| Well we should all just be using JST. If we don't then shows will get lost in transition. |
  |
| Uhm I don't think that's the problem here given that this was a day after you took over, right? Also that wouldn't happen anyway; that could only happen when you went from JST to a more western timezone, not the other way around That aside I still firmly believe that doing it in your own timezone and focusing on the airing times rather than the dates is just so much more efficient, not to mention convenient for yourself |
  |
| Well I disagree as my logic is telling me that if were not on the same timezone then shows will get lost in transition. |
|
| If you're going East then you're going forwards in time, not backwards. In other words you'd perhaps end up wanting to update shows that have already been updated, but there's no way you could end up with any series getting lost and ignored in-between the shift. If you're going west however then it's possible (actually happened to me during my very first day since I assumed Zee had updated some shows which he apparently hadn't because of this) Actually now that I think about it... if you want to update all the shows that aired during the same day in JST, and still keep all club updates in the 48-72 hour range, then during many of the weekdays you literally only have a 30-minute timespan which you can do the updates on, in other words between 23:30 and midnight JST. Otherwise you'll either pass the deadline or miss out on some shows for that date Also I still find it kind of bothering that a show airing at 23:30 will be updated 24 hours before one airing just half an hour later Really it just feels like you're being given so many awkward problems that would be completely null and void if you just used your own timezone instead. In that case all you have to do is update everything once per evening and that's it. Since almost all anime air during the night in Japan it'll get split across two different dates depending on whether it's before or after midnight, but by using a timezone further west instead then all the anime will end up on the same date since even the latest series will be before midnight your time. So therefore there won't be any issues of this nature to even have to think about to begin with. Just feels like a whole lot less work for everyone that way. I had this exact same discussion with Xia on Skype when I asked him to help and we both agreed on this logic at least |
HaXXspettenAug 24, 2015 12:13 PM
| |
| Well that aside, why are there red plusses :s Also we decided to not put up/down arrows for series who get their position changed solely as the result of another series' update. So if one series gets a positive update which pushes it above one that was previously ahead of it, only the new leader should get a ▲[+1], the old one should still be at ▶, not ▼[-1] |
| |
| If something goes up than something has come down. That is logic. If you want to add in this new feature than we need do it right. If not, get rid of it. |
|
| Personally I don't mind it. It makes sense to have it but by not doing it the way I'm going it right now defiles logic and the only reason to do it the other way is because you want to do less work. That's it. Which, you know. It did take me over an hour to the update but if we have this new feature we need to do it right. |
|
| About the time zones, i did some testing using GMT 00:00 London, for the shows we're recording this season, the earliest airs at roughly 9am and the latest at 10pm, which means that your timezone should fall between GMT -09:00 and +02:00 in order to avoid shows falling into different days, as long as everyone is within this range then we can go ahead and have everyone use their own timezone, if not we need to decide on one timezone we should all use within that range. About the ranking arrows, at the start of the season there are sporadic changes happenings in the ranks, a show could easily be at the bottom one week and rise to the top the next day, this means that every show in between would have a red arrow facing down and just makes the whole table messy, doing it the right way usually works in sports tables and such because they get updated about once or twice a week, we do ours everyday. To compensate for this is why we're using the updated sign to draw attention to a show that gets updated that day with it's change in score and rank, doing it any other way will result in a table that is convoluted and very confusing. |
 Join the MSP Club for an in-depth look at score progressions and other stats of currently airing anime |
| I actually test it out and yes, JST does screw everything up. Everyone else's Tuseday Is Overlord, Sore, Monster, Bikini Warriors JST is Akagami To-Love Ru Teekyuu Jitsu Non Non Overlord Sore Makura So by actually doing the research and presenting it. JST should be outlawed to be used. |
|
