Apex Legends & Data Analysis

Background & What You’re Looking At

I started recording the basic data you’ll see below on my Apex Legends games, specifically ranked games, queueing solo, on September 29, 2021. I had collected very similar data for years playing Apex – specifically the 60-man battle royale game – both individually and as a part of a three-man team. I’ve found I enjoy collecting this data not only for the fun in crunching the numbers and visualizing it, but also because it will (ideally) give me more information about how I play with each Legend, my strengths and weaknesses, if I play this game too often, etc.

Previously, I had gone overboard collecting data – place we finished, shield level I had, gun I was using, shield level of teammates, etc. It took me about 2 minutes after each game to fill out and became a nightmare. I decided if I kept it simple by recording just the place we finished in, whether I was jumpmaster, how many kills & knockdowns I got, and how much damage I inflicted. All of this information is on display after each game, so there was no reason to rush/hope I remembered what gun I had used. It was simple and quick, two things I greatly appreciate as my attention span has atrophied over the years. I created a form, and at 7:16 pm EST I submitted my first line of data for a dumb project that has seen me track over 1,000 games so far.

What you’ll see below is an embedded Google form displaying my stats with each legend individually, while I was playing solo, on Playstation 4. I did not include games that I played with friends, as not only would that take away from the randomness encountered in the form of other teammates, but also because they play on PC, which forces me onto the PC servers.  I also did not include any games in which I only had one teammate, as that would skew the data (and there was no spot on the form for number of teammates, and I did not want to edit it). Anything else was fair game, and is included below. This chart is live updating, so I suppose with a refresh you’ll be able to tell if I am sitting in my apartment in front of my TV. Please do not use this information to assassinate me.

BY LEGEND

To the left is the data divided by legend, displaying the Games I played as the legend as well as the Percent (%) of overall games, the average Damage when I play that legend, average Place we finished in as a team(1-20), and the average number of Kills and Knockdowns I managed per game with that legend.  In the dark grey area, you’ll find the percent difference when playing that legend vs when playing other legends in most of those same categories. In addition to that, you’ll find a custom made and very propritery statistic called Base Overall Net Adjusted Rating or BONAR. I did consider for a moment whether I wanted to put this onto the internet with a crude penis-joke so blatant, and then decided that yes I did want to do that.

BONAR is, as I mentioned, something I made up from whole cloth. It’s been adjusted slightly a number of times to find what I find is a fairly accurate Overall performance indicator. It is a weighted number that factors in kills, knockdowns, damage done and place the team finished in, hopefully a fairly representative number that indicates how well I perform with each legend.

To the top-right of the table, you’ll see two editable cells – cells P2 and P4 – with red text. You should be able to tap at those and choose how the table is sorted, and, if you’d like, instill a minimum number of games played with that legend to display.

BY JUMPMASTER

To the right is a smaller chart broken down by whether I was the jumpmaster for that particular game. That one isn’t sortable, as it’s just a Yes/No question, but it includes many of the same points of information – Games, Percentage, Place, Kills, Knocks, and BONAR. However, you’ll also see a few new numbers, including number of Wins, as well as Percentage of Games Won. You’ll also see INSTA-%, the percentage of time we land and almost immediately die (ie. finish in 17-20th place).

Presented below is the data for you to “enjoy.”

What is Apex Legends?

Apex Legends is a first-person-shooter (FPS) developed by Respawn Entertainment & published under the Electonic Arts umbrella. Apex was an early follow up to free-to-play games like Fortnite, the mega-popular battle royale style game  released by Epic Games in 2017. Free-to-play games like Apex rely heavily on in-game purchases to make money, and make money they do. Despite me (and everyone else) spending $0.00 to buy the game, players around the world have decided to willingly and by their own accord spend well north of $2 billion on the game as of May 2022.

The bread and butter for games like Apex Legends is often their Battle Royale mode, a game that pits 20 teams of 3 (60 total, no need to get out your calculator) in a “last team standing” competition as the map slowly gets smaller and smaller, pushing remaining teams together and causing more agressive end game gunfights. You can continue reading if you’re curious how the game functions (and would like some insight on what exactly some of the vernacular in the sheet below is referring to).

Prior to the game starting, each player chooses one of 24 (as of 7/13/23) playable characters (refered to as Legends), each having unique abilities or powers that allow for a different playstyle. Some, for example, may move more quickly, have the ability to fly or go invisible, or can absorb more damage before being killed. To start the game, each team of three has a (semi) randomly selected “jumpmaster” – the individual on the team who ultimately chooses where the team will start the game on the map.

