User talk:Born2cycle

Coherent reply policy If I put a message on your talk page, I will be watching that page for a reply. If you leave a message here, I will reply here, unless you request otherwise.

Observation about Page view statistics

B2C, in a recent discussion you stated Page view counts are useful evidence there. when trying to measure what readers are searching for. I don’t think that’s really true, and I’ll address that in a minute. The reason I am a bit hard over on naturalness and recognizability as titling criteria is that it puts an unnecessary and impossible burden on editors to defend a title based on what readers are thinking or reacting to. When there is no empirical method (even a grossly inaccurate one) to validate what readers are thinking or reacting to, any answer is the right answer and thus serves no useful purpose in adjudicating a title decision. When opposing editors believe they have the right answer but neither side can empirically defend it, we find ourselves in contentiousness that in the end never serves us well. Unlike Commonname, Precision, Conciseness, and Consistency which put the burden on actual title and can reasonable be adjudicated with empirical evidence from sources and such, naturalness and recognizability cannot.

Now my observation about Page Views. From my reading of Wikipedia:About page view statistics, the statistic for any given page is incremented when a reader navigates to the page. That navigation can occur in a number of ways:

• A reader searches Google or such for something and a link to a WP article shows up in the search results
• A reader uses the WP search function, gets the article title perfectly and navigates directly to the article
• A reader uses the WP search function, gets the article title imperfectly or clicks search instead of Go and gets a results list. A click on one of the results will take the reader to the article
• A reader is not searching for anything, but instead is reading an article and clicks on a wikilink in an article and is taken to another article.
• An editor knows the name of an article and navigates to it directly to do some editing.

If we believe paragraph two of information about Page Views: There is no way to know why people have seen the article. Perhaps they wanted to read about it, but there are other options. Perhaps they were seeking something else with a similar name, perhaps the name was interesting or similar to something else and called their attention, perhaps they were browsing articles seeking some specific information, perhaps they were doing maintenance... then the only thing that can be deduced from Page Views is that some unknown % of readers were searching for the page in either Google or WP, but we cannot deduce exactly what there were searching for. Also, because readers navigate to pages via wikilinks embedded in articles, I find the statements such as readers will be confused if the article title is not exactly what they are searching for or something to that effect, in complete contradiction to actual practice. One, I know of no empirical way to determine the level of confusion a reader might be experiencing, but more importantly we don’t make that same argument for wikilinks embedded in the prose of an article. Statistics about page views are useful, but they don’t give any information about what readers are searching for or how they react when they navigate to an article. BTW, were you confused when you clicked on the second and third link to: Wikipedia:About page view statistics in this post? I believe the intent of Naturalness and Recognizability are valid, it’s just that as titling criteria they do far more harm than good and their intent can be dealt with through other criteria. --Mike Cline (talk) 17:18, 31 January 2012 (UTC)

Sorry I missed this earlier. I won't get into a semantic argument defending the readers will be confused point. I don't believe I've ever said that. But I do think you're missing something. While it's true that page view stats tell us nothing about why any one person made it to a given page, it certainly tells us quite a bit about trends. For example, we know Google orders results by popularity. So if someone Google searches for a given term, looks through the results (which typically include enough content from the page to indicate what it is), and clicks on the 12th item in the results, that slightly increases the rating of that page. If enough people do that, it starts moving up in the results. So if someone arrives at a page after clicking on it in Google results that's likely to be what they are looking for. Further, if they're not looking for that, then they're likely to either click back to the Google results and click on another result until they do end up at the page they are looking for, or getting to the dab page via a hatnote link and then getting to their desired page, thus eventually bumping up its page view hit count one way or another. Yes, that one user bumps up the page view counts of the pages he hit inadvertently just as much as he bumped up the page view count of his desired page, but with multiple users the more popular page will over time get a higher count.

Consider five uses of a given term A, B, C, D and E. Say C is what most (say 80%) people are looking for, and the other 20% are divided up evenly among the other pages, but the first page they go to is via a method that turns out to be totally random. So of the first 100 people, 20% end up initially at A, 20% at B, 20% at C, 20% at D and 20% at E. But each of those people eventually finds their way to their desired page. So in the end C gets 20 (initial - of which 80% or 16 are at their desired destination and 4 are looking for something else) + 80% of the others (16 of each of the 20) for a total of 20 + 16 *4 = 84 page views, while A, B, D and E each get 20 (initial - of which 5% or 1 is at the desired page, 80% or 16 are looking for C, and the remaining 3 are looking for one of the other non-C pages) + 5% of the others (1 of each of the 20) for a total of 20 + 1 * 4 = 24.

That is, even though we can't account for the reasons that any one page view occurred, the total page view count over time, relative to the page view counts of other articles, is a very strong indicator of how likely people are to be seeking that article relative to other articles. --Born2cycle (talk) 16:58, 7 February 2012 (UTC)

Don't disagree with the logic until you get to the last bit. Only some % of page views result from readers searching for a topic via Google, some other % are direct navigation because someone actually knows the name in WP, and some % are from wikilinks in other articles. In a recent RM, an editor actually complained that the pageview statistic was skued toward an alternative title because another editor had used the alternative title extensively in Wikilinks in a lot of other articles. All I am contending is that collective page view statistics are an unreliable indication of what readers are searching for because they accumulate from several different types of reader (an editor) activity unrelated to searching for a topic. If we had reliable, empirical information that on average X% of page views result from searching, that % was consistent across a wide range of articles, it might be a useful determinant. But absent that, it is much too unreliable because of the multiple ways page views are generated. --Mike Cline (talk) 22:17, 7 February 2012 (UTC)

But obscure cases aside, I suggest it's generally true that an article A which is significantly more likely to be sought via search than article B is also more likely to be visited via links than is B. For example, London is both more likely to be sought than London, California, and more likely to have visits from links than is London, California. So even if a significant percentage of article page views come from links it's still a valuable measure of relative likelihood to be sought.