| I can make a pretty easy distinction on why JST is a bad idea: If you look at Livechart for Saturday for example (in JST), that'll start with IDOLM@STER Cinderella Girls at 0:00 JST. Then there are a bunch of shows airing right after it, eventually ending with Symphogear at 3:10 JST. Now fast forward to the evening and we have Denpa Kyoushi airing at 17:30 JST, then a few more shows, eventually ending with Durarara at 23:30 JST. Now, the club rules state that everything should be updated 48-72 hours after its time of airing in Japan. If we now insist on updating everything on a day-to-day basis, this means that Idolmaster, which aired at 0:00 JST on Saturday must be updated at the very latest on 0:00 JST on Tuesday. Any later than that and it'll have passed the 72-hour mark. At the same time, Durarara which aired at 23:30 JST on Saturday may be updated no sooner than 23:30 JST on Monday. Any earlier than that and 48 hours won't have passed yet. So, if we put these two together, what can we conclude? If you want to update every day individually, then by using JST you must update all of Saturday on Monday, after 23:30 JST but before midnight. That only gives you a mere 30 minutes where you're even allowed to do the update according to our own club rules. By using a western timezone you can totally eliminate the midnight issue. Alternatively you could drop the idea of updating strictly one date at a time. You don't have to update one date at a time according to the club rules after all; all it says is that you have to stay inside the time range. But of course if you use western time then that'll be a non-issue anyway since all the updates will come comfortably in time before the deadline, no problem. |
| |
| or just look at the site were using like i did >.>. |
|
| Well either works, same function Though I still prefer Livechart over MAL's seasonal schedule as of yet. I've got two userscripts running on MAL's to make it a bit better but it still needs more work before I'll consider switching |
| |
| Well we officially have our first banner and I already took the liberty in being the first affiliate to the club -.^. |
|
| So before anyone asks, I am using EST now but I still need to pick up the shows that were lost yesterday from the JST. So that is why there is more shows than usual. |
|
| As long as they stay withing the allowed update time then that doesn't matter I usually update Gintama, World Trigger and DBS individually just because they air at weird times. No need to be inflexible when there's no need to |
| |
| So can I get confirmation that God Eater did not air as I eat lunch before doing the lists. As there is already 5 weeks updated and Livechart says it is episode 6 airing next. |
|
| Yeah it was on hold this week again It's seriously been delayed 3 times over a period of 8 weeks. That's just not acceptable >.> Especially given that the show sucks horribly anyway |
| |
| So someone forget to update Dragon Ball Super last week on list 3, so I just grandfathered it in and explains why there is such a big drop in movement. |
|
| That's probably my mistake in that case.. I think |
| |
Challenging Rule  |
|
| By changing the list with the positions in our update, this has changed and evolved the way that the positions should be handled on the lists and makes more evolution of club necessary to provide fairness in our ties that the current system lacks. There needs to be protection to the shows that have been going up for more than three weeks to keep their top spots against something that just started to go up for three weeks. HaXX said that it doesn't matter because our list are only tracking three weeks but I think that is completely the wrong way to look at it. Our lists are to show what shows are trending the best up and for something that has been going up for 5 weeks straight to be passed by a show that is going up for 3 weeks straight. Does not represent that. |
|
| If you want to use that reasoning then MSP score should be something you just add on another symbol every week instead of focusing on the last three. In that case of course a show with ++++++++ would be better than one with -----+++ but that's not we've decided to do here, right? I don't think it makes a whole lot of sense to first strictly say that we're only tracking the last three weeks, but then use the previous weeks as a tiebreaker. Because in that case then we actually are tracking every single week even if we only show the results of the last three. You can't both have the cake and eat it. With that same logic then ++++--- is better than ----+++ and that just feels wrong to me |
| |
| Well at a certain point there needs to be a cut-off and three weeks is that cut-off. But that completely changes when there is a tie because in three weeks they have been on the same ground so by then we need to go beyond that to find what belongs above each other and by making the new guy have to be on the bottom and not the top makes sense and is fair to all the shows. It's pretty damn simple HaXX, I'm not sure why your fighting me so hard on it. |
