User talk:ClueBot Commons/Archives/2011/September

From Wikipedia, the free encyclopedia
Jump to: navigation, search

A barnstar for you!

Brilliant Idea Barnstar Hires.png The Brilliant Idea Barnstar
let me join Kairan tumaquin (talk) 13:05, 1 September 2011 (UTC)
Join what? --Σ talkcontribs 23:20, 1 September 2011 (UTC)

The ClueBots' messages'

(When someone logs into the ClueBot Commons account, they are gonna have a CRAPLOAD of messages to read :P )

After this message was added here (which was added twice, but the second edit with that was not tagged with "repeating characters") I'm thinking that someone should edit every one of the ClueBots' code to also put a giant warning that is clearly visible saying "THIS IS A MESSAGE FROM A AUTOMATED BOT. THIS BOT IS NOT A HUMAN." with much, much, MUCH emphasis on the word "NOT".

(P.s. As if this isn't very obvious, I am really just posting this here as a joke. Though I would kind of be surprised if it actually was done) LikeLakers2 (talk) 21:30, 2 September 2011 (UTC)

The ClueBot Commons talk page is monitored by myself and a ton of other people. I technically "own" the account, but just to use as a userspace to use for the various ClueBots. -- Cobi(t|c|b) 21:33, 2 September 2011 (UTC)
Yeah. *insert me abusing my keyboard to screw up this page as much as possible here* :P LikeLakers2 (talk) 21:36, 2 September 2011 (UTC)

Ur no fun

Let me edit man--Furymaster (talk) 04:00, 3 September 2011 (UTC)

Taking a guess here: you must of been vandalising, as the bot will revert any possible vandalism. Nothing else. LikeLakers2 (talk) 04:04, 3 September 2011 (UTC)
Actually, looking at the edit, it was inserting commentary into a article, which can count as vandalism. So really, it will let you edit if you don't intentionally vandalise. LikeLakers2 (talk) 04:07, 3 September 2011 (UTC)
I've also looked at the edit and ClueBot was quite right to make it. Firstly (and most importantly) the Bot was correct because you have made allegations about this person that are unreferenced. Secondly, the way you made the edit it looks like commentary in an article.--5 albert square (talk) 22:17, 4 September 2011 (UTC)

Request to add to bot

Could you add to the bot the pattern vandalism that has occurred on Peter Cetera, Lillian Too, Jack Colvin, Philip Oakey, and Tharman Shanmugaratnam, all from IPs begining with 60 or 175? I have some filters in place, but they don't seem to be working consistently. Thanks, NawlinWiki (talk) 13:16, 3 September 2011 (UTC)

