Scan omission - heads up

lmkwin

I already sent this to support but just wanted to give a heads up.

I have a list that has 205 symbols. When I run a scan that simply ranks them by % from 52 week high, it excluded 2 of the symbols in the results. So a 205 list returned 203 symbols. The symbols omitted were BERK and PSCF. BERK appears to have a problem with its chart so perhaps the history is messed up. PSCF appears to be OK from what I can tell from a data perspective. PSCI is also on that list but it came through the scan.

I often will run a scan on a list to just sort the list and then overwrite or replace the existing list with the results. This is the 1st time that I've noticed a difference in the results. Not saying that it hasn't happened before. It's just the 1st time that I've noticed it.

markd

Have your tried re-running the scan several times?

It's possible it could be a transmission error either at your end or the SC end or somewhere in between, rather than a database error.

Also, you could try putting just those symbols in a separate list, run the scan against it and see if they get picked up.

Also, they seem to be very low volume symbols. I don't know why that would affect the results, but it seems possible - e.g., if they haven't traded since the open when you run your scan, there may not be any data? Just guessing. You could try backdating the scan for dates when you know there is data.

lmkwin

Yes, tried it several times. When I run it today, I get 204 of the 205. PSCF is in, but BERK is not. I believe that has to be a data issue. That chart is really messed up. I received no response from support but that's what I'm guessing.

[favorites list is 1] and [symbol starts with 'BER']

returns 0 results now.

Just giving a heads up. It's a little concerning though when you think about scanning lists.

markd

I don't think the chart is messed up. I think it is thinly traded - some days no trades at all, and when there are trades, the bid and ask can be very far apart because there is no liquidity.

lmkwin

I'd say that the chart has all the prerequisites for

rank by [[[Close - Max(253,close)]/close]*100]

But when I run it for a day(s) that BERK actually traded, it doesn't appear in the results either. 205 in the list. 204 in the results.

markd

what happens if you use PctDiff?

rank by PctDiff(close,max(253,close))

not sure if it needs brackets. rank by seems to be kind of finicky with complex conditions. Also, maybe reverse the order inside the parens, see what happens.

lmkwin

something really weird happens when I use that rank by statement. I'll investigate further before sharing. I've plugged that rank by into other scans and it's doing the same really weird thing for me. Closed and reopened my browswer. I'll try the log out log on thing when i get a chance.

lmkwin

Logging out and logging in resulted in the same weird thing.

When i put that Rank by in any scan, it opens a new tab on my browser (just like normal) and instead of results it shows the Technical Scan that is the same as the Technical Scan I'm trying to run. In other words, it opens a duplicate tab. If I put other rank by statements in the scan it runs fine. Opens a new tab and displays the results.

On a different note, I found this in the documentation

Warning: Scanning a ChartList will only return symbols on the ChartList that are actually in the scan engine. Some symbols are not available in the scan engine, so they will not be returned as scan results even if they are in the ChartList being scanned.

https://stockcharts.com/docs/doku.php?id=scans:advanced_scan_syntax:chartlist_scans

So I'm guessing that BERK may not be in the scan engine. Doesn't explain PSCF appearing the next day but, I'll accept that on BERK.

lmkwin

I did receive a response from support so they are aware of the situation

Response:
Thank you for contacting us. Currently, we are experiencing a technical issue with the site. We are aware of the issue and currently are working to resolve the problem as quickly as possible.

We appreciate your patience and understanding at this time.

My inquiry:
Usually when I run a scan, the results pop up in a new tab with the results. When I run a scan now, the new tab don't show the results. It opens a new tab with a duplicate of the Technical Scan workbench I'm running.

Also, it's 10:20am now and the Movers and SCTR panes are not showing any updated information. All of them are showing zeros for changes.

lmkwin

When I re-run the List Scan using the

rank by PctDiff(close,max(253,close))

It now runs OK. Still excludes BERK but at least it runs like "normal". As we get closer to the ACP rollout, I'd keep an eye out for anomalies. ACP is quite an interesting chart program. Best wishes to the developers.

markd

It seems the site has been having issues for the past couple of days. Hopefully that's behind us.

lmkwin

Got a response from Support regarding scan omissions

"The scan engine doesn't track inactive stocks, so if you have a stock in your list that didn't trade on the day of the scan, like BERK, you won't get that stock in your scan results."

I still think that it should logically perform a Rank By statement but, I did get an answer.

markd

Well, I'm not sure how the scan engine works in detail, but if there is no data for a stock today, it probably doesn't go looking for the last day there was data. That could vary widely. How far back should it go? As you can imagine, searches would add a lot of time to the query, and would be difficult to test.

lmkwin

A list of symbols that the system does recognize should be able to be ranked by Name. It shouldn't matter if it traded TODAY or yesterday or not. Like I mentioned before, I don't know if this has been happening all along and I never noticed (seems unlikely) or is a change in the system/ database programming. As they are continually enhancing the site, I'm leaning toward the latter. I used to write database reports for a couple decades. It's just a heads up if you are running scans against your lists, it may be excluding things from your list and you will never know it.

markd

Well, I would guess there are competing priorities - e.g. completeness versus response time. Also, a stock that doesn't trade very often, or trades on very light volume, is probably not a good candidate for technical analysis to begin with, so it's a reasonable business decision to omit them from searches on a site that is specifically oriented toward TA, vs. say, portfolio management, to improve performance.

lmkwin

Just something to be aware of. BERK got into that list via a scan so at some point it was worthy of TA.