Template talk:Television Rotten Tomatoes scores

There seems to be a problem
After changing the color in Game of Thrones RT scores S7 to, in order to match the color in the season article, the graph looked kind of weird (diff). Until the problem has been fixed, i am using instead.

Suggestion
Is there some way that we could control the alignment of this template in articles, rather than it only ever appearing in the centre of the page with big chunks of whitespace on either side? It could just be an alignment parametre with the default as is, and then left and right options that have text wrapping of some sort. - adamstom97 (talk) 10:28, 11 October 2017 (UTC)
 * you are the editor that I know that works with templates most. Is this something we could do? - adamstom97 (talk) 03:44, 15 October 2017 (UTC)
 * Apologies; I saw this, then forgot to reply. I'll see what I can do with it today for you. --  Alex TW 05:49, 15 October 2017 (UTC)
 * How's this? --  Alex TW 06:04, 15 October 2017 (UTC)
 * Yeah, that's basically what I was thinking. Just with a bit more padding around the left and right ones I guess, like what you would get around an image or something like that, rather than the text touching the box. - adamstom97 (talk) 06:14, 15 October 2017 (UTC)
 * All done! --  Alex TW 06:24, 15 October 2017 (UTC)
 * Cool, thanks Alex! That looks great. - adamstom97 (talk) 06:50, 15 October 2017 (UTC)

WP:BOLD
Reverting on the basis of WP:BOLD seems curiously contradictory considering the page begins by saying 'Be bold can be explained in three words: "Go for it".' I note that IPs aren't allowed to create template sandbox pages (nice job, Wikipedia). Given that Line chart fails to display at all on the mobile site, and that the mobile site accounts for |line|1-year|access~desktop*mobile-app*mobile-web|monthly half of English Wikipedia's traffic, it would seem valuable to make the graphs available to the whole of the site's audience. The lack of any actual justification for reverting suggests the more relevant page here is in fact WP:OWN. 202.159.169.45 (talk) 18:42, 9 September 2019 (UTC)
 * Well, the changes made affected dozens of articles by an IP who has never edited before (indicating either a sock or a logged out editor), and the edits were undiscussed before implementing, yes, the revert based on BOLD was valid; this is expanded further at WP:BRD. Navboxes aren't displayed on the mobile site either, and there has never seemed to be an issue with that. I would have been glad to discuss this with you (as per the D in BRD), but if you're here just to make accusations, I can now see that you're not here to build an encyclopedia, just here to throw personal attacks. Good day. -- / Alex /21  01:12, 10 September 2019 (UTC)
 * Navboxes aren't really part of the article though. They were disabled because they work even more poorly on mobile than they do on desktop. It's apparent you haven't actually checked the mobile site. If you had, you would see that the template's table does display; it is only the graph that doesn't display. This means mobile pages say 'Percentage of positive reviews tracked by the website Rotten Tomatoes'. But the parentages are not provided, because they are communicated via the graph. Which is missing. Invisible. Absent. Not there. So "your" template doesn't just make the pages incomplete on mobile, it makes them confusing and broken.