The only time this wouldn't be true is in a case of two articles A and B in which even though A has significantly more page view counts than B, it has a disproportionately high percentage of page views due to visits from links, such that the number of page view counts B has as a result of searches is higher than the number of page view counts A has as a result of searches. I suggest it would be very difficult if not impossible to find a case like that. That's why I think it's reasonable to ignore that factor, and just assume that the percentage of page view counts that result from searches is close enough to being the same to presume that are the same.

But, just in case, we do also look at link counts. In a case where A has significantly more page views than B, but B has significantly more link counts than A, it's probably a no primary topic situation. --Born2cycle (talk) 23:38, 7 February 2012 (UTC)

Please comment on Wikipedia talk:User pages

Greetings! You have been randomly selected to receive an invitation to participate in the request for comment on Wikipedia talk:User pages. Should you wish to respond to the invitation, your contribution to this discussion will be very much appreciated! If in doubt, please see suggestions for responding. If you do not wish to receive these types of notices, please remove your name from Wikipedia:Feedback request service. — RFC bot (talk) 11:15, 1 February 2012 (UTC)

Bande à part vs Band of Outsiders

In light of your previous participation in film titling issues, the discussion at Talk:Bande à part (film)#Requested move may be of interest.—Roman Spinner (talk) 11:17, 1 February 2012 (UTC)

Please comment on Template talk:Expand language

Greetings! You have been randomly selected to receive an invitation to participate in the request for comment on Template talk:Expand language. Should you wish to respond to the invitation, your contribution to this discussion will be very much appreciated! If in doubt, please see suggestions for responding. If you do not wish to receive these types of notices, please remove your name from Wikipedia:Feedback request service. — RFC bot (talk) 12:15, 3 February 2012 (UTC)

Please comment on Wikipedia talk:Verifiability

Greetings! You have been randomly selected to receive an invitation to participate in the request for comment on Wikipedia talk:Verifiability. Should you wish to respond to the invitation, your contribution to this discussion will be very much appreciated! If in doubt, please see suggestions for responding. If you do not wish to receive these types of notices, please remove your name from Wikipedia:Feedback request service. — RFC bot (talk) 13:15, 5 February 2012 (UTC)

Please comment on Wikipedia talk:Categorization

Greetings! You have been randomly selected to receive an invitation to participate in the request for comment on Wikipedia talk:Categorization. Should you wish to respond to the invitation, your contribution to this discussion will be very much appreciated! If in doubt, please see suggestions for responding. If you do not wish to receive these types of notices, please remove your name from Wikipedia:Feedback request service. — RFC bot (talk) 13:16, 7 February 2012 (UTC)

Please comment on Wikipedia talk:WikiProject Soap Operas

Greetings! You have been randomly selected to receive an invitation to participate in the request for comment on Wikipedia talk:WikiProject Soap Operas. Should you wish to respond to the invitation, your contribution to this discussion will be very much appreciated! If in doubt, please see suggestions for responding. If you do not wish to receive these types of notices, please remove your name from Wikipedia:Feedback request service. — RFC bot (talk) 14:15, 9 February 2012 (UTC)

Please comment on Wikipedia talk:No original research

Greetings! You have been randomly selected to receive an invitation to participate in the request for comment on Wikipedia talk:No original research. Should you wish to respond to the invitation, your contribution to this discussion will be very much appreciated! If in doubt, please see suggestions for responding. If you do not wish to receive these types of notices, please remove your name from Wikipedia:Feedback request service. — RFC bot (talk) 15:15, 11 February 2012 (UTC)

Your Arbitration evidence is too long

Hello, Born2cycle. Thank you for your recent submission of evidence for the Article titles and capitalisation Arbitration case. As you may be aware, the Arbitration Committee asks that users submitting evidence in cases adhere to limits regarding the length of their submissions. These limits, of User:HersfoldArbClerkBot/Length header/Words words and User:HersfoldArbClerkBot/Length header/Diffs diffs maximum, are in place to ensure that the Arbitration Committee receives only the most important information relevant to the case, and is able to determine an appropriate course of action in a reasonable amount of time. The evidence you have submitted currently exceeds at least one of these limits, and is presently at 939 words and 50 diffs. Please try to reduce the length of your submission to fit within these limits; this guide may be able to provide some help in doing so. If the length of your evidence is not reduced soon, it may be refactored or removed by a human clerk within a few days. Thank you! If you have any questions or concerns regarding the case, please contact the drafting Arbitrator or case clerk (listed on the case pages); if you have any questions or concerns about this bot, please contact the operator. On behalf of the Arbitration Committee, HersfoldArbClerkBOT^(talk) 06:02, 13 February 2012 (UTC)