Comments and Suggestions

Go read Medalist

Offline

Apr 2007

285

Caitlin8898 said:
I have got one question if you know, though– with the second last diamond, is that including just your active team, or it is the show with the most points (under 85,000) on your team as a whole?

Active team only. For the purpose of checking if an ace is valid, shows that are on the bench or over 85k users are treated as if they don't exist.

FAL Enthusiast :: Reviewer :: Interest Stack Creator

Oct 8, 2022 3:54 AM

#153

Mytrackersea

Offline

Dec 2021

moozooh said:

Active team only. For the purpose of checking if an ace is valid, shows that are on the bench or over 85k users are treated as if they don't exist.

What happens if you ace an anime that surpasses the 85k limit after you do it, does it un-ace itself, do you get no points or does it count as invalid, so you actually lose points (which I think would be terrible, since if you're not available on Sunday and an anime you aced on Sat hits the limit, it seems unfair to punish people for choosing earlier)?

Or do you still get them, so it's worth it to ace an anime that it looks like it might reach the limit before it does so? Are the points cumulative and don't reset weekly, so once an anime hits the limit, it becomes un-aceable for the entire season?

Oct 8, 2022 5:44 AM

#154

Go read Medalist

Offline

Apr 2007

285

Mytrackersea said:
What happens if you ace an anime that surpasses the 85k limit after you do it, does it un-ace itself, do you get no points or does it count as invalid, so you actually lose points (which I think would be terrible, since if you're not available on Sunday and an anime you aced on Sat hits the limit, it seems unfair to punish people for choosing earlier)?

It's all in the flowchart, see special case 1. This is currently the only real source of risk in this mechanic, otherwise it would be completely trivial. But unless you're somewhere with no internet or mobile reception for the entire weekend, it shouldn't ever become a problem; you only need to make a couple clicks to check which show you should ace.

Mytrackersea said:
Are the points cumulative and don't reset weekly, so once an anime hits the limit, it becomes un-aceable for the entire season?

For acing purposes, only the show's weekly points count. Once it has become unaceable (for either hitting the audience cap, or you already acing in one of the previous weeks), it becomes unaceable forever (also in the flowchart).

FAL Enthusiast :: Reviewer :: Interest Stack Creator

Oct 15, 2022 4:15 PM

#155

theslickasian

Offline

Mar 2020

Are the score of anime in mal not a factor for this? As it seems to emphasize more on users watching.

Oct 17, 2022 7:05 AM

#156

IsaBelle33

Offline

Feb 2019

636

hi hi Can anyone give me a rough idea on how score is calculated in this competition. Is it just the average of the 5 selected shows? I'm a total noob when it comes to competitions, this is my first one ༼ つ ◕_◕ ༽つ

My Candies:

“Even when you were so depressed and emotionally unstable, the taste of your coffee jelly stayed the same." Kusuo Saiki

Oct 17, 2022 7:17 AM

#157

Samii

Offline

Jul 2012

1009

IsaBelle33 said:
hi hi Can anyone give me a rough idea on how score is calculated in this competition. Is it just the average of the 5 selected shows? I'm a total noob when it comes to competitions, this is my first one ༼ つ ◕_◕ ༽つ

It's the sum of all 5 in your active team. So for this week, we have watching and score counted. Looking at the Points Criteria in the rules page, we have this:

Users Watching (every week):
• +0.5 points for each user on MAL who has the anime under "Watching" or "Completed" that week. If the anime has not aired yet, these will be assumed 0.
• +0.25 points in Weeks 2, 4, 6, 8, 10, 12.

Score (Weeks 3, 7, 10, 13):
• +20,000*(score - 6.00) points for the anime's score on MAL.
• +40,000*(score - 6.00) points in Week 13.

So, let's make Urusei Yatsura as an example. It has 26,484 users watching currently and it's score 7.46. So you would multiply watching with 0.5 and score would be (score-6.00)*20000 which would mean:

Points = 26484*0.5 + (7.46-6.00)*20000 = 42,442 pts

You do that for all 5 active anime, add them up, and that would be your points for that week.

Oct 18, 2022 1:41 PM

#158

IsaBelle33

Offline

Feb 2019

636

Samii said:

It's the sum of all 5 in your active team. So for this week, we have watching and score counted. Looking at the Points Criteria in the rules page, we have this:

Wow! That was a REALLY good explanation thank you so much! You're an awesome teacher!! 💕

My Candies:

“Even when you were so depressed and emotionally unstable, the taste of your coffee jelly stayed the same." Kusuo Saiki

Oct 18, 2022 1:44 PM

#159

IsaBelle33

Offline

Feb 2019

636

Mahdhi1235 said:
So I aced blue lock on Monday, when it was below the 85000 threshold. Atm, its above the threshold.. Should I keep blue lock aced (will I get the 75k because I aced before) or should I change it?

If it's selected as your current ace and it's past 85,000 members (highlighted in red) you can't ace it and will lose 5,000 points.

My Candies:

“Even when you were so depressed and emotionally unstable, the taste of your coffee jelly stayed the same." Kusuo Saiki

Oct 19, 2022 4:57 AM

#160

Kaylin-chan

Offline

Jan 2021

moozooh said:
Active team only. For the purpose of checking if an ace is valid, shows that are on the bench or over 85k users are treated as if they don't exist.

Thanks again! I just saw you were at the top of the leaderboard– congrats!

Should have un-aced it, oof. ;_;

My MALoween✟Mansion Candies (2024):

x18

Oct 21, 2022 10:28 PM

#161

Chris2theJ

Offline

Nov 2018

Just curious, since this season is the first time I've participated in FAL, will I still be penalized for acing a show if the shows that score above it have already been aced? I know that, at least for this season, any show that exceeds the 85000 user benchmark is no longer ace-able and I believe doesn't count against you if you picked the next highest. Just wondering if that remains true for any aced show on my team that isn't the "highest" scored show, but the show/s above have already been aced? Hope the question makes sense

Oct 22, 2022 12:14 AM

#162

Samii

Offline

Jul 2012

1009

Chris2theJ said:
Just curious, since this season is the first time I've participated in FAL, will I still be penalized for acing a show if the shows that score above it have already been aced? I know that, at least for this season, any show that exceeds the 85000 user benchmark is no longer ace-able and I believe doesn't count against you if you picked the next highest. Just wondering if that remains true for any aced show on my team that isn't the "highest" scored show, but the show/s above have already been aced? Hope the question makes sense

If you already aced the next highest, you need to wait until it goes over 85k to ace the next one or another anime takes it's place.