 * It is impossible to discuss what it is you find problematic about the use of the new graphs when you haven't stated what that is. I was bold. You reverted. But your explanation for doing so was essentially "I'm reverting because I'm reverting", leaving nothing of substance to discuss. And after your reply this is still the case. Sock puppetry should be taken seriously, so I hope you'll be opening a new case at WP:SPI to allow an investigation to proceed. 202.159.169.45 (talk) 20:23, 10 September 2019 (UTC)
 * No need. I'm arguing with someone who isn't here to build an encyclopedia, just here to make personal attacks. Thank you for your contributions! -- / Alex /21  23:27, 10 September 2019 (UTC)
 * I would like to take this opportunity to say that EasyTimeline doesn't display on Television ratings graph either, as i had reported in February 2019. I am not an experienced template editor, but i believe using a Graph:Chart template has been suggested as a solution to this problem. I took a quick look at the documentation page before Alex 21 reverted your edit and i didn't see anything wrong with it (i actually preferred it to the previous version). Basing the revert on BOLD is not a valid reason. BRD doesn't say that a bold edit should be followed by a revert. It says that a revert after a bold edit should be followed by a discussion. Failing to explain why the edit was reverted though, didn't provide the IP user with a subject for a constructive discussion. Regarding the navbox comparison, and as the IP user pointed out above, there is no text introducing these templates, contrary to this one and the ratings graph template i mentioned before. This disrupts the page layout and mobile readers will find it confusing. Radiphus (talk) 23:28, 10 September 2019 (UTC)
 * "It says that a revert after a bold edit should be followed by a discussion." Indeed, that's why we're here right now, doing what we do best: discussing. I'll see if I can find the correct template to hide this template on a mobile viewer; same as a navbox. This template has existed for three years without any issues; I doubt there's going to be any anytime soon. -- / Alex /21  23:31, 10 September 2019 (UTC)
 * Is that what we do best? Discussing on a template's talk page about sock puppetry suspicions? You didn't do it in your edit summary, so now this is the appropriate place to explain why you don't like the change the IP user made to the template. If there is no reason and the only other alternative is to completely deprive mobile readers of valuable information, i am inclined to restore the IP's version of the template. I am off to sleep. Radiphus (talk) 23:46, 10 September 2019 (UTC)
 * No. You stated "It says that a revert after a bold edit should be followed by a discussion." That's what's happening - "a revert after a bold edit", a bold edits was reverted, "should be followed by a discussion", which is what is happening here. The suspicious IP account was just an interesting note, an IP who's never edited before and comes with unsupported claims that "WMF want to remove EasyTimeline". Says who? Where?
 * Again, this template has existed for three years without any issues, even after thousands, if not millions, of readers have accessed pages using this template and not raised an issue about it. Even your February 2019 discussion has no replies or concerns raised. It seems that the Phabricator report is still underway, so that should conclude first before we make such hasty decisions. Looking at the "new" code, it seems that the image is not generated in the graph.
 * By the way, the edits and reverts were almost two weeks ago, and you saw the edit before the revert, but are only just commenting now? Was there a reason for the delay? Also, if you aim to revert, I recommend gaining a consensus first. -- / Alex /21  13:08, 11 September 2019 (UTC)
 * And WP:SQS says that when you say "discuss first" you are expected to explain your revert, otherwise the other editor may find it difficult to assume good faith. We are currently having a discussion about having a discussion. This is not building consensus, it's stonewalling and BRD misuse.
 * It's the first thing you noted in this discussion, and the only possible valid arguement you have presented so far. However, this talk page is not the appropriate place to talk about SP, and your unproven suspicions should not affect the template and the reader's experience on this site.
 * What's with the ostrich effect? Are you actually questioning that there is currently a problem with this template (and EasyTimeline in general) on mobile view?
 * So, we should continue smoking, until a cure for cancer is found? What kind of logic is that? Shouldn't it be the other way around?
 * I don't know why the graph doesn't display in the old revision directly. You can see a working version if you try to edit the source of the old revision and where it says "Preview page with this template" put "Template:Television Rotten Tomatoes scores/doc" and hit "Show preview". This is what you will see.
 * Yes. There was no discussion until the IP started one two days ago. Who would have thought, right? Are you asking me why i didn't open the discussion myself Alex? Radiphus (talk) 17:09, 11 September 2019 (UTC)

''NOTE: there was an edit conflict when I went to save this page. I haven't changed this reply to account for Radiphus' latest reply.''

Are you really suggesting 'No need. I'm arguing with someone who isn't here to build an encyclopedia, just here to make personal attacks. Thank you for your contributions!' constitutes a discussion about the content of the template? Because it's really more of a refusal to discuss.

Anyway, as you have, at last, raised some issues that can be discussed - though not in replies to me! - I shall attempt to discuss them:


 * WMF want to remove EasyTimeline: MW:Extension:EasyTimeline says 'The Graph extension aims to replace this extension' and there's a ticket to do this (created by a WMF employee, I note) at https://phabricator.wikimedia.org/T137291. While WMF internal politics/ADD/incompetence means ET isn't going to vanish anytime soon, 'it's the way it's always been done' is a poor argument given ET stuff will at some point have to be migrated anyway.


 * The bug you've found with the image generation is caused by (the graph extension version of) the graph being absent from the current revision. The broken image can be fixed by looking at the page in preview mode. I think this might be related to the extension's behaviour where the version of the graph in page preview is slightly different from the one when the age is saved. Note that if the graph is still present in the current revision, it will still display when looking at a prior revision, for instance at https://en.wikipedia.org/w/index.php?title=Internet&oldid=913175238#Performance. But of course, since the ET stuff has returned, that isn't the case here.


