Act! Addict! Actors! — Analysing Past Event Results
Which event was the scariest bloodbath for A3! rankers?
Hello! This is the third part of my A3! x Data series, and we’ll be travelling back in time to take a look at previous ranking results for both the A3! EN and JP servers! If you’re not familiar with the A3! game, do check out this series’ prologue for a quick explanation as well as the first part of the series, which goes through the less intimidating task of compiling and analysing all of the characters’ basic information. Meanwhile, the second part did a deep dive into A3’s scoring mechanics in an attempt to find the most meta character for teambuilding. This time, I decided to play around with the previous event results as I wanted to see if JP trends could help to predict future bloodbaths in EN. As you might notice later on, this article will assume that you are somewhat familiar with the A3! characters, but if you aren’t the wiki is always a good reference!
Before we jump right into the data, a few notes:
- Most of the ranking data was from the A3! subreddit, Yaycupcake Wiki pages and this amazingly comprehensive fanmade spreadsheet which works as an hourly event border predictor and ranking data archive.
- Making these tables was relatively simple, but tedious as I had to get the data from all of the separate event pages. Afterwards, simple tabulation, formatting and chart plotting helped with the analysis! Please view the full spreadsheet here!
- A3! EN and JP vary wildly in popularity, which will explain some of the later findings! For context, the busiest event in EN so far was the 1st anniversary which had an estimated player base of 17,000 players, while A3! JP is estimated as the 4th most popular game of its genre in Japan and recently passed 7.5 million downloads… then again please keep in mind that JP is about 3 years older than EN, and that this kind of game genre is much more established in Japan!
- Reading the raw tables: The darker the cell colour, the higher the tier cutoff for that event in the raw tables below.
Analysing EN Results
Interestingly enough it seems like the later events were more competitive (1st anniversary onwards), with the exception of Into the Night (because Itaru, the event SSR, is by far the most popular character in the game for both servers). Probably because A3 wasn’t as well-known back then? Sorting by T1’s (top 0.1%-0.5%) performance gives a fairly obvious gradient, meaning that in general if T1 is scary the rest of the tiers are probably pretty scary too. Knights of the Round IV On Stage (fans call it Kniroun for short) could be seen as a bloodbath for almost every single tier, no surprises there…
I think Kniroun severely skewed the colors/average points, because the average points jumped by 100k points for quite a few tiers when I updated the Reddit info to include newer events. 3 Cheers for Mankai Company was pretty competitive too, but in the mid-tiers… I guess people wanted the SR token prize? The event format might’ve inflated the point cutoffs too, similar to the Training Camp events (disclaimer that I didn’t start playing A3! until Valentine’s Magic, so my knowledge about earlier events mostly comes from the Wiki). The Training Camp events (these events give players the chance to earn a very powerful card as an advanced event reward, and have a slightly different mechanic from most events) tend to have fairly high cutoffs for the lower tiers, although I think that’s because of the Training Camp Leaders’ rather significant stat boost making life easier for us mere mortals during the events… the average T1 performance makes me think that these events weren’t insanely popular but their formats helped to inflate the cutoffs.
Interestingly enough, it seems like One Day Princess was the Kniroun of Seasonal Events (like normal events, but the scoring system is measured in items instead of points) in the sense that it was really competitive across all tiers. I probably can chalk it to SSR Tsumugi + SR Itaru as the event cards… Various Vows of Love was also fairly competitive, but I suppose that was because of the double SSR prize. Homare and Tenma aren’t quite as popular as Itaru and Tsumugi and people were probably saving for Kniroun, so maybe that’s why it wasn’t more competitive. I was wondering if Liber changed the mechanisms to make scoring points easier for new events, but I don’t have enough data and the lower competitiveness of Valentine’s Magic seems to disagree with that theory (not that I’m complaining though, I love that Hisoka card).
Another pretty interesting thing I realized was how scary Into the Night was… Hardcore top tiers but the tiers below 40% (cutoff for 1 copy of SR Azuma) were some of the least competitive overall. Into the Night was a pretty early event which might have had a smaller player base, and casual players or non-Itaru fans might have just decided to give up after seeing the bloodbath. The rankers really did anything to get that bunny Itaru SSR card…
If the tables are scary, I hope that these charts might help:
The charts showed me another interesting trend, which is how the top 3 tiers closely mirror each other in shape, while the other tiers are considerably flatter. #1 players are in a league of their own, but in general competition in tiers after T3 is significantly lower, even if the event is a bloodbath. Guess that’s the power of whales/rankers…
One more point in the EN result analysis: Out of curiosity, I grabbed the estimated EN player base sizes for each event (up to Die by the Sword) from the hourly tracker spreadsheet. If I’m not mistaken, the estimates can be calculated by counting the number of players in the top tiers and multiplying to calculate the total active player base size for that event. For example, 11 players were in the top 0.1% for Die by the Sword, so multiplying that to 100% would give a rough estimate of 11,000 players for the event. The spreadsheet gave a value of 11,310 players, with a small margin for error of +/- 10 players. I realised that EN’s active player base has been shrinking since the 1st Anniversary event, 3 Cheers for Mankai Company. Uh oh.