ClueBot NG uses machine learning. There is no list of patterns that it prevents against. It takes in a few hundred statistics into a black box and out comes an answer. The old ClueBot was heuristic based, but much of the advantage of that was replaced with the Edit Filter. There were a few things that the old ClueBot could do that the Edit Filter couldn't, but all of those have been eclipsed by ClueBot NG. I would recommend that you add that to the Edit Filter. It seems simple enough to catch with the Edit Filter. -- Cobi(t|c|b) 21:17, 3 September 2011 (UTC)
I have an idea to improve the AngryOptin feature that would help.
The idea is to add the option of a lower vandalism score threshold for specific ip ranges from which substantial abuse is received. The new, Super AngryOptin page would contain a column for IP ranges, a column for corresponding articles, and a column to specify whether or not to respect the one revert rule.
Both, the ranges column and the articles column could be left blank for each row, but not both. If the ranges column was left blank, the low threshold would affect any range from which an ip is editing the specified article. And if the articles column was left blank, the low threshold would affect any article in mainspace edited by an ip from the specified range.
It will be something like:
IP range Article Ignore 3RR George W Bush yes no
Barack Obama no
Keep in mind, this feature will only lower the threshold, it will not affect the way the bot currently works, like not reverting non-mainspace edits or not reverting users who contributed x number of edits. Of course, this new page will be edited only by admins, so it should be an addition to , not a replacement of the old feature. Sole Soul (talk) 16:38, 4 September 2011 (UTC)
There is currently no consensus for disrespect of the 1RR. --Σ talkcontribs 18:14, 4 September 2011 (UTC)
Regarding 1RR, nothing in my suggestion is different from what the bot currently does. The bot already ignores 1RR in all pages listed here. Sole Soul (talk) 18:22, 4 September 2011 (UTC)
In any case, half of the pages listed there are userpages. In 1RR, that is not very different from the optin pages, which I have no strong feeling toward, but what you're asking for is the assumption of bad faith to an IP range, which as usual, may be used by more than one person. --Σ talkcontribs 19:14, 4 September 2011 (UTC)
The 3RR is not the essential part of this suggestion. This suggestion could lead to the unblocking IP ranges. You could interpret the current optin as assumption of bad faith to any IP editing a specific page, which as usual, may be edited by more than one person. Sole Soul (talk) 21:07, 4 September 2011 (UTC)
Doesn't semiprotecting the page accomplish the same thing? --Σ talkcontribs 21:25, 4 September 2011 (UTC)
No. I will give an example to clarify. Let assume we have a "bad" range. And let assume that the general threshold is 0.9 and the lower threshold is 0.8. Now:
* A vandal from the bad range makes a bad edit, the bot give it 0.85 as a score, it will be reverted.
* A vandal from a range not included in the optin page makes a similar edit, the bot give it 0.85 as a score, it will not be reverted
* A good user from the bad range makes an edit, the bot give it 0.6 as a score , it will not be reverted.
* A good user from the bad range makes an edit, the bot give it 0.85 as a score , it will be reverted as a false positive. Sole Soul (talk) 21:49, 4 September 2011 (UTC)


False Positive: 569432

(And your report page sucks.) — Preceding unsigned comment added by (talk) 23:46, 5 September 2011 (UTC)

If you tried putting that number into one text box and clicking one button (as the instructions state), maybe the false positive would be processed faster. --Σ talkcontribs 23:57, 5 September 2011 (UTC)
What makes you think I haven't done that? -- (talk) 00:03, 6 September 2011 (UTC)
Why tell us then? In other news, I've checked your report. --Σ talkcontribs 00:15, 6 September 2011 (UTC)
Have you seen the report page? Putting a number into a text box and clicking a button is not enough. -- (talk) 00:26, 6 September 2011 (UTC)
How so? If you can't fill in one box and client a button then possibly the issue is not with the usability. Once in the report interface it can be reviewed, it cannot if it never gets in there. - DamianZaremba (talkcontribs) 00:28, 6 September 2011 (UTC)
Have you seen the report page? Filling in one box and clicking a button is not enough. -- (talk) 00:44, 6 September 2011 (UTC)
Seeing that you already found the revert number, I don't see how hard it is for you to go here, scroll down, and put the number in the box and click a button. --Σ talkcontribs 01:02, 6 September 2011 (UTC)
Wow, you are really not listening, aren't you? -- (talk) 07:14, 6 September 2011 (UTC)
I was listening up until the point where you ignored my question and repeated your non-sensical statement. - DamianZaremba (talkcontribs) 07:17, 6 September 2011 (UTC)
Actually, nobody answered my question yet. Let me repeat it: Have you seen the report page? Have you logged out first? -- (talk) 08:34, 6 September 2011 (UTC)
Yes, it looks exactly like it should - if it didn't then I would have commented on that rather than asking you what you are referring to. As pointed out numerous times previously there are VERY clear instructions on the report interface's homepage with a BOX to put your ID in and a BUTTON to click. If we hadn't of looked at the report page we wouldn't know these things so in fact your question is redundant. - DamianZaremba (talkcontribs) 08:37, 6 September 2011 (UTC)
  • The instructions page should be updated. In the "from your talk page" way, you don't need to click the edit button, and you don't need to find the number. The link in the warning message which leads to the reporting page, already contains the revert number in it, so all that you need to do is click "report it here" in the message. Sole Soul (talk) 10:00, 6 September 2011 (UTC)
IIRC User:ClueBot NG only links to the report page on the first warning. After that it is just the usual template with the ID as a comment hence the instructions on the page. If you spend 2seconds reading the page you are probably never going to see the instructions as when you click on the link all you have to do is press one button and optionally enter a comment - even simpler! I still don't see the point User: us trying to make though. - DamianZaremba (talkcontribs) 11:02, 6 September 2011 (UTC)

Warning reverted

I'm curious - why did ClueBot revert my warning at IP194.80.20.151 (talk) at 10:34 today? Denisarona (talk) 12:52, 5 September 2011 (UTC)

ClueBot didn't revert you, the diff shows that you and the bot edited at the same time. I'm willing to bet that's what caused your warning to disappear - had that happen myself!--5 albert square (talk) 22:36, 7 September 2011 (UTC)

Report page down?