 * As for no one reporting the problem in three years...so what? An unreported problem is still a problem. And it speaks volumes about how little most Wikipedia editors care about how articles look and work on the mobile site. As you said, 'thousands, if not millions, of readers have accessed pages using this template'. The response of the editing "community"? Crickets. And it's no longer true that no person has identified the problem, because I have. But I didn't just complain and leave it to others to do the work, I fixed it myself per WP:sofixit.

'Tis regrettable you don't want to report me to WP:SPI. 202.159.169.45 (talk) 18:38, 11 September 2019 (UTC)


 * Seeing as Alex's terse non-replies have now turned into complete silence, I'll proceed to the next stage. I'll shortly be reinstating my edit. I am certain this will be reverted by Alex 21 with claims that I need to obtain broader consensus. The idea that consensus needs to be established to fix a serious problem that makes the template useless to half the readership is absurd, but establishing broad consensus is what Alex 21 will demand. This will of course not be an attempt to OWN, filibuster or dissuade, but a sincere attempt to protect the template from being useful to halt the readership. Given Alex will insist that broader consensus must be obtained and there is absolutely no chance of Alex 21 proving me wrong about that - and thereby finding themselves emerging victorious, I have already prepared material to begin a larger discussion. 202.159.169.45 (talk) 19:16, 14 September 2019 (UTC)
 * Apologies. Back! Can you back up your claims that the "useless to half the readership"? I'm not seeing any comments, complaints or discussions. You think it's useless. That's what you need a consensus for. But hey, keep putting words in my mouth. I look forward to your discussion! (Hopefully you can even do it logged in for a change?) -- / Alex /21  23:06, 14 September 2019 (UTC)

Errors
Template 1: Previously working template

Template 2: Faulty template

Per the documentation, included is One parameter for each episode, and parameters may be included but not set - these will include an entry on the episode axis for the episode, but the entry for the episode will remain empty (i.e. for currently-airing seasons). This is no longer the case. This can be seen above in Template 1. For the above example in Template 2, the template has ten parameters ready for values, but only one is set (given that only one episode has aired). All episodes should be displayed on the x-axis, with empty values for episodes 2 through 10. This is no longer the case. Now we have a ridiculous view of only one episode and one value. Fix this, else the template will need to be restored to its previously-working state. -- / Alex /21  12:45, 19 September 2019 (UTC)
 * Please create a sandbox page for this template (use the link in editing mode). As mentioned, unregistered users aren't allowed to do that.202.159.169.45 (talk) 22:55, 19 September 2019 (UTC)
 * ✅ -- / Alex /21  01:43, 20 September 2019 (UTC)
 * I created Template:Television Rotten Tomatoes scores/testcases for you as well. -- / Alex /21  02:23, 20 September 2019 (UTC)

I've created a new  parameter (which can be removed once the season has finished) to accommodate this. It inserts scores of 100 for the number of episodes specified and is set to be fully transparent & thus invisible. The problem is that the symbols don't honour the opacity settings. I've posted messages at the Graph:Chart talk page and VPT to alert editors to this problem. Anyway, here's the current state of play.

Using :

NOT using :

202.159.169.45 (talk) 04:05, 20 September 2019 (UTC)
 * Good solution, but best to wait to see if there's a solution for the opacity until we implement anything. -- / Alex /21  04:07, 20 September 2019 (UTC)
 * 2 days; any updates? -- / Alex /21  07:08, 22 September 2019 (UTC)
 * For full opacity #FFFFFF should be used as the color instead of #3C230D. Partial opacities, like the 25% example on Template talk:Graph:Chart is an different matter. I could go into the details, but there is no point.--Snaevar (talk) 09:00, 23 September 2019 (UTC)
 * , white would not work, given that the background of the graph isn't white and the white dots along the top would become obvious. See the above examples; I've restored the "fake" 100 scores and you can clearly see the white dots. We need full transparency, which does not seem to be supports with the showSymbols parameter as of yet. -- / Alex /21  09:08, 23 September 2019 (UTC)
 * Nevermind; I've modified the sandbox further and found a hack-fix for our problem. -- / Alex /21  09:28, 23 September 2019 (UTC)

Data presentation
I've noticed a few articles using this template and I have fundamental concerns that this is an inappropriate and inaccurate way to present the data. Specifically I don't think this is an appropriate use of line graph, because of the way it joins episodes and discrete chunks of information together. A bar chart seems like it would be a more appropriate way to present the same data.