|
| By not doing it my way does not incorporate fairness in what we are doing and devalues us as a statistical haven on MAL. It makes our second list defective that something that only been going up for 3 weeks straight is would be above something that has gone up 5 weeks straight and does not give the right information to our members on what has been the best show trending. So it is also giving a false message to them and makes our efforts flawed. |
|
| My point is that we already have a functioning system. If we're only looking at the last three weeks and have the same MSP, then just look at what series have gotten the bigger plusses. In other words a +++ consisting of +0.01 +0.02 +0.01 would be worth less than a +++ consisting of +0.03 +0.04 + 0.04 etcetera. Hence why you just look at the Score Change table in that case instead since whichever series is above on that has evidentially had a bigger increase than the other. Like you said yourself, "Our lists are to show what shows are trending the best" and thus whichever show go up more with the same MSP is the one which is trending the best. So I don't see why you need to include facts based on earlier weeks when you already have a perfectly logical way of sorting it based solely on the weeks we're actually tracking Don't get me wrong, I'm not really against your idea either. I just think it's needless since we already have a working system that actually stays within the time period which we're actually listing Though I will agree that the MSP and Score Change lists oftentimes end up looking disturbingly similar with the way it has worked up until now as a result of this, so I guess that might be a better reason to change it |
| |
| But nothing your saying challenges me back that a show that has gone up something that only been going up for 3 weeks straight is fair to be above something that has gone up 5 weeks straight and the points of the wrong message being sent as send on reply 175. This is not needless if it provides fairness and the right message to our members on what is trending up. |
|
| Because let's consider this scenario: Anime A Week 1: +0.01 Week 2: +0.01 Week 3: +0.01 Week 4: +0.01 Week 5: +0.01 Total: +0.05 MSP: +++++ Anime B Week 1: -0.01 Week 2: +0.02 Week 3: +0.04 Week 4: +0.05 Week 5: +0.05 Total: +0.15 MSP: -++++ With your system, A would be considered trending better than B. And I just can't see why that would be the right way to look at it. B is clearly on a way more impressive hotstreak than A here |
| |
| Well now this discussion is turning more into challenging the concept of the MSP entirely as 5 pluses over 4 pluses- does put Anime A above anime B. So before we get any further are we in agreement that at least with the first and third list that the anime that achieved the score of the tie first deserves to be above the one that just achieved that score recently? |
HibbingtonAug 25, 2015 3:02 PM
|
| Well... I have never done that at least. For Score Change I guess I'd be fine with it though because it doesn't really have any good alternative either. For MAL Score though... that's a pretty terrible idea. If one anime was initially ranked higher but is currently plummeting (like Dragon Ball Super atm) and another one started lower but is climbing steadily (like let's say Prison School then since those two are almost tied), then I think the latter one should definitely be considered better. Which date you happened to update the respective series should be less relevant than their actual trends. If there's a tie in MAL Score at least then I would definitely go by the Score Change list as a secondary tiebreaker. There is no way it's fair that a series with -0.36 over the last three weeks can be considered 'better' in that case |
| |
| Well that way of thinking is also not fair. Let's say a show has been going down three weeks. Anime A -0.02 -0.03 -0.02 and is sitting at a 7.89 Then we have this show going up three weeks Anime B +0.02 +0.02 +0.01 and is sitting at a 7.89 I believe the show that is going down should still be over the anime that just achieved being a 7.89 in a tie breaker. Just because the show is trending down doesn't mean that next episode Anime B might stall out and Anime A goes back up +0.05 as we've seen in this season. Your are right that we do update things daily on what airs but that even MORE helps my argument as the show that is getting contested should have a chance to fight back before being overtaken by the other show in the tie. |
|