Questions About Ace
The anime with the highest points in your active team can be declared as an Ace if:
1. you have not already declared it as Ace
2. it has not passed 85,000 users in "Watching" + "Completed"

If you already declared the anime as an Ace and it still has the highest points in your team, you must wait until either:
a. a different anime has the highest points, OR
b. the already declared Ace has more than 85,000 users in "Watching" + "Completed"

You cannot Ace the anime with the second highest points until the one with the highest points has >85,000 users watching+completed.

Oct 22, 2022 10:12 AM

#163

Chris2theJ

Offline

Nov 2018

Samii said:

If you already aced the next highest, you need to wait until it goes over 85k to ace the next one or another anime takes it's place.

Thanks, figured that was the case but best to get conformation

Oct 24, 2022 10:18 PM

#164

Go read Medalist

Offline

Apr 2007

285

Okay, so there's been a Discord discussion about restricted shows because this is the second season in a row where one of the three restricted picks (DAL IV last time, Uzaki-chan S2 this time) is plain not competitive, period. To the point where just having either on active throughout the season would be a clear bad move as they would be outperformed by bench shows most of the time.

At the same time, both seasons have had picks that were so overwhelmingly safe and shoe-in that they should've been the ones restricted instead. I'm talking Aharen-san for spring and Kage no Jitsuryokusha for fall. Even though Aharen was slammed by drops which made it sag in the latter half of the season, it had been the top unrestricted show up until the last week, dunking on DAL throughout and still finishing above it in the very end despite a comparatively insane penalty on week 13. Even in top-20, there were teams without either Paripi Koumei, Summertime Render, Tomodachi Game, or Gaikotsu Kishi, but the first Aharen-less team can only be found at #64. That's just how safe and powerful the pick was, and everyone knew what they were doing when they picked it because numbers did not lie.

So the topic of the discussion is how to make the restricted show selection more competitive between each other; in other words, how to get the likes of Kage/Aharen into that group, and the likes of Uzaki/DAL out of it.

Up until now, the choices have seemingly been decided by PTW numbers prior to the team-picking period. I.e. if it's over 100k, or just a tad under, it's gonna be restricted. The period itself seems to have different duration each time; I distinctly remember it being significantly shorter in spring, but this season it was going on for over two weeks. So, for the sake of consistency, let's assume that from now on it will always start exactly two weeks before the deadline.

The problem with only going with gross PTW is that it inherently penalizes shows announced years in advance, which, as you may have guessed, is exactly what happened in our case: both DAL and Uzaki spent over two years accumulating PTW numbers between their announcement and team-picking. It may not seem like it matters much, but even if a show puts on a meager couple hundred users a week, having 70–80 extra weeks to do so is inevitably going to make numbers add up. On the other hand, the list of shows eligible for the season at all can only be compiled roughly a month before it starts, as many shows cut it close with regards to deciding their season and airing dates.

In order to level the ground before early and late announcements, we may want to look not just at the resulting PTW, but the rate it changes over time. With regards to FAL, since we're limited to a month before the season start, and we need two weeks to pick teams, that would leave us with two weeks (from 4 weeks to 2 weeks before the start of the season) to track changes. Thankfully, that was one of the things I did this time, so let's see what results we would've gotten.

These are ordered by their PTW by the start of the team-picking period, which would be the default order you'd see on the roster page. I've added Bleach and HeroAca S6 (both blacklisted) numbers to the chart for reference because that's relevant.

Right off the bat we see that both Blue Lock and Kage were in one league with the big boys, with Fumetsu S2 trailing around the middle of the next big group of 5–6k, and Uzaki S2 not even managing that much and losing itself among the mid-sized crowd. We also immediately see why picks like Fuufu Ijou had already proved themselves as competitive a couple weeks in despite having relatively modest PTW numbers at the start, and KanColle being... well, being nothing at all so far, but about to be a complete opposite of what Fuufu represents down the line. If you were wondering how experienced players picked the bulk of their teams, this should illustrate how. (I'll also write a more in-depth guide later for a deeper insight into a FAL spreadsheet warrior prep phase.)

This way, by adding just one more data point, we can get a much clearer picture of how all the shows will scale, and hence which shows need restriction—and which don't. Going by the numbers we get by the end of the tracking period (which ends when team-picking starts), we can formulate the inclusion guideline as the following decision matrix.

This would exempt Uzaki and include Kage, as well as keep the current blacklisted shows safely blacklisted.

FAL Enthusiast :: Reviewer :: Interest Stack Creator

Oct 31, 2022 4:36 PM

#165

Go read Medalist

Offline

Apr 2007

285

The current meta made aces extremely powerful, and that puts a massive inherent value on the extra swap wildcard, owing to how significantly it increases the chance of squeezing in another ace. It also makes bombers even worse than they normally are, so at this point they're negative value to the ones who use them, and can only work to grief others. So there's basically one main option and one fallback option—the booster—that's mostly just there to provide an opportunity cost for the other one.

So I decided to brainstorm some wildcard ideas to provide some other competitive options for the current meta and even off-meta playstyles. Numbers are subject to discussion, but I made sure to balance them to be roughly competitive with the extra swap and never be universally profitable (because that's what the booster is for).

Blood Sacrifice
Mark target anime on your active team for sacrifice. Sacrificed anime will receive a fixed bonus of 50,000 points per week instead of its weekly score until the end of the season and cannot be aced or swapped.

Can be used as a safety net option for weak teams or for niche optimization.

Transmogrify
Target anime is replaced by a random non-restricted anime that is currently not on your team. The anime and its weekly statistics will not be revealed until deadline.

A great (horrible) way to replace a useless stinker on your team with a potentially worse stinker without knowing what it will be until it's too late. But hey, you have a ~1/35 chance of hitting that missing hidden gem!

All for One
At the penalty of –35,000 points, declare an anime that you haven't aced before as your ace for the week. This will be considered a valid ace regardless of the anime's weekly score and watching/completed numbers. You cannot ace anime until the end of the season.

An option to force a half-ace on any show, including ones that couldn't be aced due to growing too quickly, at the cost of permanently disabling further aces.

Niche Audience
At the penalty of –20,000 points, lower target anime's ace eligibility threshold from 85,000 to 50,000 viewers until the end of the season.