I strongly suggest reading "The Visual Display of Quantitative Information" by Edward R. Tufte, or any data scientist or statistician, or any kind of relevant expert at all. -- 109.78.215.166 (talk) 16:57, 24 November 2020 (UTC)
 * I realise this was initially mentioned a couple of years ago without anyone else commenting, but I completely agree. A line graph is inappropriate for this kind of data - it should be a bar chart. (Hohum @ ) 16:04, 1 March 2023 (UTC)

Dashed red line at 97
Does anybody else get a dashed line when hovering over 97? Emir of Wikipedia (talk) 20:14, 21 February 2021 (UTC)
 * , that's intentional, to set up positions in the graph for blank values. Previously, it never lit up red and just stayed white, that's new and (I'm guessing) a result of the graphing template. -- / Alex /21  13:08, 6 March 2021 (UTC)

Number of episodes to be increased?
It seems you can only add up to 20 scores, I'm wondering if this can be increased? This will be an issue for The Walking Dead (season 10) as the season will now contain 22 episodes as opposed to 16. Thanks. Drovethrughosts (talk) 20:52, 22 March 2021 (UTC)
 * Pinging you as you're the creator of this template and I'm not sure if this page is on your watchlist. Thanks. Drovethrughosts (talk) 16:06, 24 March 2021 (UTC)
 * , ✅ -- / Alex /<sub style="color:#008">21  20:29, 24 March 2021 (UTC)

Compressing the Y axis makes this template useless at a glance
This is generally an excellent template with one major flaw: It compresses the Y axis without strongly emphasizing that it has done so. Unless every episode or season has a perfect score, by compressing the Y axis it gives the reader the sense that there was a drastic dip in critical approval at one point. Only by viewing small, unemphasized text on the left side does one realize that it's compressed.

Given that these graphs serve as eye catching shorthand to the casual viewer, I think this needs to be addressed. Some possible solutions:
 * Just do away with compressing the Y axis entirely.
 * Add bold to the top and bottom values of the Y axis, and maybe even make them slightly larger.

Thoughts? Other ideas? -- sarysa (talk) 22:20, 3 May 2021 (UTC)
 * Neither of these options are possible, as the graph (and hence, y-axis) are generated through the standardized template Graph:Chart. -- / Alex /<sub style="color:#008">21  23:59, 7 July 2021 (UTC)
 * There seems to be a parameter called yAxisMin in the Graph:Chart that could help with showing the whole Y axis. —El Millo (talk) 00:51, 16 July 2021 (UTC)

I think this graph is generally a terrible idea and far worse than "useless at a glance". The distortion caused by the compression of the Y-axis is merely yet another reason why it is terrible and should not be used. There's "Lies damned lies and statistics" but this graph never even says anything interesting or insightful, a good distortion of the data would at least be interesting. It's bad science and should not be acceptable in an encyclopedia on that basis alone, but most editors should be able to recognize that the graph isn't actually being used to saying anything worthwhile. Maybe this template could be salvaged but I can't help thinking it is fundamentally flawed. Is there someone from WikiProject Mathematics with a background in statistics who could take a look at this and better explain why I'm wrong and it isn't actually terrible (or not)? -- 109.79.161.25 (talk) 00:38, 16 July 2021 (UTC)
 * This isn't related to science whatsoever. This is just a way to show the Rotten Tomatoes scores individual episodes got within a season of a TV show. —El Millo (talk) 00:47, 16 July 2021 (UTC)
 * Maybe I'm skipping a step and not explaining clearly enough but statistics is a science, and bad graphs that don't say anything in particularly are bad science. An encyclopedia should not include bad graphs that do not serve a clear purpose. (Similarly it would be bad writing to include a paragraph listing all the Rotten Tomatoes scores for every episode, unless there was some comment being made about those scores.) Graphs should be used to make a point, to say something particularly meaningful. Different graphs should be used for different sets of data. For example a good line graph might highlight an upward trend, or sudden downward trend. This graph doesn't meaningfully encapsulate the fact that the critical response from each episodes is separate from the last, and the percentage is based on a different number of reviews each time. If for example it was a graph of Nielsen ratings you might implicitly be making the point that a bad episode would cause lower ratings the next week, and in that case the slope might have some meaning. The slope connecting each RT score to the next episode is not meaningful, the RT scores might be better presented as a bar chart. Back to the original point made by sarysa, if the purpose of the graph is to show that the review scores are consistently high then compressing the Y axis undermines that (but maybe a horizontal line at the 60% fresh/rotten threshold might somewhat compensate for that distortion, it depends what the underlying purpose of the graph). A simple table of the score data would be dull but it would be clearer and simpler, something like the Television critical response table.
 * I don't need to be a chef to tell you that this does not smell right and that it needs improvement. Maybe someone behind Template:Graph:Chart could explain how to make better use of graphs for this very particular dataset. -- 109.76.132.146 (talk) 03:06, 19 July 2021 (UTC)
 * These graphs do serve a purpose of displaying episode scores in a much more readable and understandable way than a list in prose. If we could enforce a minimum value of 60 I think that would be an easy solution to the biggest issue with them rather than making a drastic change to a lot of articles. - adamstom97 (talk) 03:05, 19 July 2021 (UTC)
 * My point is that the graphs are not meaningfully "readable" or "understandable", they are inconsistent and distorted on the Y-axis for one, and the slope does not mean anything. We cannot even use them make reasonable comparisons between different season of the same show. Doing something to address the Y-axis would be a good start. I'm no expert but I sincerely believe that an stats and graphs expert should be consulted if this type of graph is going to be widely used in an encyclopedia. -- 109.76.132.146 (talk) 03:15, 19 July 2021 (UTC)

