Surprised there wasn't a ticket for this so I created one. If it's already in here, sorry in advance.
RCalhoun@paragon-adv.com
I submitted a request to correct the data issues surrounding the equity screener a few months back. Is there anyway of following up on that? It was originally submitted to investigate any ratio related to companies with no debt. For example, when ranking stocks with the lowest debt-to-capital ratio, a company with 1% debt will be awarded high points due to low debt levels, a company with 40% debt will be awarded average points, and a company with no debt is awarded no points (the worst possible score). Thus, I cannot use any metric within a ranking system that uses debt since those companies with the least of it (zero) are penalized the most.
In June alone I spent over a week analyzing 100+ equity metrics to determine which ones were viable within the scoring model, I ended up only being able to use about 15; all the others we littered with "bad data" rendering them useless (from a scoring perspective). Regarding the other metrics, I have TONS of examples where the numbers are just flat out wrong. I understand it would be impossible to catch these manually, but is there any systems in place to check for logical values. For example, a normal ROIC should range from 0 to 20%, possibly up to 60% for the very most profitable companies. Please see below for "bad data". AZPN is flat out wrong when you cross-check it against Morningstar's data.
I like this idea
Good idea. LOTS of work....just need to know if worth it.
Good idea. LOTS of work....just need to know if worth it.
Hi Han,
I know we talked about this when I was there, but I'm trying to visualize this problem. I created a saved screen with only long term debt, lower is better and got this:
https://ycharts.com/screener/company/#/?metricFilters=&page=1&securitylistFilters=method:add,name:all_equities,group:company,,method:intersect,name:major_us_exchanges,group:company,,method:intersect,name:index_sandp_500,group:company&ratings=1-attractive,2-neutral,3-avoid,4-unrated&screen=&sort=648686&sortDirection=asc&onlyPrimaryShares=false&orderedColumns=name:currency_code,type:infoField,,name:total_long_term_debt,type:metric,,name:648686,type:score
Putting aside the unrelated UAL issue.... this looks fine to me.... Can you give me a link that illustrates what he's talking about?
Hi Han,
I know we talked about this when I was there, but I'm trying to visualize this problem. I created a saved screen with only long term debt, lower is better and got this:
https://ycharts.com/screener/company/#/?metricFilters=&page=1&securitylistFilters=method:add,name:all_equities,group:company,,method:intersect,name:major_us_exchanges,group:company,,method:intersect,name:index_sandp_500,group:company&ratings=1-attractive,2-neutral,3-avoid,4-unrated&screen=&sort=648686&sortDirection=asc&onlyPrimaryShares=false&orderedColumns=name:currency_code,type:infoField,,name:total_long_term_debt,type:metric,,name:648686,type:score
Putting aside the unrelated UAL issue.... this looks fine to me.... Can you give me a link that illustrates what he's talking about?
Requested by Mike Chen - Byrne Asset Mgmt $7,500
"Hi kyle,
In the ycharts website’s equity screener tool that has customizable “scoring models”. It seems like for the data points that do not exist, the default value is -1. When we use this for our scoring models, these data points really skew the final rankings. Is there a way to set these blanks to 0 instead?
It is not so bad by itself. But if we have a group of 8-10 data points in a scoring model, and a few of them are missing, it really screws up the score.
Thanks,
Mike"
Requested by Mike Chen - Byrne Asset Mgmt $7,500
"Hi kyle,
In the ycharts website’s equity screener tool that has customizable “scoring models”. It seems like for the data points that do not exist, the default value is -1. When we use this for our scoring models, these data points really skew the final rankings. Is there a way to set these blanks to 0 instead?
It is not so bad by itself. But if we have a group of 8-10 data points in a scoring model, and a few of them are missing, it really screws up the score.
Thanks,
Mike"
We are planning on finishing this by end of year, it will only be available in the redesign, not the old site.
We are planning on finishing this by end of year, it will only be available in the redesign, not the old site.
Replies have been locked on this page!