A niche optimization that forces the cap on shows that grow slowly but still easily block others on most weeks at the end of the season, allowing a player to squeeze in an extra ace or two at a hefty penalty. Note that if the target show is already at >50k by the end of the week you're using this and you haven't aced it earlier, that's also one less ace for you.

Polarity Change
Target anime will have its weekly score turned positive if it is currently negative, and vice versa, capped at 80,000 and –80,000 points respectively.

This one's fun! Basically, it's something to play around with if you have a very bad show on your team (or, like @MeowPoi_, would rather prefer a show to do worse instead). 80k seems like a lot at first, but you need really good team planning to get that much out of it and not run out of swaps, and the effect only lasts one week.

Good Teamwork
On the final week, every anime that hasn't been aced receives an extra 40% bonus to its weekly score. Receive a penalty of –25,000 points for every anime that has been aced, including those on your bench.

This may look simple on paper, but in order to make the maximum of it, you need to know your week 13 composition well in advance to avoid acing corresponding shows. The more you ace, the worse it gets, but you still want to ace your worst performers somehow to be competitive with conventional teams. It's also great for those who don't want to bother with acing at all.

Swap Hoarder
Swaps are disabled until the end of the season. Receive 25,000 per every unused swap you had at the start of the week, and an extra 50,000 if no swaps have been used since the start of the season.

Another off-meta option for players who would rather go with their best picks from the start and use swaps as little as possible (or not at all).

Booster mk. II
Receive 30,000 points if used on week 10, 25,000 if used on week 11, 20,000 if used on week 12, and 15,000 if used on week 13.

Replacement for the current booster that makes it more interesting and competitive.

Edit: Some clarification.

moozoohNov 1, 2022 9:21 AM

FAL Enthusiast :: Reviewer :: Interest Stack Creator

Nov 1, 2022 1:46 PM

#166

Offline

Jul 2012

Hmm, well I think something to improve mid game slump would be nice. People make few swaps and aces early, but after that they are mostly forced to wait long until shows reach 85k threshold and save valuable swaps late for maximal gain.

Having few more wild cards choices and given 2 of them would help this. If wildcards were usable mid game then it should not require perfect forethought. That would be too much headache for players to plan around. Something to aid player in present without having to worry too much about past decisions.

There could be 2 sets of wild cards for mid game (week 6-7-8?) and late game (week 10-13).
New acing change gives wildcard swap lot more value than previously, so points suggested here are bit all over the place.

Hat Trick
+40,000 pts if you aced 3? shows in row in previous weeks
(Might make players more aggressive in early game and take risk with swapping and acing)

AFK
+30,000 pts, you cannot make changes to team for 3 weeks (acing and swapping is disabled)
(For lazy players or if you are confident about predicting future)

Odd Duck
+30,000 if you have unique team
(You could choose unconventional pick for your bench in order to guarantee it but you would be at disadvantage with wasted team slot. Would be also nice boost to casual players with more varied teams, or players who take risks when predicting dark horse of season)

Lone Wolf
+15,000 if you are not tied with anyone in ranked leader board

Imposter
temporarily change anime in your team into any non limited show for 1 week [cannot be aced]
(Didnt choose hidden gem? well you can have it for 1 week. Probably too wild and annoying to code. But there are endless possibilities)

Didnt pick really popular show? you can change your lowest point show into huge +85k watching show to help with acing while gaining lot of points
Use it on show that has high points but prevents you from acing (this would be stronger than wasting swap)

Misfit
ace show with negative points
(helps players who are stuck with horrible show in their team)

Tiny Titan
choose show below 60,000 watching which will receive *1.5 points for this week
(this would be mid game wildcard, if used on w7 would net about +30,000 pts, having this on w13 would be too silly)

Audience Boost
temporary +20,000 watching for 1 week (does not give points for extra watching)
(helps when low watching show is blocking ace, similar to moozooh idea "Niche Audience" in above post but this would be temporary change and would be available as early game wildcard choice)

feedback on @moozooh wildcard ideas
Niche Audience is really good idea and would help worse performing teams to finish their aces in final weeks but this could be possibly very powerful when you have high scoring show with low watching
Transmogrify lot of people have suggested something similar but rng mechanics are not really fun. Which is why I suggested my version of this with 'Imposter' being only temporary for 1 week. Maybe someone will come up with even better way to do this.
All for One is pretty cool as you can use it on Blue Lock in final week without having downside.
Booster mk. II is simple and easy to understand, makes booster more dynamic.
Polarity Change is pretty fun and useful for teams which aim for lowest points but little gimmicky

NMx13Jun 13, 2023 1:00 AM

Nov 1, 2022 2:30 PM

#167

Go read Medalist

Offline

Apr 2007

285

(Funnily enough, I'd be one of the few who wouldn't be able to use All for One on Blue Lock since I'd already aced it, but that just means I could use something else!)

We also had some discussion on Discord prior to NMx13's post, and at some point people started expressing wishes on making wildcards available earlier and/or making more than one available per season.

While I think this should be treated with great care, as some of the powerful ideas suggested above can compound their power above their respective individual levels when combined, one potential way to address this would be to separate the total pool of wildcards by weeks when they are exclusively available.

Say, for instance, players have up to 3 wildcard uses per season, and there are 10 different wildcards in total which have different themes and functions to them.

Wildcards A, B, C require the player to commit to a particular strategy or playstyle, and are only available weeks 1 to 4.

Wildcards D, E, F would be those that help or reward mid-season acing or something else like that, available weeks 5 to 8.

Wildcards G, H, I provide powerful late-game effects such as swaps and similar functionality, available weeks 9 to 13.

Wildcard J is some basic one-time effect that is available on any week (but probably not the booster because it's more interesting to give it a time incentive as I suggested earlier).

Whenever a player chooses a wildcard, it is removed from the option pool until the end of the season, so you can't stack the same one three times. You can potentially allow or disallow stacking wildcards within the monthly sets depending on how well they synergize.

moozoohNov 1, 2022 3:39 PM

FAL Enthusiast :: Reviewer :: Interest Stack Creator

Nov 2, 2022 10:19 AM

#168

turtleinshell

Offline

May 2013

318

Highly recommend nerfing forums down heavily after this season. It feels very busted right now. Score/quality should be the most impactful factor. It's cool for forum buzz to count more, but 2.5x the value of last season feels very busted. Score should be buffed up closer to what it was before too. Just my 2 cents.