| Well of course we don't know what's going to happen next week, but that's for the next update to reveal in that case, not the current one. The current one is solely based on the last three weeks, and in that case I think it's only logical that a positive current trending should be considered better than a negative current trending as a tiebreaker. And even if you want to think about next week, that also suggests that it's much more likely that the series going upwards will be higher than the one going downwards once next week comes around Well whatever this isn't going anywhere. I'll just leave it for Zee to decide |
| |
| Where I'm getting at is there needs to be a absolute pass in position for something to dethrone the other anime's position. Even if it was Anime A was 7.81 last week on let's say week 7 and Anime B was 7.77 and now they are both tied at 7.89 on week 8. Anime A should have the advantage as there has been no absolute pass in position and not the other way around like current system dictates. Anime A deserves protection of its position until proven guilty, regardless of what show gets updated first. |
|
| Whenever you have a list like this in any sort of point-based ranking then there are different kinds of tiebreakers but I don't think I've ever heard time being relevant. Like in a sports division then if you have the same amount of points then it'll be decided by goal difference, goals scored, head-to-head results etcetera, but which team had the position first or last doesn't matter at all. And in the same way I don't see why that should be the case here either. Of course in a sports division then whether two tied teams won or lost their last games doesn't matter either, so in that sense the "trending" we're using is also irrelevant. So the real problem is that we don't really have any good equivalent to goal difference and stuff in an anime ranking like this, hence the awkwardness. You could look at it totally differently and look at popularity instead: whichever show has the more viewers could be considered better since it manages to appeal to a wider audience while still maintaining the same rating. Or you can just go by MAL's official ranking and see which show it ranks higher according to its rating formula Idk I'm just brainstorming. Either way I'm going to bed now so I'll have to put this discussion on hold |
| |
| But when there isn't something like goal difference than what was there first should stay first until being properly dethroned by an absolute change, and there doesn't need to be something like goal difference for the MSP. The simplest and fairest way is what I'm proposing by there needing to be an absolute change for something to get passed in position. If Soccer/Football didn't have goal difference, there would just be a 1 game playoff anyways. But Anime Shows aren't living things. We can't ask them to face against each other and knock the other one down. Absolute Change is what should be adopted in. |
|
| Current Rule Ties: If there is a tie in any of the categories, the show will further be ranked on a secondary level in the following order: Score Change, MSP Score and finally MAL Score until at least a difference is arrived upon. Proposed Change. Ties: If there is a tie in any of the categories, the show that has been in higher position has advantage over the the newly updated show until an Absolute Change has occurred. |
|
| Well I'm going to sleep now but I've gotten a pretty good idea I think. I'll post it tomorrow morning or so |
| |
| Well okay first of all, just to add to what we last talked about; even if our list doesn't have any good parallels to stuff like goal difference, matches/sets/legs won, head-to-head record, away goals, penalty points and stuff like that which you can find as tiebreakers in various sports, there is still one thing we can learn from it. And that is that all forms of tiebreakers like that are based on the accumulated results, not the journey to that point. In other words it doesn't matter what happened last week, last month or during the first round of the season; all that matters is the overall statistics in that exact moment So if we are to draw a parallel between sports and our list, then what we can deduce is that we shouldn't think about what anime had a certain score first, or what its current trending is or anything like that. We should be looking at secondary statistics. Now of course since anime isn't a contest in the same way, we don't have the same kind of obvious statistics to look at, but there are in fact some things that would make sense I think. For example, like I already mentioned I think popularity would be a logical secondary statistic to use as a tiebreaker. If two anime have the same score, but one of them manages to do so whilst appealing to a much larger audience, then I think that's more impressive and thus could be an argument for ranking it higher. Another thing that is most certainly real on MAL is the so-called sequel effect, as every sequel will get an artificial overrating by default. So you could argue that if one anime's currently airing first season is tied with another anime's currently