A Data Driven Approach to Battlefield V

The purpose of this project is to analyze my play style and identify ways to improve. The data for this experiment comes from manual entering values after each match. If you happen to know how to gain access to the battlefield tracker API please let me know. The critical factors are date, map, gamemode, offense/neutral/defense, class, weapon, assignment, assignment type, practice range, full game, win, game score, kills and deaths. Additionally, it will be interesting to construct models based on metrics for maps and weapons that could be used to predict how well I might play with a new weapon or on a new map as the Tides of War continue.

A Snapshot of My Play Style and Skill

Here is a brief overview of what kind of player I am. Mostly, I play for fun and the challenge the game offers and I would consider myself a mediocre player. I typically play about 8 rounds on weeknights and around 10 rounds on the weekend.

## [1] "I've tracked a total of 158 rounds. During that the interval I've averaged a K/D of 0.993 and a total K/D of 0.825 (Kills = 1286, Deaths = 1558)."

Here we can see the modes that I play and what weapons I use on each map.

A key part of the draw to gaming is the competitive nature and the challenge. Therefore, it was interesting to see whether or not I’ve been improving with each round. First, we can look at K/D as a metric of performance. The plot below indicates that the variation in the outcome for matches varies greatly but the overall trend is clearly showing and increase in Kills per Deaths with increased lifetime rounds played.

Next, we can look at gamescore as a metric of performance. The plot below indicates that the variation in the outcome for matches varies greatly but the overall trend is clearly showing and increase in Kills per Deaths with increased lifetime rounds played.

Maps

Now let’s take a closer look at the maps that I play. An interesting question is how well do I do on the maps that I play the most frequently as well as whether there are any correlations with the maps that work best with my playstyle.

Inspecting my average K/D per map it is evident that Fjell 652 and Panzerstorm are definitely my best maps although the maps I play most frequently are Hamada, Rotterdam and Narvik which are my 6th, 5th and 3rd best maps.

The above data are a bit strange as there doesn’t seem to be any relation between my play and the type of map. Fjell 652 and Panzerstorm are my two best maps yet complete polar opposites in terms of map layout and gameplay, or are they? To see if there is a simple correlation between whether or not my performance is related to the number of time I’ve played the map I plotted my average KD for each gamemode and map vs times played. The result is that there is a weak correlation.

Up to this point we’ve determined that I am clearly getting better as I continue to play and that my performance on each map is loosely correlated to the number of times I’ve played a game mode on a given map. Now we will dig a bit deeper into my gameplay for each class.

Medic Class

The medic class was a natural fit for me as I overwhelmingly played assault in BF1 and the SMG’s are my go to. However, the medic class is exceptionally difficult to play in BFV due to the limited range of the SMG’s and the 5 hit kills required with most of them. Let’s first see if my medic play has been improving over time.

It also appears as though my average is increasing over time while playing medic. When we take a look at my gameplay as a medic for each map we see that my gameplay is fairly consistent although for some reason I do very well on Fjell 652 and Panzerstorm.

Assault Class

The assault class I found to be challenging as I like the high rate of fire but the recoil is difficult to manage. In a similar sense the self-loading rifles are powerful but they require a bit better aim than I’m throwing down right now. Let’s see if my assault play has been improving over time.

Looking at my play on each map shows some interesting deviation from my play as medic.