The report page gives me a 403 error, so I'll just dump a report ID here: 578293 ∫eb²+1(talk) 09:07, 10 September 2011 (UTC)

Same problem here: 577248. – hysteria18 (talk) 14:03, 10 September 2011 (UTC)
Same problem here: 579267. Comment: This is an extremely frequent problem when editing dab pages. (talk) 20:03, 10 September 2011 (UTC)
I've fixed the problem with it and added these reports in. - Damian Zaremba (talkcontribs) 21:40, 11 September 2011 (UTC)

False Positive

On this page ( today the bot made a false positive. What happened is I had clicked the Edit button for one section of the article, but the html that was included on the edit page also included the code for the section which followed that one. It looked odd, and when I saved my edits the result was that that second section appeared twice in the article, in succession. I then did a second edit by removing the duplicate (the one that had appeared on my edit page). That is what triggered the bot into thinking there had been vandalism. There was none. Coastwise (talk) 17:33, 10 September 2011 (UTC)

Er, from what I can see the Bot didn't revert your edit. This is the edit the bot reverted, the edit summary by the bot says that it's reverting the revision by the IP and going back to the revision by you, it does not say that it's reverting your revision. If it had reverted your revision you would know about it because the bot would leave a message on your talk page. Whilst the reversion by the bot is not obvious vandalism, I would not look to overturn it because it's original research --5 albert square (talk) 16:18, 11 September 2011 (UTC)
Thanks, 5 Albert Square. Also, I tried to report the false positive by the link that was provided, but got an error message. — Preceding unsigned comment added by Coastwise (talkcontribs) 16:38, 11 September 2011 (UTC)
That's ok, if you look at the section above I think a few people have had this :)--5 albert square (talk) 20:30, 11 September 2011 (UTC)

References removed

I gotta compliment your Bots work, I have yet to see it cause a problem. I want to suggest an additional, more than occasional clue. In order to mask their dirty work, vandals frequently place an edit note. "References removed" is the favorite one. Maybe it is not an automatic clue, but the vast majority of edits that claim to or actually remove references, in wikispeak, sources, have no business being removed and the remains of the article after such vandalism will be a poorly or improperly formatted article. Add those two together, (before, properly formatted article; after, formatting errors) and the edit note "references removed" and you've probably got a vandal. Just a thought. Trackinfo (talk) 17:44, 11 September 2011 (UTC)

As the bot uses an ANN it will pick up patterns like these but on a munch larger scale - one that is difficult for humans to see. The old ClueBot use to work off patterns like this which where manually maintained but they are very difficulty to ensure accuracy with. If there are certain types of entries/vandalism that the original ClueBot works better at we could look at letting it run just for these but the ANN results will improve as the dataset does. Thanks for your ideas though! - Damian Zaremba (talkcontribs) 21:43, 11 September 2011 (UTC)
I think the old ClueBot should be enabled in the portal namespace. --Σ talkcontribs 21:56, 11 September 2011 (UTC)