Nov 3, 2022 10:32 AM

#169

Offline

Jul 2012

To combat abuse from forum post spamming for easy FAL points, from FAL players but also some freak accidents like RobertBobert, where he made 100 forum posts alone for episode 1 of gundam and 60 for episode 2. Now RobertBobert is not participating in FAL as player but this is very much troubling precedent. There was also user who made bunch of bot accounts called Maltest01, testmal02 and so on to spam forums with gibberish but those posts were clearly deemed as spam and forum moderator removed them all.

With that said I suggest to use count of users who voted in poll.
There are 10 times more users voting than posting. So instead of current 100 points we can divide it by 10. 100/10 = 10 pts, which is lot more in line with other point metrics such as favorite being 15 pts.
10 points might still be too much if we remember that forum posts used to be 40 pts before this season and from looks of it they make more than estimated goal of 15% distribution coming from forum post will be way over.
We can see in picture below that first episode gathers more discussion but afterwards ratio between posters and voters is relatively close.

Stats as of writing (earlier episodes have had more time to accumulate more votes and posts)

Other solution would be to only count unique posters in thread, but this would not solve problem of forum posts giving disproportionate power to users to skew results. See below for image example of close race between Tensei and Fuufu. Also Bocchi vs Gundam was very close until end. Few bad apples could screw over ton of people when trying to ace.
Solution would also need easy way to view in advance how many unique posters there are in each thread. This would be good thing in general for users who do not excel sheets to calculate forum posts but would require extra effort to implement into FAL.

Just before deadline

After deadline

Nov 4, 2022 7:39 AM

#170

Offline

Feb 2011

While the above discussion is ongoing - can I kindly ask you guys what is the exact method by which we get points for episode discussion? And I mean exact, out of all the scoring criteria this is the only one where I fail to replicate results from the official FAL page. I know that we look at two last weeks for episode discussions but what do we count as two last weeks (ie. do they start Monday two weeks prior?), do we count topics for all episodes (regardless when they aired) etc.?

For the sake of argument, let's take a simple example of Punirunes from last week's scoring:

When we consider week 4 was timed from Monday Oct 24th to Sunday Oct 30th, then apparently there should only be 1 post in episode discussions of that series. That is indeed true, if we consider week 4 only. There has been another post on October 18th (one week prior) in the episode 1 thread and 1 more in episode 2 thread on October 10th (two weeks prior). Can anyone help me understand this?

Also, on a completely different note - why was Arknights: Renmei Zensou not considered for the initial listing? From it's anime page it seems to be a TV series so not an OVA, ONA, Movie or Special. Just trying to figure out why we have KanColle but no Arknights :D

Nov 4, 2022 11:26 AM

#171

Offline

Jul 2012

@Grimiku New posts in all episode threads made in previous two weeks. Punirunes has currently 3 episode threads but main/opening post is not counted as you can see in picture below (ep 2 and 3 says 0 posts). In your imagine it says total of 400pts from forum posts which matches to 4 posts in episode 1. As example, if someone makes new post to ep 1 before deadline of week 6 then punirunes will gain 100pts for 1 new post made in last 2 weeks.

Arknights was announced after FAL registration opened so it could not be included.

Nov 4, 2022 11:53 AM

#172

Offline

Feb 2011

Thanks so much for the info! Seems a bit odd that opening posts are discounted but it makes sense from what you're saying. Too bad for Arknights as well.

Just wanted to drop it here - I've been on MAL on-and-off for over 10 years but it's my first FAL and so far having much more fun than I thought! Big props to everyone involved in organizing ^^

Nov 4, 2022 2:21 PM

#173

Go read Medalist

Offline

Apr 2007

285

NMx13 said:
With that said I suggest to use count of users who voted in poll.

On the one hand, if the individual users' contribution is appropriately small, there will be less incentive to abuse this system. On the other, the abuse itself, if it happens, will be essentially invisible to other forum users, so somebody very determined on creating, say, 200 fake MAL accounts using temporary mailboxes and random name generators to prevent surface-level detection could easily put in thousands of points for their anime of choice with everyone else being none the wiser. It would also lose the main distinguishing feature forum discussion bonus has at the moment: additionally scaling off of the show's content as opposed to mostly the number of people watching.

That being said, I do agree that the current abuse must be dealt with by the next FAL season at any cost. Not only it puts too much power in the hands of rogue individuals, it also enables easy griefing as you've demonstrated with week 4 Fuufu/Tensei situation. I think any solution that relies on an absolute count of anything users are permitted to do is doomed to be abusable by those who can just spend time to do more of it.

The hypothetical optimal solution both solve both that issue and avoid inherently rewarding more popular shows. The current implementation of the rating score formula (used for weeks 3, 7, 10, and 13) is great for this because it's a direct reflection of what the majority of active audience thinks about the show in question and is also very nearly impervious to any abuse by FAL players. Perhaps there's a way to make the discussion score formula similar.

FAL Enthusiast :: Reviewer :: Interest Stack Creator

Nov 4, 2022 5:08 PM

#174

Offline

Jul 2012

moozooh said:
On the other, the abuse itself, if it happens, will be essentially invisible to other forum users, so somebody very determined on creating, say, 200 fake MAL accounts using temporary mailboxes and random name generators to prevent surface-level detection could easily put in thousands of points for their anime of choice with everyone else being none the wiser.

Good discussion on forums. Just to understand your position on matter, if this was just something you wanted to highlight in forums or is there deeper reason why you think using number of votes in polls would be bad? It is true we cannot verify if votes in polls are made by bot accounts but there is nothing to stop people from making duplicate accounts to make legitimate looking forum posts. Writing 'legimate ish' forums posts does take more effort but it is worth effort due to high point value given in return (even if points were reverted to 40). I'm not 100% sure polls scale only by audience size, there does look to be pretty good correlation between forum post activity and polls. Gundam would definitely take hit but if we have to remember that RobertBobert made 100 of those 577 posts in ep 1 thread. Gundam poll activity matches Fuufu which is much bigger show and Bocchi (with little more audience than Gundam) out performs Fuufu. I think this should be something to look into further with more shows and later weeks.

Other metrics like drops and favorites are also hidden and it might be hard to detect if there is foul play. Drops and favorites do have linear scaling so something could be detected if it does not hold up but would be hard if foul play is spread out among weeks. Something that has been suggested for other metrics to help less popular but good shows is idea of percentage scaling. Not sure if this applicable to polls or posts but this would help out even playing field and giving proper attention to shows which have large amount of discussion in relation to size. As we would want to avoid this being only popularity race.

Lot of data and more thoughts in detail inside Google Sheet I made. You can also copy it and play around values.
Here is also picture of sheet but reading directly from sheet is easier

NMx13Nov 4, 2022 5:16 PM

Nov 4, 2022 7:28 PM

#175

Go read Medalist

Offline

Apr 2007

285

NMx13 said:
Gundam poll activity matches Fuufu which is much bigger show and Bocchi (with little more audience than Gundam) out performs Fuufu. I think this should be something to look into further with more shows and later weeks.

Here:

These are the values for the first four episodes of the shows in your table above. Kage is taken as 100% for relative calculations. Since we the data for vote tallies for their respective FAL time periods is forever lost, I applied a correction assuming the same difference as between FAL-eligible post counts for the first four weeks (taken from the leaderboard) and total post counts.

Moving on from posts to votes removes a huge chunk of relative value from Gundam, the smallest show, and only barely helps Bocchi, which doesn't even need the help since it's about to catch up to Fuufu in audience size by the end of the season, anyway.

FAL Enthusiast :: Reviewer :: Interest Stack Creator

Mar 6, 2023 5:50 AM

#176

Konan_Edogawa_

Offline

Jan 2018

554

Can we make discussion posts not be worth 100 points. I feel like a reasonable number would be 50 or even 40 like in the spring 2022 season. This will cut down on the amount of people spam posting in the last hour or so of the FAL week.

Apr 3, 2023 11:23 AM

#177

Looping_Fox

Offline

Mar 2023

can we know which anime is having the largest number of watchers this season. umm this is my first time participating in FAL.

Apr 3, 2023 12:18 PM

#178

Go read Medalist

Offline

Apr 2007

285

Looping_Fox said:
can we know which anime is having the largest number of watchers this season. umm this is my first time participating in FAL.

I believe some of the community members are working on a public FAL tracker like the one we used in the previous season, but it's not available yet.

Other than that, if you go to the season page and sort the titles by total members (which is the default sorting), you'll eventually be able to see which anime has the largest number of watchers. Right now, since most of the shows haven't started airing, it mainly reflects the number of plan-to-watch listings, but after a couple weeks the ordering should more or less align with the real number of watchers.

FAL Enthusiast :: Reviewer :: Interest Stack Creator

Apr 10, 2023 2:50 AM

#179

Offline

Feb 2011

Hey folks, wanted to clarify episode discussions with you again this season.

Any reason discussion points for early starter series like Yamada and Tengoku which managed to air 2 episodes already in past two weeks have less points than they should?

Example:

Yamada's ep 1 discussion as of now: 233 posts
Yamada's ep 2 discussion as of now: 124 posts
Total: 357 posts
Yamada's indicative number of replies counted in FAL as of now: 223 posts

Even if we count in the rule that states "Episode discussion forum posts that were made before the FAL season starts will not count" (discounting 94 replies made on April 1st before season started) that still leaves us with quite a gap.

Would appreciate if someone could clarify, thanks guys.

Apr 10, 2023 5:16 AM

#180

Samii

Offline

Jul 2012

1009

Grimiku said:
Hey folks, wanted to clarify episode discussions with you again this season.

Any reason discussion points for early starter series like Yamada and Tengoku which managed to air 2 episodes already in past two weeks have less points than they should?

Example:

Yamada's ep 1 discussion as of now: 233 posts
Yamada's ep 2 discussion as of now: 124 posts
Total: 357 posts
Yamada's indicative number of replies counted in FAL as of now: 223 posts

Yamada actually had 136 posts on season start, It's less now probably because forum mods deleted some of them that were against guidelines, but they still count. I think they shouldn't but FAL system doesn't seem to account for deleted posts.

Here are all discussion posts made before FAL started:

Apr 10, 2023 8:13 AM

#181

Offline

Feb 2011

That's the answer I was looking for, thanks so much Samii!

The table is also sthg I didn't know I needed but I do now ^^

Apr 16, 2023 6:41 AM

#182

cruelpyro

Offline

May 2014

I honestly think that the episode discussion point system should be rethinked or even scrapped completely.

I looked over the Oshi no Ko episode discussion list and in 1 hour i found more than a 100 comments from fake accounts. By my estimates about 30%-40% of the comments on Oshi no Ko at the moment are from fake accounts, inflating the number of points received immensly.

This made me think that this category could easily be influenced by botting taking away from the sportmanship of the whole thing.
Simply imagine i prepare 1000 accounts in advance and i spam the ep discussion few hours before we receive the points. That means that whoever does not have the anime that i spamed in their team loses 75k points/anime spammed. This could easily add up to more than a million points across the whole FAL season.

I feel like the best way to evade this scenario is to completely scratch pointing ep discussions. Or to rectify the points after the Admins had time to clean up the forums.

Update: Just for reference after the points were distributed around 800-900 comments were deleted from Oshi no Ko's ep discussion, this means that everyone that had Oshi no Ko got 60k-70k points from bots.

cruelpyroApr 17, 2023 12:50 AM

Apr 19, 2023 1:57 PM

#183

FuckJJK

Offline

Feb 2017

cruelpyro said:
I honestly think that the episode discussion point system should be rethinked or even scrapped completely.

I looked over the Oshi no Ko episode discussion list and in 1 hour i found more than a 100 comments from fake accounts. By my estimates about 30%-40% of the comments on Oshi no Ko at the moment are from fake accounts, inflating the number of points received immensly.

This made me think that this category could easily be influenced by botting taking away from the sportmanship of the whole thing.
Simply imagine i prepare 1000 accounts in advance and i spam the ep discussion few hours before we receive the points. That means that whoever does not have the anime that i spamed in their team loses 75k points/anime spammed. This could easily add up to more than a million points across the whole FAL season.

I feel like the best way to evade this scenario is to completely scratch pointing ep discussions. Or to rectify the points after the Admins had time to clean up the forums.

Update: Just for reference after the points were distributed around 800-900 comments were deleted from Oshi no Ko's ep discussion, this means that everyone that had Oshi no Ko got 60k-70k points from bots.

Does that mean that anyone having oshi no ko on the main roster will lose points on the next week if it doesn't surpass its previous peak?