This template is basically a specialized version of Graph:Chart. That template has the parameter yAxisMin, which is filled with code that makes it automatically adjust the Y axis to the values included. We can very easily just add the possibility to override that if desired and manually put the Y axis minimum, just to be able to do it in specific articles. If we want that to be the new default, however, we must have an RfC for it. —El Millo (talk) 03:49, 19 July 2021 (UTC)
 * I implemented it at Template:Television Rotten Tomatoes scores/sandbox. Examples collapsed below:

Here's an example with the bottom at 60:

and with the bottom at 0:


 * —El Millo (talk) 04:29, 24 July 2021 (UTC)
 * I don't think having the bottom at 0 is going to be helpful for series where the scores are mostly high. The version with the bottom at 60 solves the issue we are talking about while still being readable, so I would support making that change. - adamstom97 (talk) 05:08, 24 July 2021 (UTC)
 * And for series with more mixed scores, do you think the bottom at 0 would work there? Or should we just leave it automatic there? —El Millo (talk) 05:51, 24 July 2021 (UTC)
 * Now that I'm thinking about it, if there's consensus for it we could make it automatically be at 60 at the highest. That is, from 60–100 if all scores are within the range, and automatically becoming lower with the equation currently used. —El Millo (talk) 05:54, 24 July 2021 (UTC)
 * Yeah I think it should just move down from 60 with the lowest score. - adamstom97 (talk) 06:27, 24 July 2021 (UTC)
 * How do we go about implementing this? Do we wait for other editors or do I just WP:BOLD it and see if anyone opposes? —El Millo (talk) 18:49, 24 July 2021 (UTC)
 * Just because I was seeing this as a problem at Watchmen (TV series) (where the bottom is 90 due to the high scores), might I suggest adding a user-entered lower and upper bound parameter for the axis, otherwise defaulting to the current behavior? I agree that where most problems are, using 60 as the "maximum" lower bound is a good place, but having a bit more end-use flexibility would be good too. --M asem  (t) 14:21, 1 August 2021 (UTC)

A little late but I did it. Templates with only fresh scores (60–100) will not go above 60 as the minimum value of the y axis. Templates with scores below 60 will function as it did already, the minimum set 10 numbers below the lowest score. yAxisMin can be used to customize the number as long as the value entered is lower that what would automatically be set. —El Millo (talk) 03:18, 6 November 2021 (UTC)

Continued disturbance of sitewide articles
This template has been continuing to not function on all integral television and movie media articles. I do not have the coding experience to suggest changes, but I suspect this template will be a future liability without action to find, address and fix code problems. I do not know who this will go to, but I hope that this may inspire action. 3vvww661 (talk) 23:31, 30 June 2023 (UTC)


 * It's an issue with all charts on Wikipedia, not just this one. - adamstom97 (talk) 00:27, 1 July 2023 (UTC)