asking.

Jim
Print the post


Author: agehlot   😊 😞
Number: of 3959 
Subject: Re: RunRadisJM
Date: 08/24/2023 10:05 AM
Post Reply | Report Post | Recommend It!
No. of Recommendations: 0
Pardon my ignorance but from where I can download original Radiscreen spreadsheet?
Thanks in advance,
agehlot
Print the post


Author: lizgdal   😊 😞
Number: of 3959 
Subject: Re: RunRadisJM
Date: 08/25/2023 12:11 PM
Post Reply | Report Post | Recommend It!
No. of Recommendations: 0
The original Radiscreen download site is no longer hosted.

A 2019 post: "Radiscreen files. I would suggest downloading all as Phil's site will not be hosted forever."
http://www.datahelper.com/mi/search.phtml?nofool=y...

In 2023, Google Sites says:
"Site not found
We're sorry, we were unable to locate the site /TheGreatRadish/."
http://www.datahelper.com/mi/search.phtml?nofool=y...
Print the post


Post New
Unthreaded | Threaded | Whole Thread (24) |


Announcements
Mechanical Investing FAQ
Contact Shrewd'm
Contact the developer of these message boards.

Best Of MI | Best Of | Favourites & Replies | All Boards | Followed Shrewds