Apr 19, 2023 3:53 PM

#184

cruelpyro

Offline

May 2014

As far as i know, no, they will not lose points.

Apr 20, 2023 3:22 AM

#185

Go read Medalist

Offline

Apr 2007

285

The scoring system doesn't remember past values. It's only concerned with values it can see at the very end of its respective week.

The only time past values are used are when weekly results are added to players' and titles' totals for the leaderboard placements.

FAL Enthusiast :: Reviewer :: Interest Stack Creator

Apr 24, 2023 3:51 AM

#186

Ocleal

Offline

Dec 2016

I've joined FAL this year and am thoroughly enjoying it so far, reading Moozooh's guide, looking at spreadsheets, etc.

However I can't help but feel that the point distribution is just off. I understand that "Oshi no Ko" is doing great, and thus it should grant the most points, but it becomes a lot less fun when the main determining factor in my rank is not whether my team as a whole is doing well, but rather whether I picked this one anime.

My suggestion would be to keep the point scoring the same for the anime rankings, but to change the way the players get points:
Assuming there are 50 anime in the season, the #1 anime would grant 50 points, #2 would grant 49, #3 would grant 48, etc. Aces would grant something like 25 points (whatever number maintains the current balance). That way the best anime still grant the most points, but they don't render other picks entirely meaningless.

It makes sense to me that someone who picked the #2,3,4,5,6 best anime should get more points than someone who picked #1,8,9,10,11

In the particular example I gave, the system would make it so that there are a lot more people with the exact same scores, so it needs work, but the general idea is there.

What do you guys think? Would you prefer a more balanced point system? Or am I just being salty because I didn't pick "Oshi no Ko"?

Apr 24, 2023 10:00 AM

#187

Go read Medalist

Offline

Apr 2007

285

The thing with Oshi no Ko is that nobody really expected it to blow up this hard. Jigokuraku ended up blacklisted, making the choice between restricted picks trivial for experienced players, but prior to that the Discord regulars were debating whether it was a better pick than Oshi if both were available as restricted picks. I'd been in the Oshi camp from the beginning, but even I was hesitating, knowing that Jigo would start much earlier and be easier to get into (given that it's a well-produced shounen with many endorsements and no ambiguity to its genre or premise). Myself and others who were in the Oshi camp all thought that if (if, not when) Oshi would come out ahead, it wouldn't be by very much, and the gain would likely only materialize late in the season. It was considered the riskier of the two options.

Oshi blew past Jigo's totals in less than three weeks (while being idle half the time due to late start), and it came as a shock to all of us. This is not how restricted picks are supposed to behave. This is by all means a performance you'd see from a blacklisted title, and should give you a taste of what it'd feel like if no titles were blacklisted. There was no indication that something like this would happen. This is a mistake on the devs' part, but there was nothing they could do to avoid it. It's just what it is. This season sucks in general because there are five shows that are clearly better than the rest, so every experienced player picked almost the same team, and they've all capped in week 2 so now all we have for the rest of the season is finding a way to squeeze everything out of our bench. Those who had Cheat Skill don't even have to think about it since the smartest thing to do was to rotate the bench during weeks 2 to 4 and afterwards just check in once a week to see if any of the bench shows somehow popped off and eclipsed one of the big five.

Which brings us to your suggestion: I'm not sure it would achieve the results you're hoping for. If you look at the current leaderboard, the top five shows have a substantial lead over the rest, meaning they would always give the same maximum amount of points every week, and there are well over a hundred players who have all of them. Without Oshi but having #2 to #6 best picks (i.e. Tengoku, Mashle, Konosuba, Yamada, Cheat Skill), you'd still be somewhere in the middle of three-digit placement at best because of how many users would've still had a team superior to yours: 1-2-3-4-5, 1-3-4-5-6, 1-2-4-5-6, etc., would all be higher than 2-3-4-5-6. You would've been able to pass players with poor teams that happened to have Oshi by luck, but it wouldn't let you beat players with strong teams who had picked them knowing how to pick a strong team.

More importantly, it would've also reduced both strategic and tactical depth of the game. Now, instead of potentially picking titles by different roles they can fulfill, you'd always pick all-rounders that'd maintain a high overall rank and have totally risk-free bench rotation without worrying when to do it. In the end, you'd have hundreds of players tied in clusters separated by 1-2 points. Sure, it'd potentially reduce the cost of catastrophic team-picking mistakes, but in return it'd drop the skill ceiling for everyone and remove any avenues for high-risk/high-reward play. It'd make the game more awkward and eventless than it already is. I'm not sure if that is what you want.

I do, however, think that the scoring system needs a better way to handle obvious outliers and possibly provide some sort of catch-up mechanics for players who have made wrong picks. Just not at this cost!

moozoohApr 24, 2023 10:08 AM

FAL Enthusiast :: Reviewer :: Interest Stack Creator

Apr 24, 2023 2:22 PM

#188

Ocleal

Offline

Dec 2016

An interesting analysis, but I think you might be thinking too much about the top 1% of players. I'm sitting 4418 (down from a 1608 peak on week 1), and expect to drop down to about 7000 as literally everyone with "Oshi no Ko" overtakes me. It might be competitive at the top level, but in the middle where I and most people are, no anime pick, swap, ace or other decision has even close to as much impact as having picked this one anime. All strategy is reduced to "did you pick that one anime?" I almost wish my team was doing worse so all the "Oshi no Ko" people would overtake me faster.

While I'm not sure I agree that the strategic and tactical depth of the game would be reduced, I'm at least glad to hear that you agree that we need a better way to handle outliers. Since that's exactly what I would like, regarless of how it's achieved. Another option to at least mitigate it would be to award points non-linearly, like taking the square root of the anime's score, rather than the score itself. This would impact outliers significantly, while having less impact on anime with similar scores. The disadvantage is that it may be a bit more confusing, especially for casual players.

Thinking about it a bit more after reading your post, I realized that perhaps a bigger issue than the outliers is the predictability of the scores from week to week. This is because, with the exception of discussions, all the scores are cumulative. The points accrued in past weeks will also be awarded in future weeks. Wouldn't it be more interesting if only new viewers/ratings/favorites/drops for only that week were considered? This would lead to a more dynamic anime ranking leaderboard where anime would swap places more often. There might be issues with that specific implementation too, but the general idea is that the anime rankings should be more dynamic based on the anime's performance that week, rather than static as they currently are.

Catch-up mechanics could be nice, though just having a few more actions to take would be enough for me, ideally I'd like to have at least 1 action to take per week. This would give the me something to think about on calmer weeks where nothing really happens.

Apr 25, 2023 12:19 AM

#189

Go read Medalist

Offline

Apr 2007

285

Ocleal said:
An interesting analysis, but I think you might be thinking too much about the top 1% of players. I'm sitting 4418 (down from a 1608 peak on week 1), and expect to drop down to about 7000 as literally everyone with "Oshi no Ko" overtakes me.

Balancing around strong players isn't a bad thing in principle. If you care about the game, a strong player is what you aspire to be, and having a high skill ceiling is what keeps you in the game at that point, because you surely wouldn't spend dozens if not hundreds of hours twice a year to do all the prep work described in the guide otherwise.

Your first attempt is pretty much guaranteed to be a failure. You shouldn't be too discouraged; use it as a practice season to try out a risky strategy and see what the top players are doing to maintain their lead or try to pass each other.

Ocleal said:
It might be competitive at the top level, but in the middle where I and most people are, no anime pick, swap, ace or other decision has even close to as much impact as having picked this one anime. All strategy is reduced to "did you pick that one anime?"

That's how it is with restricted picks most of the time, unfortunately, because of their inherently high scaling potential. There's always the right pick. I do think that not having Oshi in the pool would've made this season a lot more interesting, but that's a thought made in hindsight, and then the remaining "right" pick would've been Konosuba because the other two don't stand a chance, as you can tell from the leaderboards. If you want to compete, there's no getting around picking the strongest anime, and if you're competing for one of the prize spots, you'll want at least the top four (sans mutually exclusive restricted picks, of course). This would still be true even under your proposed scoring system mind you. But, again, that's not necessarily a bad thing because picking the right team is also a skill you can and have to get better at.

Ocleal said:
Another option to at least mitigate it would be to award points non-linearly, like taking the square root of the anime's score, rather than the score itself. This would impact outliers significantly, while having less impact on anime with similar scores. The disadvantage is that it may be a bit more confusing, especially for casual players.

That is definitely something to consider, but rather than for score, I'd apply that to watching/completed members and perhaps favorites, since these tend to have the most runaway scaling and are unlimited at the top end. Discussion points just need to be reworked entirely; there's just too much stress involved for everyone looking for spam and boosting posts.

Ocleal said:
Thinking about it a bit more after reading your post, I realized that perhaps a bigger issue than the outliers is the predictability of the scores from week to week. This is because, with the exception of discussions, all the scores are cumulative. The points accrued in past weeks will also be awarded in future weeks. Wouldn't it be more interesting if only new viewers/ratings/favorites/drops for only that week were considered? This would lead to a more dynamic anime ranking leaderboard where anime would swap places more often.

Scores aren't cumulative; metrics are, and this is an important distinction because the game (or the site itself) would first need to have a way to somehow record and meaningfully present the difference between past week metrics and current metrics—and this is already a big problem with tracking discussions. As things stand now, unless you've recorded the previous week's state yourself or know where to look for this data (which nobody sans the top 1–2% know or care to do), you're going into each week completely blind. On the other hand, this would give a lot more power to spreadsheet warriors, since all they'd have to do to see future weeks' results would be a simple subtraction from their predicted values. I can tell you in advance that this wouldn't hinder me even one bit, for example, but somebody who isn't so keen on data forecasting would be utterly annihilated, lol.

But you know what? That'd be an interesting experiment to go for in an unofficial FAL that the Discord regulars conduct during off-seasons. Coming to think of it, we certainly haven't been using it to experiment with new features and gameplay changes as much as we could have. Drop by if you want to discuss and/or participate (to get to the #fal channel, visit #bot-commands to acquire the necessary role).

Ocleal said:
Catch-up mechanics could be nice, though just having a few more actions to take would be enough for me, ideally I'd like to have at least 1 action to take per week. This would give the me something to think about on calmer weeks where nothing really happens.

There's been a call for more wildcards, as well as many interesting suggestions made on the previous few pages of this thread. The main problem is they require a lot of work to code and properly balance against each other. That's pretty much the bottleneck right now. Having more wildcards and having them become available earlier would open a ton of interesting options for everyone. Maybe if we bug the devs insistently enough, they'll cave in and do it. :D

FAL Enthusiast :: Reviewer :: Interest Stack Creator

Apr 25, 2023 1:27 AM

#190

trampler8324

Offline

Jan 2012

204

Ocleal said:
It makes sense to me that someone who picked the #2,3,4,5,6 best anime should get more points than someone who picked #1,8,9,10,11

Well, don't want to disappoint but actually those who picked 2,3,4,5,6 have more points than 1,7,8,9,10. Sure, sometimes high-rated shows kill a lot fun, but that is not a main problem with FAL. For the longest time the main problem is the importance of "currently watching". They should really buff the amount of points for rating/favorites/posts or smth. While I don't have the right answer how to make it better, it is pretty obvious to me, that unless you just pick shows you like, it's really easy to be in top 300/500 on your first week.

Jun 12, 2023 8:20 AM

#191

Martinch

Offline

Jul 2014

I was going to write this post after the season ended, but given the sudden announcement of Summer FAL, I had to act quickly, so excuse my English if something is unclear or grammatically incorrect.

To give my feedback more credibility, I finished 1st in Fall 2022 and I'm currently looking to finish 3rd this season. While I think FAL is a very fun game and has the potential for some really cool moments, there is one aspect of it that completely ruins the experience for me and many other players - discussions. The way discussions currently work in FAL is bad for several reasons, so I will try to break it down to demonstrate that as best as I can.

In the current rules, each comment on the forums is equivalent to 75 points. These points are calculated on even weeks, where users watching bring you 0.25 points per user. This means that during each even week 1000 FAL points are equivalent to 4000 people watching an anime or... 14 comments. Simply put, that's ridiculous.

Because comments are so overpowered, each season we see people go on the forums and "engage" in the discussion just to boost the points an anime will get. They do this to influence aces or make up for it performing worse than they expected. Just this season alone I've had to report like 50 blatantly obvious comments by alts of FAL players whose only purpose was to rig the results:
https://files.catbox.moe/dmc09z.png

While over 13 weeks the total damage these spammers can inflict could be negligible, because FAL is a game where the results each week matter, as such it should be looked at from that perspective.