Something interesting to note is how the player base size increased by nearly 2,000 players when Kniroun came around, possibly because of the hype surrounding the event, and maybe from fans who took a break from the game before returning to rank for Lancelot Itaru.
Here’s the summary of what we’ve seen so far for EN: Itaru fans are terrifying, good luck to all those who are preparing to tier for his events…
Comparing EN and JP
I told myself that I wouldn’t be insane enough to collate the ranking data from all the previous JP events… but apparently I can lie to myself as well as a certain cabbage-head. For this part, I’ll be comparing the past EN data from the previous part together with the JP data until SHINOBI Adventures, which was the most recent EN event at the time I made the spreadsheet. Since then, the sheet should have been updated with newer events! Note that non-ranking events aren’t included here, since I’m only interested in ranking data.
Here’s the raw data:
The color scales follow the same conditional formatting I did for the EN events, but now the colors take both EN and JP events into consideration. The darker the green, the higher the points. A few things I realized from the raw data and charts:
- Wow. JP is very scary. The color is basically concentrated on JP’s side… which makes sense, considering the significantly higher popularity and player base of A3! JP.
- Looking at the chart, we can see that the ranking horrors we’ve experienced in EN are nothing compared to JP. To put things into perspective, the highest-ever event point score on EN (Kniroun), is on the low end of JP’s 1st place scores… and JP Kniroun was basically off the charts.
- It seems like at the time JP was playing EN’s recent events (Kniroun onwards), JP’s top rankers were becoming more insane… In fact, the trend of events being more competitive as time passes was observed in both servers.
- Sorting the events by T1 performance, we can confirm what the charts were implying: JP players are really on another league altogether, because the most competitive events across both servers all happened in JP. Something interesting to note is how EN Kniroun managed to be more competitive than average by even by JP standards, and that events which were JP bloodbaths were also bloodbaths/very competitive in EN, indicating that EN and JP fans have pretty similar tastes. I’m pretty sure that someone made a player survey which ranked all the characters based on EN popularity… I think it was this one! It found that EN really loves Itaru, Misumi and Tsumugi, which tracks with how scary their events are on both servers.
One more thing I did when comparing both servers was to get the average cutoffs for all JP and all EN events and compare them to see exactly how scary JP was…
Basically, JP beats EN for virtually every single tier, except for T10 which is the top 50%-60% where EN very very slightly beats JP. The average cutoff differences increase for more competitive tiers, which basically says that JP competition is scary but casual JP players might be able to get similar rewards as casual EN players. This average point difference might be helpful when trying to estimate future EN cutoffs but do note that EN hasn’t had a lot of events compared to JP, so scary events such as Into the Night and Kniroun would inflate the average cutoffs in EN more than JP!
Summary of this section: All hail JP players!
Into the JP-verse (Be warned, JP spoilers from here onwards!)
This part will go into all the events which have been released in JP so far, as well as my attempt to predict future bloodbaths. All the raw data and charts for the JP data can be found here!
First I’ll go through how I collated the raw data for JP events, since I wasn’t able to find a previous event compilation unlike EN’s results.
Step 1: Filter out all non-ranking events.
I made a list of JP events, based on the structure of my existing EN event list. It was pretty simple since all I had to do was change the EN event names/links to the JP equivalents, and reorder some of the events to get the running order right since JP had a slightly different event order. Next, I just took out the Etude, Promotion, Rerelease and Campaign events since they had no ranking.
Step 2: Hunt down all the results from Yaycupcake’s Wiki.
Bless whoever runs the wiki for all the data… After a lot of copying and pasting, I made a table similar to the one I used for part 1’s EN data. Again, events which used the point system are separated from the item system, and the darker the shade of green, the more points relative to other events in that particular tier.