I was wondering if someone can please answer a question I have about ClueBot in regards to it making a mistake. Yesterday while I was on Wikipedia looking at the page for the upcoming Superman Movie The Man of Steel Directed by Zach Snyder I all of sudden get a message saying I visited and edited a page called Videos and audio recordings of Osama bin Laden and I was then sent a warning message by ClueBot not to do that type of editing again. But here is the thing Yesterday and any time before that I have never done any editing type work on Wikipedia from my personal computer I own I just visit the site thats all. But all of sudden this ClueBot says I visted this certain page and did editing work on and even my history page says I visted the page. But the actual real this that happened is I never visted that page noir did I edit the page so can someone please explain to me why ClueBot says I did something that I actually never did and never visted on Wikipedia. An yesterday just to make sure I didn't read it wrong I did one small editing work on the TV Show Supernatutal Season 7 page to make sure I read the right User Name that was given for my computer. Now I followed the process of how to report a problem if ClueBot made a mistake and I know the editers who watch this page perfer no one to leave messages here if there about ClueBot making a mistake and causing a problem. But the User Name I was given was User: because I'm not a registored user and it only has two contributions list the page title Videos and audio recordings of Osama bin Laden which says I visted and edited which is absoultly not true at all because I've never visted that page ever and the page for Season 7 of The TV Show Supernatural. An my talk page has similar type information. If someone could please answer me back or at least fix this problem I would be very thankful and if it means getting getting rit of all of my information contributions, talk page and even blocking me please do that if means it will fix the problem. The reason I'm asking this question and leaving this message is I don't like being warning and accused for editing a page I didn't even vist or edit by ClueBot. I understand why ClueBot exist but could someone please answer my question and fix this problem.

From User: —Preceding undated comment added 22:55, 12 September 2011 (UTC).

You have a dynamic or shared IP address. --Σ talkcontribs 02:02, 13 September 2011 (UTC)

instead of removing my post, or atleast just correcting it because i was drunk, why dont you watch the movie and post it yourself that barton aka hawkeye makes a cameo. sheeeit. not looking for fame, its just wiki, christ. could atleast update that article to include his appearance in thor. — Preceding unsigned comment added by (talk) 19:25, 13 September 2011 (UTC)

Not archiving

Cluebot hasn't archived User_talk:Citation_bot for some months. The automatic archiving helps me to keep on top of bugs with my bot — I miss it! Any update on when the bot might be working again? Can I run the script myself from my toolserver account? Thanks, Martin (Smith609 – Talk) 20:00, 17 September 2011 (UTC)

Seems the bot wasn't running properly - it is catching up on the backlog now so your page should get archived pretty soon. - Damian Zaremba (talkcontribs) 20:11, 17 September 2011 (UTC)

What happened here?

Not the end of the world, but I'm just wondering why CBNG didn't recognise this warning, and issued a (second) l2 warning. HJ Mitchell | Penny for your thoughts? 17:13, 22 September 2011 (UTC)

It might not have recognized the template, I'll dig though the logs later and try to see what happened. - Damian Zaremba (talkcontribs) 17:22, 22 September 2011 (UTC)
{{subst:test2}} ≠ {{subst:uw-test2}} Dynamic|cimanyD contact me ⁞ my edits 00:06, 24 September 2011 (UTC)

no comprendo

I am not stupid. And I want to report a false positive. But I cannot get the ID right (now what would an unexperienced editor meet).

All I can do is a diff: [1] (and do not start telling me how to do it next time. Just create a good backoffice , I'd say). -DePiep (talk) 21:54, 23 September 2011 (UTC)

Might as well not have users. :O --Spidey665 00:38, 24 September 2011 (UTC)

Still not archiving

Further to User_talk:ClueBot_Commons/Archives/2011/September#Not_archiving, the bot still doesn't seem to have visited User talk:Citation bot. Martin (Smith609 – Talk) 02:30, 22 September 2011 (UTC)

Hmm the bot is running, I'll have a look though the logs and see why it isn't touching the page - might have to get User:Cobi to look at it. - Damian Zaremba (talkcontribs) 17:23, 22 September 2011 (UTC)
Thanks – I appreciate you looking into this. Martin (Smith609 – Talk) 03:52, 25 September 2011 (UTC)

Protection of User:ClueBot NG/Run

At the moment, this page is semiprotected, but there's a request for full protection at WP:RFPP, and the requesting editor has noted that the "Documentation" section of the userpage refers to those who edit the page to stop the bot exclusively as admins, rather than as editors. Against this I would support the current format: the page's history shows that it's never gotten significant vandalism (except for one editor that was warring over it), and we've only had two incidents of vandalism all year. To me (I'm saying this as an admin), it seems reasonable that average registered users should be able to stop the bot temporarily. Moreover, I don't see the point of an admin-only page to stop the bot: we admins can already stop the bot by blocking it, so the only real use of a page like this is if it can be edited by non-admins. Nyttend (talk) 04:07, 23 September 2011 (UTC)