To illustrate my point, let's take a look at Yamada's and Konosuba's points this season. On week 4 the difference between Konosuba and Yamada was 242 points:
https://files.catbox.moe/iydora.png

This is equivalent to just 4 comments. Or, to just put it in perspective again, 1000 people watching. We already saw this problem happen multiple times in Fall 2022, where players engaged in forum spam wars, where 1 side was trying to boost an anime to ace it, while the other side was boosting something else just to stop the first side from getting their ace. The mods promised they would keep a closer look on the discussions this season and while they've done a tremendous job and I'm really grateful for their work, there are still cases that slip their sight and frankly they shouldn't have to bother with these types of problems. To make things even worse, while the stats show that Yamada had more points that week, I'm still not sure if that's accurate, due to one idiot who decided to spam as many comments as they could just before the deadline. Here's a screenshot from that forum thread 30 seconds before the 23rd hour after which the stats are frozen:
https://files.catbox.moe/rkojpp.png

Because these comments were removed a few minutes after the stats were frozen, I'm still not sure if they didn't have an effect on the week 4 leaderboard. And even if these points were removed from w6 points to balance that, it would be pointless, because the damage would have already been done.

Looking at it as a total amount, the difference between the two was even more of a key factor. In week 2, some of the teams had to make a tough choice - bench 1 anime between Mashle, Yamada and Konosuba for 3 weeks - meaning you had to predict which of these would have the least total points throughout week 2 to 4 combined. In the end the total week 2-4 difference between Yamada and Konosuba was 217 points. That's again equivalent to 4 comments - something that nobody would be able to predict and would be up to absolute chance. While there is always a gamble element to fantasy league games, this is not something any player could or should even remotely predict.

So how can we fix this?

Well, there are several ways this could be done.

Nerf discussions - given that 1 drop is equivalent to -4 points and 1 favourite is equivalent to 15, discussions should be equivalent to a number in that range. I'd say 20 would already be a big improvement.
Stop discussion boosting - to achieve this, the only comments that should be counted are from people who have at least 25 other comments and an account age of at least 6 months. Also, if the account is participating in FAL, their comments shouldn't count towards the results. Also it would be a good idea to count each commenter only once - 2 people posting 20 comments replying to each other is not representative of the overall audience engagement.
Don't count discussions altogether - instead of counting discussion, why don't you count poll results? Given that MAL automatically asks you to rate the episode you watched every time you update, the total number of votes is a good indicator of engagement and is also something that isn't easily manipulated.

Personally I'd go with 3, but if you insist on having discussions, 1+2 (just 1 isn't going to fix the issue) is a good alternative. If you go that route, please only count the comments on the last 2 episodes - why are comments that were posted in the last week in Episode's 1 discussion thread counted towards the points for w10? Also, count double the discussion points for both w12 and w13 (why should anime which have 10/11 episodes or start earlier get punished by missing the double discussion points on w13?)

I hope this explains mine and many other top players' frustration with the way discussions currently work. I hope you take measurements against it as soon as possible.

Thank you for reading my feedback and for organising FAL in the first place. I know this is just a side project MAL is running, but you have a passionate community, so we want to make it even better and more enjoyable for everyone involved!

- Mart

PS:

Bombing is the worst mechanic in the game, both from a gameplay and out of gameplay perspective, please remove it
We need more wildcards that are available since early on - it would both increase the skill cap and we would have less situations where people don't have anything to do for 2 months
Why do we need to manually select what we want to ace, this should be automated... (the scenario where you purposefully don't want to ace an anime that's eligible is 1 in a billion and is related to bombing anyway)
Boost back anime score, we're predicting the best anime of the season after all. Also increase the threshold from 6.00 to 7.00

MartinchJun 12, 2023 8:59 AM

Jun 12, 2023 6:52 PM

#192

Rumpel1408

Offline

Jan 2021

Personally, I think having more FAL season will result in fatigue, but I think you polled on this last season so I'm probably just in the minority

But, I think it's an bad idea to get rid of the limited selection brackets

Jun 18, 2023 2:53 AM

#193

Infernalization

Offline

Dec 2020

627

This season's FAL shouldn't have included Oshi no Ko. Just look at the point differences Oshi no Ko has in comparison to Heavenly Delusion, the #2 of the individual anime point list. It almost triples the points.

Jun 24, 2023 6:08 AM

#194

MonstroGH

Offline

Jun 2022

952

My sugestion.
Fantasy Anime League in all seasons.

Jul 2, 2023 4:36 PM

#195

J3_Omega

Offline

Sep 2016

483

Infernalization said:
This season's FAL shouldn't have included Oshi no Ko. Just look at the point differences Oshi no Ko has in comparison to Heavenly Delusion, the #2 of the individual anime point list. It almost triples the points.

Wouldn't have made any sense to do that. At the start of the season, there was no indication that Oshi no Ko would blow up like it did, so there'd be no reason for them to not include it at the start of the season. It'd be like saying Bocchi The Rock shouldn't have been included last Fall because of how good it ended up doing, even though at the start of the season it was under everyone's radar and barely discussed. Some anime just explode out of nowhere and sweep the season, and they should definitely not be excluded for doing as well as they do.

Jul 3, 2023 12:26 AM

#196

nick_el_son

Offline

Jul 2011

336

Hey guys, is it possible to somehow filter the results and see how will be the rankings if we exclude all participants with "oshi no ko" in their lists? :D

Jul 14, 2023 10:12 AM

#197

Konan_Edogawa_

Offline

Jan 2018

554

nick_el_son said:
Hey guys, is it possible to somehow filter the results and see how will be the rankings if we exclude all participants with "oshi no ko" in their lists? :D

No because that would change when anime should be moved from the active to the bench and would affect aces of said shows. Its really not that simple.

Jul 17, 2023 2:05 AM

#198

eplipswich

Offline

Nov 2010

1162

So yeah @Kineta, would strongly advise removing the discussion points altogether or at least hugely nerfing the discussion points as suggested by most users here for the next Fall 2023 season (75 per post is still too high, 40 was actually fine back then). The thing about discussion points is that they're too easily abused unlike the other metrics (because 1 user can just easily make many posts unlike the other metrics where you need multiple accounts to get more points instead).

eplipswichJul 17, 2023 2:12 AM

Jul 17, 2023 2:29 AM

#199