*1

[Gragg9]

Yes, I am still alive! Barely! What better way for me to write a comeback media than to pick one of the most tangentially related things to the league and put an incredibly disproportionate amount of time and effort into it only to get enough pay for a few dotts packs. That’s right the title isn’t misleading, I have put together an in-depth report of the communities counting-bot channels. The next paragraph or two will be a bit of an explanation as to why and some background regarding this project but if you don’t care about any of that stuff you can skip below to the line for the good stuff.

If you’re reading this the Minnesota counting room will have recently hit 20,000. I hoped to drop this media the night it happened, but wouldn’tchya know it life tends to get in the way. The Minny counting bot has been around since June 2020 but really started to pick up around the time I entered the league (S28 class best class). Such a simple little game has been a lot of fun and led to many hilarious moments. To finally reach 20k is a really cool achievement, making it one of the highest active counts in the world.  Minny isn’t the first to accomplish this in the league however, with the London counting room achieving legend status of its own by hitting 20k back in May of ‘21 and is still going, leading all other ISFL locker rooms. That London counting room was the first added in the ISFL in April 2020 but now there are 17 such channels including the ISFL gen chat #counting.

So why the hell did I decide to take a deeper look into counting bots of all things? Well long story short I’m making a career change into data practitioner/data analyst work and have a lot of learning to do to become qualified. My current job has me rapidly improving my database management and sql skills but data aggregation, data visualization, and the many related tools I’ll need to teach myself. For me the best way to do that is through projects like this where I pick an interesting subject that poses a lot of problems to solve. Counting channels for example have a ton of data to gather in a mess of data formats (text, integers, date, datetime, etc.). You heckers using math in your counts certainly don’t make it any easier. Anyway, the idea is to build a little presentation on this subject to hone my data skills and gain some familiarity with Microsoft Power BI, which I used to build all of the visualizations below. On that note if you have any questions or suggestions, particularly if you have experience in the matter, please please please send me a message.

---

Let’s start with the ugliest chart! Normally the user would filter this down to a few teams but since I have no way of sharing this with interactive features, y’all will have to deal with a bit of a jumbled mess. Regardless, there are a few things we can immediately see from this visualization:

London has been the king for a very long time. Since September 5th, 2020 to be exact. Just a month later they were leading all other servers by 8,000+ until very recently.

But wait! What’s the weird drop in London?? Well that’s an interesting case. Very briefly the London count got reset by a bug in the counting bot. Fortunately they were able to get it restored to the previous count.

Colorado, Sarasota, and New York have some incredibly long streaks. If they were compared to anyone except London they might be the stars of the show.

If London is the Tom Brady, Minny is the Patrick Mahomes. Despite being younger than many other active streaks, the Minny run has been climbing at the highest rate in the ISFL. So long as they don’t regress to the mean, they will likely catch London sometime late this year. Assuming of course neither of them pull an Andrew Luck and retire early.




Remix time! This time instead of overlying every server I stacked them on top of eachother, because it’s not a contest right? This way it’s easier to see the total count between all servers, and how devastating some resets have been. A few points from this chart:

As we see the count reach nearly 65k it’s hard not to set the goal of 100k between all ISFL servers!

Almost all at once the counts for Chicago, ISFL, KCC, and DAL were ruined, resulting in a visible crash which we only recently 'recovered' from.




Moving along we have a visualization of counting bot activity over the past couple years. The hayday was certainly late 2020/early 2021 but there is a definite resurgence in the past few weeks.

Surprisingly, the average counts per day from all servers combined is 255.

The most counts in one day is nearly 3,000. That’s more than the record of 5 current servers. If those counts were all done by one user, that user would nearly be in the top 10 counters of all time. 2,500 counts in a single day was achieved only a couple days later.


[/img]

Next let’s take a look at who sucks at counting. This view shows what percentage of all counts are of any type of error, whether that be a reset (RUINED), guild save, or user save. The bookmarks visible allow toggling between each specific error and if you’re interested in seeing those leave a comment or shoot me a message and I can provide.