Personally I see it like this - admins are invited to stop the bot by blocking it if it is doing bizarre things (as they should and would), having a run page that the bot 'soft checks' (by soft I mean if it is functioning correctly then it will follow) is useful for many purposes. Due to the lack of vandalism I don't personally support it being protected - I've used it a few times and find it useful when debugging/tweaking the bot. If the page was being vandalised often then I would support the check being removed over the page being protected - there is no point having 2 places to do one thing! - Damian Zaremba (talkcontribs) 04:14, 23 September 2011 (UTC)
I've never considered that the page was meant as a method for non-admins to shut down the bot, but that sounds reasonable. As Nyttend says, the page isn't actively vandalized, and even if it is, nothing drastic results from it (e.g., the bot will not malfunction). Someone will just spot it and change it back to "true". So I withdraw my request for full protection of the page. Goodvac (talk) 15:16, 23 September 2011 (UTC)
I actually raised the concern about this on ANI. I do admit it isn't that vandalized. So instead, we could create an edit filter that automatically prevents any vandals that changes the page to anything other than "True" and "False".OpenInfoForAll (talk) 22:14, 23 September 2011 (UTC)
I don't see the point in the page being fully protected. I've looked at the page tonight and over the course of the last few weeks I think the page has only been vandalised once. If it was being constantly vandalised by a number of editors that would be a different story. At the minute though any request for full protection would fail though simply because of not enough vandalism on the page. --5 albert square (talk) 22:18, 25 September 2011 (UTC)

Teach ClueBot NG to recognize something new?

The article Overblood has been getting persistent vandalism since last January — almost always including the same bit of garbage (albeit with a few variant spellings). You can check the article's history to see what I'm talking about. Is there any way to get ClueBot NG's cluefulness level raised so it will detect this? Richwales (talk) 02:55, 26 September 2011 (UTC)

User:ClueBot NG/AngryOptin. →Στc. 05:10, 26 September 2011 (UTC)


Is the point of this page just to complain about the bot? — Preceding unsigned comment added by JeffClarkis (talkcontribs) 02:43, 28 September 2011 (UTC)

As stated at the top of the page "This page is for comments on or questions about the bot." - Damian Zaremba (talkcontribs) 02:45, 28 September 2011 (UTC)

"Recent history" cutoff for warning escalation

I've seen a few times lately where an IP vandalizes a few times and gets one or more cluebot reverts with talk-page warnings. Then a few days later, it happens again. Lather, rinse, repeat, with cluebot restarting at level-1 each time (example: [2]) I know we need a cutoff because it's likely that an IP would be reassigned after a while or be a public machine. But seems like the current cutoff is too conservative. DMacks (talk) 17:07, 28 September 2011 (UTC)

Yes, it's very annoying to see walls of long ClueBot warnings under multiple identical headings. Maybe the cutoff should be between half a month and a month. Dynamic|cimanyD contact me ⁞ my edits 17:16, 28 September 2011 (UTC)

Litnarovich Ruslan Nicolaevich

This text does not carry any meaning, perhaps just an inexperienced user. Sorry for bad English Blind GM (talk) 21:54, 28 September 2011 (UTC)

Strange ClueBot edit

In this edit, ClueBot added a warning, but also removed all the other warnings that were on the page. Is this a known issue? I've never seen ClueBot do that before. Thanks, BMRR (talk) 04:38, 30 September 2011 (UTC)

Interesting... not sure how that one happened. I'll go ahead and assume the variable that holds the talk page content which the warning gets appended too didn't get set properly for some reason. I'll keep an eye out and see if something needs tweaking (: - Damian Zaremba (talkcontribs) 13:56, 30 September 2011 (UTC)

ClueBot NG malfunctions

My experience is all the ClueBot NG's. reverts are false positive.-- (talk) 13:17, 30 September 2011 (UTC)

That is not the case, feel free to go manually review a few thousand edits here. - Damian Zaremba (talkcontribs) 13:54, 30 September 2011 (UTC)