airing second season, then the first one should be ranked higher since it's a much more impressive statistic. Same way if you compared a second season to a third season, etcetera. Another possibility would perhaps be to look at the episode number; if an anime has a certain score after 10 episodes whereas another anime has the same score after 2 episodes, then it's pretty clear that the first one has a much more stable and reliable rating, and should thus be considered better (I'm not really a huge fan of that particular idea but it's a thought at least). Or you could of course simply look at the series' respective overall MAL rank to see which one the site itself ranks higher. The only problem with that would of course be the fact that that list is only updated every 12 hours. Again, this is just me brainstorming here and I'm not sure it's really a better solution or not but it's a new way of looking at it at least, so it might be worth giving some sort of consideration Anyway that's not really what I wanted to talk about here; when I said that I had an idea I wasn't talking about the MAL Score list, but rather about a potential overhaul to the entire MSP Score list and how it works. If you do it this way then:
So what does all this mean exactly then? I got the idea from looking at the MSP Score we have listed next to every series in the data thread. That MSP Score is essentially just the sum of the total number of positive and negative score changes over the course of an anime's entire airing. So I was thinking how about you do the same in the club layout... except you only count the current trend, not the entire airing. To be precise, it'd mean that an anime's MSP Score would be equal to the number of weeks which an anime has gotten the same kind of score change in a row since it had a week of the opposite type, or alternatively since the anime started airing. Weeks without a change in score are simply ignored. By doing this, you incorporate the exact amount of weeks which an anime has been trending (like Hibb wants), but the entire list will have much more accurate, diverse and interesting data Just to clarify, let's take some examples from the data thread and show you what I mean: Gate: Jieitai Kanochi nite, Kaku Tatakeri Week 01 - 7.64 Week 02 - 7.65 (+0.01) Week 03 - 7.69 (+0.04) Week 04 - 7.74 (+0.05) Week 05 - 7.76 (+0.02) Week 06 - 7.79 (+0.03) Week 07 - 7.82 (+0.03) Week 08 - 7.87 (+0.05) Gate has had a consistent positive trend ever since the anime began. In this case the MSP Score would simply be [+7] since it has gone upwards for 7 weeks straight Gangsta. Week 01 - 8.16 Week 02 - 8.23 (+0.07) Week 03 - 8.26 (+0.03) Week 04 - 8.29 (+0.03) Week 05 - 8.30 (+0.01) Week 06 - 8.29 (-0.01) Week 07 - 8.28 (-0.01) Week 08 - 8.28 Gangsta had a strong start, but has lately been going down a bit. In this case, despite the overall greater number of positive score changes compared to the number of negative score changes, the MSP Score would be [-2] because it currently has a trending of two negative score changes in a row. The very latest week had an unchanged score however, and will thus be totally ignored (just pretend it isn't there) Okusama ga Seitokaichou! Week 01 - 7.04 Week 02 - 6.92 (-0.12) Week 03 - 6.86 (-0.06) Week 04 - 6.83 (-0.03) Week 05 - 6.83 Week 06 - 6.85 (+0.02) Week 07 - 6.85 Week 08 - 6.86 (+0.01) In this case, Okusama started out poorly but since then has slowly staggered upwards at least a little bit. In this case the MSP Score would be [+2]. It has been 4 weeks since the last time it had a negative score change, but only 2 out of those are actually positive ones. The other 2 were unchanged, and shall hence be ignored Teekyuu 5 Week 01 - 6.99 Week 02 - 6.97 (-0.02) Week 03 - 7.01 (+0.04) Week 04 - 6.98 (-0.03) Week 05 - 7.01 (+0.03) Week 06 - 6.98 (-0.03) Week 07 - 6.97 (-0.01) Teekyuu has a rather... strange graph to say the least; going up and down haphazardly every week. Here, the MSP Score will be [-2] however, because it's currently on a negative 2-week streak despite the overall chaos Gakkou Gurashi! Week 01 - 7.70 Week 02 - 7.76 (+0.06) Week 03 - 7.78 (+0.02) Week 04 - 7.78 Week 05 - 7.76 (-0.02) Week 06 - 7.76 Week 07 - 7.76 Finally, let's look at Gakkou Gurashi which also has a very unusual graph in the sense that it has barely moved whatsoever since it started airing. It would have an overall MSP Score of [-1] however, because it does still have one negative score change as its current trend since all the weeks without change are ignored In other words, this system would be based on an anime's current trending directly and nothing else. And given that that's kind of the point of the MSP Score list to begin with (or so I presume), I think this change would be a step in the right direction. Plus like I said, with the current system it ends up looking awfully similar to the Score Change list more often than not