From there, the game begins. Players start weaponless, and need to scout the map to find weapons, ammo, and their opponents. Inflicting enough damage on a player knocks them down, allowing them to be picked up by a teammate or further shot/grenaded/punched to death. Once all three teammates have been eliminated, that team is removed from the game. Those are the broad strokes, and I’d advise you to read the Apex wiki if for some reason you’d like to learn more.

What is Data Analysis?

Hmm, interesting question. You know like, numbers? And how some are larger than others? So like, what if you look at one number and it’s bigger than the other? That like, tells you something, right? Congratulations, you’ve just done a data analysis.

To be more specific, according to Coursera (more on this to the right), data analysis can be defined as “…the practice of working with data to glean useful information, which can then be used to make informed decisions.” More on what data I’m working with and what I’m hoping to glean can be found on this page. I’ll be doing some very very light analysis here, in Sheets. In fact it’s more of a Sheets chart-building exercise.

Why Are You Doing This, Ryan?

Good question. I do not know, but I have some guesses below.

I like numbers and data. You know this if you’re on this site. If you don’t know this, go read Ryan Woerner by the Numbers.

I started collecting data on Apex Legends in 2019, when the game came out, initially with the fairly straightforward task of figuring out which character I was best when playing as. Since each of the (currently) 19 legends feature unique abilities, no two are played identically. While people will argue endlessly online about which Legends are better or worse, which are over powered and which are so poorly built they’re rarely picked, one or two would likely better fit my personal playstyle. I decided that collecting information on how well I play (what place we finished in as a team, how many kills/knockdowns I got, how much damage I did, etc) could paint a somewhat clear picture of who I am best with, and who I should avoid picking at all costs.

The Data & Summary

Below you’ll see the editable (P2 & P4) chart displaying the data. Scroll beyond it to see some conclusions.

Early Conclusions

While the data below is live-updating (meaning, the early conclusions/data may change), most conlusions drawn at around 1,200 games prove to be interesting to look at, if not statistically significant.

As I mentioned, the information I’m collecting points can shed light on two main variables: the legend I’ve selected for each game, and whether I was the jumpmaster. Each of the two variables here appear to affect the outcomes differently — while legend choice impacts my style of play and how successful I am, the jumpmaster tends to have a much more pronounced impact on placement specifically. I’ve detailed some conclusions below.

Jumpmaster

The first data point that jumps off of the [virtual] page when looking at the jumpaster comparison is the breakdown of Yes/No responses. With three players per squad for the jumpmaster to be selected from, you’d imagine that I would be assigned the responsibility about a third of the time. However, about 1,200 games in, I’ve contolled my team’s landing spot almost two-thirds of the time (65%). While this may seem like a result of poor tracking, it’s more a result of how the game determines the selection of jumpmaster.

Each of the three members of a squad is given the opportunity to select which legend they’d like to play as in succession, and only each legend can be chosen once. Each player has about 10 seconds to decide their legend before it passes to the second member of the team, and then the third. Because the game is a 60 player online game, if a player does not select a legend within their 10 second window, the game selects the player’s “default” legend for them before passing to the next player. The player randomly determined to select third (ie. last) is given the opportunity to be jumpmaster as a small consolation for having the third selection of legend. However, this only takes effect if the player selecting third actively selects their legend — if the timer instead runs out and the “default” legend is selected, the jumpmaster designation passes to the last player that did actively select their character. While there is more to this process that isn’t worth explaining, it’s important to know that actively selecting a character (rather than playing on your phone and letting the game pick your favorite legend for you) greatly improves your chances of being Jumpmaster. Because most people playing this game have a character they play as most often (ie. a “main“), many are content to allow the game to default to choosing that legend for them. Since I actively try to play as many different legends as possible, and therefore often actively choose my legend before each game, the tendency for me to be jumpmaster is greatly increased. So that explains that.

Perhaps the most pronounced and illucidating set of data I’ve collected so far are the improvements in score and placement when I am jumpmaster. My personal preference is — and has been — to attempt to find a spot on the map where there are few, if any, other teams landing. I would rather use the first minute or two of the game finding a weapon than use it running from another team who happened to get one before me. Others do not necessarily feel the same way, choosing to drop (ie start the game) in more clustered/congested areas, leading to higher-action (and often quicker-ending) games. Insta-Die %, a measure of the percent of time the team finishes in 17th place or worse paints this picture adequately. My team insta-dies about 20% of the time (one fifth) when someone else is jumpmaster, compared to just 7.6%, a reduction of almost two-thirds. Likewise, percent of time finishing in the top 5 teams raises from 24% to 38% when I choose where the team lands.