Unsurprisingly the general ISFL server has the highest rate or errors and it’s not particularly close. In second is the OCO server though likely due to its relatively low sample size. The big counters in the London and Minny servers actually mark the middle of the pack. Each server had rocky beginnings with lots of fooling around before streaking hard. NYS at the bottom is an amazingly accurate server that never really had a wild phase + a relatively low sample size.




Now we know who sucks at counting but HOW do they suck at counting? Well let’s just say the ISFL and Dallas are no strangers to the number 0. There is rarely a safety net in these two servers with most errors resulting in a rest.

Personally I’m a bit shocked by the average user save rate in London. I fully expected them to be near the very top in their attempt to preserve their legendary run. I guess their wild early years are still visible.

A few servers are still figuring out what a guild save is. For those that don’t know, users can donate 5 personal saves to get one guild save. If anyone in the server messes up the count and doesn’t have a save of their own, that guild save is used. It's a good way to keep inexperienced counters from resetting the streak. Colorado and Philly think that kinda fanciness if overrated, and everyone knows donating guild saves to ISFL is an absolute waste.




If you’re still reading this media for some reason you probably have a vested interest in counting somewhere in the ISFL (you’re a weirdo like me). If that’s the case you may just find yourself in this next viz. Here is a scatterplot with total counts and total errors (resets+guildsaves+usersaves) with users over 1000 total counts. Naturally users that count more will eventually have more errors so who’s really accurate?

In this view users like Lucid and Bigred are well above the crowd, with 82 likely taking the title for most accurate counter. A few counters below 1k have a higher accuracy but maintaining that to nearly 8,000 counts is the real challenge.

Unsurprisingly perhaps, deleted users tend to be trolls and hooligans. Their error count stretches the field. Other users low on accuracy are quite a surprise to me as they are still very careful. Just goes to show how a simple game can be so tough.




The keen eyed among you might have noticed a filter applied to that last visualization and the absence of a certain prolific counter: The Count of Counting himself, Katarn. It’s a bit of a cliche to say, but he literally breaks the scale of the chart. No one is even in an adjacent league and Katarn is still maintaining activity. All with fairly high accuracy to boot.




Here we have the user dashboard which is by far the coolest thing here but also of little use to y’all since you can’t interact with it. Instead you get to see a detailed breakdown of my counting stats (Such as the one I’m most proud of, 3.5k without ever resetting a server). It’s very easy for me to grab other’s data so if you’d like to see your own (or someone else's) report please don’t hesitate to leave a comment or send me a message. 




If you don’t care for visualizations and prefer to see some stuff in table format, there’s a quick shoutout to you. If you’d like to see any of this sorted a certain way, again, please don’t hesitate to ask. I enjoy seeing who’s used the most saves or seeing how these stats change over time with the slider. Also that curse word counter is one of the more interesting bits of code I’ve ever written.




Finally, we have my pièce de résistance. I don’t think anyone has ever created a chart that communicates its point so clearly without any ambiguity. As this visualization is perfect I won’t be taking any comments or suggestions on it. 


Ok I’ll give it a break now. Just to reiterate it one more time (and earn myself a few thousand bucks while doing so cause I’m’a be so underpaid for this media), if you have any questions, suggestions, or want to see another view from the report, please let me know. I spent way too much time putting this together, y’all might as well get some good use out of it. ‘Till next time!

[MattBake12]

another banger

[124715]

2

[Gragg9]

2

3

[domffl]

4

[DREAMSLOTH]

5

[Memento Mori]

Finally, we have my pièce de résistance. I don’t think anyone has ever created a chart that communicates its point so clearly without any ambiguity. As this visualization is perfect I won’t be taking any comments or suggestions on it. 

lmfao

[LB3737]

I gotta get back to counting i thought id be higher

[Gragg9]

6 @domffl

[firstfray]

7