which makes the data itself a bit redundant, but that will no longer be the case with this You can compare this idea to sports once again, in other words by looking at streaks. "Manchester United is on an impressive 7-game winning streak", "Southampton is on a depressive 4-game losing streak" etcetera. Those are the kind of things people normally talk about when looking at a team's overall current form. Of course they might look at how many goals they scored in the last few matches, how dominant they've been playing etcetera but it's really the consistency in the results which speak the loudest, and this is essentially the same thing Of course this is still just a theory and I'm sure there are plenty of things that could be changed. For example you could incorporate the weeks without change in some way to make it more along the lines of "weeks without victory/defeat" instead if you think that's an improvement. And though it should go without saying, I'm obviously not forcing this idea on you, and if you find any obvious problems with it that I haven't thought about then please share them. It's just a proposition of something I came up with before I fell asleep last night after all |
HaXXspettenAug 26, 2015 2:28 AM
| |
| Interesting idea Haxx, but rather than being an overhaul I'd very much welcome it as yet another category for ranking anime, if we all feel that the MSP score table on the wall is truly redundant in comparison to the score change list then I wouldn't mind us scrapping MSP score from the wall and replace it with this new trending rankings that you're proposing. As much as a good idea that it is i don't think it solves our problem of ties, if you're proposing this to also be used as a potential tie breaker then I don't see why any of the other categories cannot also be used instead. Whether we makeuuse of other categories or time to decide a time breaker I believe either one has aspects or use cases where they'll either be unfair or favor one show rather than another and I don't really think there is truly a perfect system that will be just to all. The trending ranking depends wholly on length if a 1 cour show is trending for 7 weeks and gets a negative the next week then it's impossible for it to make a comeback before it finishes airing. Using any of the other categories as Hibbs has pointed out doesn't allow the show that was on top to defend it's position before losing out and similarly even if we adopted time and the newer show happens to out rank the other show in all other categories then it's also unfair. What I did previously for ties was to identify which of the rankings best depicted How a show was currently doing and back then I deemed the most important rankings to be Score Change --> MSP score --> MAL Score. All we need to do now is revise this order taking into consideration what you've both proposed; time and trending, and find out amongst these 5 categories which should be considered first when wanting to identify the state of an anime at any given time. |
Join the MSP Club for an in-depth look at score progressions and other stats of currently airing anime |
| Well of course that idea wasn't really related to the tiebreaker issue, it was just something I happened to think of since we were talking about the concept of MSP in its entirety. But if you want to make it a 4th list then I guess that works too even more work for us though :s |
| |
| Well yeah that'd take like two seconds per entry to update. A glance is all you need |
| |
| So in a compromise HaXX can bring in his upgraded system of MSP and I can get my rule change on Ties? |
|
| This is a diplomatic negotiation now? :3 |
| |
HaXXspetten said: This is a diplomatic negotiation now? :3 Well you'd be happy. I'd be happy. As such that would be mutually beneficial for everyone in the club. |
|
| Sigh..very well then, all proposals accepted. Since Haxx suggested it you can go ahead and remove the MSP score and add the trending list. Let's see if it's really as diverse as you say, save the MSP just in case. |
Join the MSP Club for an in-depth look at score progressions and other stats of currently airing anime |
| Want me to change it or? |
| |
| Lets hold off on HaXX's idea until it's his turn so he can properly implemented it and we can observe it through his duty. Thus he should do the first week of September. |
|
| Mean swapping mine and Zee's upcoming weeks? Works for me I guess |
| |
More topics from this board
» [Week 07] Fall 2024 Preliminary DataZee530 - Sep 29 |
2 |
by Zee530
»»
Sep 30, 3:51 AM |
|
Sticky: » Spring 2024 Final DataZee530 - Apr 4 |
3 |
by Zee530
»»
Aug 23, 1:50 AM |
|
Sticky: » Summer 2024 Preliminary DataZee530 - Jun 28 |
2 |
by Zee530
»»
Jul 1, 10:09 PM |
|
» Winter 2024 Preliminary DataZee530 - Mar 14 |
0 |
by Zee530
»»
Mar 14, 3:53 AM |
|
» Fall 2023 Preliminary DataXiaraith - Sep 28, 2023 |
3 |
by Johan
»»
Dec 20, 2023 8:08 PM |