The first thing you might notice is that unlike the EN event data, the JP data has some gaps… it seems like only the earlier events tracked the bottom 2 tiers, and the top score for some events was unavailable. An interesting thing to note is how the events with 1st place data gaps seem to be clustered together… I wonder what happened then…
The next thing you might notice is how JP’s ranking system is pretty different from EN’s. JP uses a fixed rank for the cutoff, while EN uses percentage tiers. This legend helps you with converting them, but seeing how the tier rewards are the same across both servers, I’ll argue that the percentage for JP is a bit similar to EN (probably smaller though, considering how JP is super competitive).
The last thing which you probably wouldn’t notice unless you’ve seen the Event List before would be that the ranking data includes the event SRs while the list doesn’t… hopefully this might be somewhat helpful if you’re looking out for more cards of your favorite characters!
Time to start the analysis… At first glance, I couldn’t really discern a trend over time for events. EN showed a fairly strong trend of newer events becoming more competitive, but curiously enough, JP seemed to be the reverse: Bloodbaths tended to happen during events which were quite a long time ago… of course, the huge cutoffs for bloodbath events also threw off the color scale, which made comparisons more difficult. That’s where these charts come in:
Looking at the charts, it seems like my guess was partially right: Recently, JP event points have been seeing a decrease in the insanity of #1 rankers, as well as a slight decrease in insanity across most tiers. However, the point cutoffs for T2 and below have been fairly consistent the entire time, so I think it’s safe to say that JP’s competition levels are still pretty consistent. On the other hand, it seems like competitiveness for seasonal events might be increasing… However on closer inspection I realized that the recent peaks were for events where Itaru was the event SSR (Fantastic Fairy Tail and Magical Fragrant Dreams), meaning that the downward trend might have continued to persist if it wasn’t for the power of Mankai’s resident game addict… The recent release of Act 11’s gorgeous art might revive JP’s ranking fervour, so who knows?
The general observations I made with the EN data continued to hold with the JP data: Competitive tiers mirror each other in terms of shape, while the lower tiers don’t fluctuate as much. However, it seems like the competition in JP is across more tiers than EN, since the mirroring effect seems slightly more pronounced. The other observation was pretty simple: Everyone will go insane for Itaru.
Sorting the events by T1 performance shows that if T1 is scary, the chances of the other tiers also being scary is pretty high. A notable exception is Starring Party (3rd anniversary event), which has an insanely high T1 cutoff but otherwise average cutoffs for T2 and below. Looking at the featured SSR characters, we can see that Itaru basically owns the top of the list, while Citron shows up a lot at the bottom. Ouch. Curiously enough, the order the characters appear in this sorted event order seem to echo the findings in the A3 EN Player Base Survey… of course, the player bases from both data sources are different, but it’s interesting how the characters in the extreme ends of the lists seem to match. If you’re curious about the data, I grabbed the results table from the Reddit post and sorted based on JP rank, do let me know if you see the similarities too or if I’m just seeing things…
Taking a closer look at the top events:
- Fantastic Fairy Tail’s champion managed to get more than a million event items. Considering how the average T1 cutoff is 34,000 points… HOW.
- Kniroun, I Come with the Night and Into the Night (Walk with me, Talk with me! in JP) were the scariest bloodbaths as expected. What I wasn’t expecting was Risky Game. Hisoka?! I was really confused, especially because the rest of Hisoka’s events weren’t exactly super competitive… but I guess Scarlet Mirror was no slouch either. Gekkagumi fans, good luck (especially considering how Chikage’s only SSR event, RE:FebMAGIC, ended up as one of the most competitive seasonal events)…
- The next thing I realized was this: A lot of the most competitive events happen during the next round of Training Camp events. Risky Game and Fallen Blood would enjoy the seasonal bonus from the fully bloomed SSR Training Camp Leaders, and Risky Game in particular would have seen a lot of people using the fully bloomed Masumi from Let’s go Samurai Cinema Romance, since people would’ve been after the SSR Itaru. Training Camp events are scarily useful and I salute the person who posted that insanely high score for Risky Game…
- For seasonal events, it seems like Itaru, Banri and Tsumugi reign supreme, with events like One Day Princess and Twilight Gunman together with the Itaru events being especially competitive.
So… future bloodbaths? It’s actually a really simple answer: Look out for any event with Itaru as the event SSR. The next one for EN is I Come with the Night (1st mixed play), and if JP is any indication we’re in for an event which could give Kniroun a run for its money… other events to look out for would be the next round of Training Camp events, and anniversary events! All the best, everyone!
If this series was interesting to you, more resources made as a result of my A3!-induced insanity can be found here! Additionally, I made an Instagram-friendly version of this post which can be found in this Instagram Guide together with the rest of the series, which is part of my quest to make spreadsheets Instagrammable! For easier reference, here are the images:
