War Robots WR Data and Analysis

By the Numbers: At the Gates of Hell

"Abandon all hope, ye who enter here" comes the warning as Jay zips through Android Bronze. Here's the statistical breakdown!

My rise through the ranks of Android Bronze League has been mercifully swift. While I recognize I’m in a tier with lower-skilled pilots than the ones I’m used to (Diamond 2 on iOS), it nevertheless feels a little bit different to me. Like when you see something familiar, but it’s off just a touch and you can’t quite put your finger on it.

Part of this is undoubtedly down to the Anniversary Event, which is the motive force behind most sudden shifts in general player behavior. Mainly, a beacon-centric focus, leading to a higher than expected occurrence of Cossacks, Stalkers, and, yes, the fearsome double-Ecu Stalker. But while I’m withholding judgment until I’ve had more experience and can compare apples to apples, it certainly feels a bit like there’s more built-in variance in these Android matches.

Good players can be quite good here, but bad players can be abysmal. Like the RDB Griffin who did nothing to save my Ancile-Zeus Fury right next to him from getting mauled by a Plasma Gareth. These games have much more of a crapshoot feel, but I can’t isolate whether or not that’s down to player pool or just the level I’m currently playing on. With a current 68% win rate, I’m not even stopping for gas.

This explains why I’m continuing to experience smaller sample sizes. I’m just not playing enough matches in each of these lower leagues to get to my preferred sample size of 100 data points before graduating up. It’s still useful data, but maybe data taken with an additional grain of salt.

If there’s an upside I’m already seeing, it’s a lower occurrence of pilot names like FlexxNutzz and PRO BR CANAL.

On to the data!


Average Player Level (Bronze 2): 19.18 (+0.24)

Average Player Level (Bronze 1): 21.89 (+2.71)

A modest nudge up from Bronze 3 to Bronze 2, and a nice little hop up to Bronze 1. This is certainly to be expected- as players play more and advance within the game, they will begin to be more effective and skilled in the game. It’s also worth noting that the Workshop opens at level 20, so many pilots will start to have access to weapons like Magnums and Tarans, both staples of the game at any level.

Median Victories (Bronze 2): 53.00 (+5.00)

Median Victories (Bronze 1): 81.50 (+28.50)

As I explained in our previous Android stats post, I’m using median rather than mean here because of the ability of just a few experienced tankers to throw off the validity of the data. No 10,000+ winners this time, but people playing at a level they were clearly overpowered for was not a rare sight.

Again, a small jump to Bronze 2, and a larger one to Bronze 1, consistent with our earlier findings. Pilots are moving up the leagues, as intended.

Average Trophies (Bronze 2): 57.39 (+14.25)

Average Trophies (Bronze 1): 77.50 (+20.11)

So more active players tend to move up the leagues. Again, nothing mindblowing here, but it’s never unwise to check your statistics against your assumptions. Because this is a snapshot stat instead of a historical one like Victories, mean (average) works just fine.

Player Divisions

Here’s a look at what each league level is pulling to stock its matches. I’ll keep this historical so we can see any trends as they form over time.

Fig. 1: League distributions

A few things I thought sort of stood out here. First, the finding that there were more Bronze 2 players proportionally when I was playing in Bronze 3 than in Bronze 2. This could be attributable to any number of factors including online player pool at my time of playing, to smaller-sample size variance.

Next, there was a huge jump in Silver 3 representation from Bronze 2 play to Bronze 1 play. In Bronze 2, Silver 3 made up just 3.17% of the population. Now you might consider the fact that there’s a two-league gap between them, but the previous two leagues (Private 1 and Bronze 3) had fewer qualms about bringing in higher-league players (about 15% on average). This seems to underline a sort of “break” between Silver leagues and the ones beforehand, which were much flatter in terms of upward mobility.

In other words, in Bronze most everyone is equally weak with slight tracks of improvement. In Silver, which needs a lot more league points to ascend, the tier as a whole is a lot more stratified. You can also see this reflected a couple of other ways in the data. For instance, the “band” of reach decreases from six leagues in Private 1, to only four leagues in Bronze 1. This might potentially be a factor of available population, we just don’t have any way to know. Note also that in Private 1 more than 50% of the pull was from upleague. By Bronze 1, just under 20%.


Average Hangar Slots (Bronze 2): 3.12 (-0.02)

Average Hangar Slots (Bronze 1): 3.25 (+0.13)

A small step back in Bronze 2 underlines that it tends to have more in common with Bronze 3 than Bronze 1. Things look to start moving upwards in earnest in the latter.

Average Robot Level (Bronze 2): 3.67 (-0.22)

Average Robot Level (Bronze 1): 4.76 (+1.09)

Average Highest Weapon Level (Bronze 2): 4.60 (+0.06)

Average Highest Weapon Level (Bronze 1): 4.99 (+0.23)

Bronze 2 about the same as Bronze 3, Bronze 1 with a big jump. Hmm…where have we seen that before?

What bots are people running?

Fig. 2: Bot distributions

The higher up the ladder I go, the more familiar the metagame becomes. Big drops for the Boa, Golem, and Vityaz, those old stalwarts of the lower tiers. Both Natasha and Griffin took a curious dip in Bronze 2, but rebounded right back up to more expected levels alongside the Leo.

As we’ve seen in higher-league iOS, the Griffin should only continue to climb in prevalence, blotting out some of the others like kudzu. Definitely something to keep an eye on.

Now here’s the same data, sorted by Bronze 1 prevalence.

Fig. 3: Bot distribution by prevalence

This affirms the possible conclusion that chest-won  higher-end bots rampaging through the lower levels of the game might- just like clubbing- be something akin to sharks, where the perception of threat exceeds the actual threat.


The matches at this level continue to be very swingy and unpredictable affairs. Coming back from a 3-bot deficit is not uncommon, which is a factor of the shallow hangars most pilots a this level still field. When you’ve only got 60% of the hangar I have on a per-bot basis, you’re simply going to mech out a lot more quickly. It almost feels sometimes like I’m playing three games in one.

Game One: The first four minutes

Game Two: Reset, get a look a the new battlefield situation

Game Three: Be a part of the cleanup crew, or the last man standing as the tide of battle crashes ashore

I can’t say it’s not fun, though it’s not nearly as fun as my iOS Diamond play when comparing game to game. But this is the end of the road for where I was told things stay easy, for I am now standing at the gates. And my comrades on the Wiki Forum have done their level best to keep me optimistic about the next round of research!















0 comments on “By the Numbers: At the Gates of Hell

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

%d bloggers like this: