War Robots WR Data and Analysis

By the Numbers: Striking Gold

Jay reviews the data from his first set of League matches this month. What bots are people running? How is the meta shaping up? Let's find out!

It’s been awhile between stats updates here on the Diary, not because I haven’t had the data, but simply because I haven’t yet located that elusive 25th hour of the day. Between the other updates here, running Aurora Serica, and maintaining our clan’s website, it’s been a wonderfully hectic few weeks.

When the season last reset and dropped me to Gold 1, I started taking notes on the players I saw in the game for a stats post. War Robots had other ideas, however, and quickly promoted me back to Diamond 3. This forced me to scrap my old data, and begin afresh, but I did record all of the players in my first ten matches in the league. That gives us my standard sample size of 110, which isn’t colossal but enough to draw some conclusions. Let’s jump right in and take a look!

Here are my own numbers at time of writing.

Level: 30

Cups: 202 (-22)

Victories: 804 (+123)

Onto the numbers!


Average Player Level: 29.70 (+0.36)

I may go ahead and drop this stat, because it hasn’t changed much and seems unlikely to. Something really drastic would have to happen to knock this down in any significant way.

Average Victories: 824.86 (-16.70)

Stable. I’m continuing to approach parity here, and for all intents and purposes pretty much have. The last time I recorded this statistic having a sub-800 value was two By the Numbers posts ago, on 09 February. Pilots in my peer group have had a fairly consistent level of game experience.

Average Trophies: 315.62 (-115.00)

Now that is a substantial difference. Back in February, the average trophy count was 384.28. I’m hesitant to draw an obvious conclusion- players are simply playing less- but then we’d need to construct a parallel hypothesis that has some entirely different factor as its conclusion. Is the matchmaker simply smart enough to match me with my peers even this granularly? I’m skeptical of that point, but only Pixonic knows the truth. We can just guess.

For my part, my activity is down, but for overwhelmingly positive reasons. I’m more engaged than ever with War Robots, and liberty time I may previously have spent battling has instead gone into clan leadership and the clan’s website design.

As an aside, designing the Aurora Nova website has been a really interesting experience. The site looks fantastic (I’ll admit to some bias here), but I made a valiant effort to publicly proclaim to major War Robots social media outlets that I would be happy to provide my expertise and guidance to any clan that wanted to do something similar.

Know how many clans took me up on the offer? Zero. Not a one. And yet, thanks to that website we’ve seen our recruitment explode. That’s how I took my first Chapter, Aurora Nova iOS, from 8 to 35 in a week. It’s how we grew past that, setting up two more Chapters in the days that followed. While I know a number of other clans are struggling for recruitment and retention, we’re rocketing up.

At this point, the offer for free guidance is off the table. Not because I don’t want to be helpful, but because my dance card is overwhelmed. In addition to the duties of Chapter Leader and clan infrastructuralist, I’ve got no less than eight articles in development right now, as well as writing the curricula for the upcoming Aurora Nova Leadership Academy.

That’s right. Our clan has it’s own West Point.

Maybe I’m boasting a little bit here, but I am extraordinarily proud of what we’ve accomplished. But yeah, my activity has gone into the toilet. I just don’t want anyone to take away the impression that I’m losing interest or burning out. I’ve never been more engaged, and a large part of that is what’s going on outside the battlefields. When people ask, “what’s the point of clans,” that’s certainly one of them.

Player Divisions

In my last stats piece, we introduced this metric to see what the matchmaker was putting me up against. I found it curious at the time that I seemed to be playing downwards rather than upwards so much of the time. Has that trend persisted?


Nope. Gold 3 players- never common to begin with, have fallen off the map. Gold 2 and Gold 1 are both down, rather significantly in the latter case. In addition, I’m facing more opponents in my same league, Diamond 3.

If you were designing a system from scratch, you’d probably expect a player to play primarily in their own league, with an approximately equal proportion of pull-ins from one level up, and one level down. Two weeks ago there was an imbalance that favored me, pulling more from below and much less from above. Clearly, the matchmaker this time around has it dialed in. I’m facing solid opposition, and go into every game feeling like I have a good chance to win it.

I understand this is not a universal experience, but it’s the only one I’ve got.


In my last piece, I noticed with curiosity that the proportion of players who had five-slot hangars had actually dropped by just over 7% from the previous sample. This week, however, it does seem that the ship has righted:


The proportion of three-slot hangars has nearly halved, and not before time. It’s difficult to imagine being a full contributor to the team effort at the level I’m at with only three bots to choose from, so I’m not surprised to see that number dwindle.

The five-slot hangars went up by around 50%, which is what I’d been expecting to see for precisely the same reason I’d expect to see a decrease in three-slot hangars. More slots mean both more options and “extra lives,” as I covered extensively in a recent DREDDGUIDE.

That growth comes at the expense of the four-slot hangar, of course, which takes solid hit. I’ve recently become aware of a school of thought held by a small minority which purports that four slots is actually better than five. I have yet to see any real data to substantiate this- the argument seems to be economic in nature, since it costs less to upgrade five slots rather than four. Then again, you could make the same argument for three slots. While I’m intellectually curious to see what might underpin this, I have a suspicion it’s a little less than effective than advertised.

And now, a word from our sponsor…

Average Robot Level: 8.12 (+0.48)

Average Highest Weapon Level: 8.56 (+0.38)

This is also on track with what I’d expect to see. As I improve my gear and perform better on the battlefield, so too are my enemies improving their gear. Or rather, I’m simply facing enemies with higher gear. These are comfortable numbers, nothing seismic or volatile but rather a steady progression. Don’t they give you the warm and fuzzies when they do that?


Okay, maybe just me, then.

On a lark, I thought I’d calculate my own scores to compare. Though my hangar varies as I continue to tinker, the common roster I’m presently running with has an average robot level of 8.00 and average highest weapon level of 7.20. That’s encouraging, because it suggests I’m doing a little more with a little less (especially since this data is from the start of the new season and my own stats are current at time of writing). I didn’t realize I was so upside-down with my upgrade path, however, so I’ll be focusing on bringing those weapons up. Behold, the power of knowledge!

What bots are people running?

In Magic: the Gathering parlance, I’m not sure that the meta has been fully “solved” but it’s certainly settled. As with some of the data above, there’s just not a lot of volatility. Here’s what I saw in my sample matches.


Not a lot to take away, but a couple things caught my eye. We still see a near-complete absence of the Wild Bunch bots. Spotting a Jesse, Butch, or Doc is a rare treat. Is it because the common misconception about them is that they’re poor? Is it because those most inclined to drop the coin needed to get them are invested players playing at a higher league than me? Or is it simply a case of scarcity given the distribution vector. Hard to say, but we’ll have another crack at them in about a week or so- and I’m not letting Butch pass me by this time.

There’s a slight increase in Rhino sightings, and this corresponds with my personal impression in other games. The Patton has nosedived, and given the resurgence of interest in the Aphid Patton on community platforms as late, I’m not entirely sure what’s behind that other than variance.

If we set a prevalence threshold, say 5%, that gives us a pretty good idea of the current meta at my level of play:  Carnage, Galahad, Griffin, Leo, Rhino, and Stalker. Especially Griffins, Griffins everywhere.

Here’s the same table, sorted by prevalence rather than bot name:



I think the most noteworthy takeaway here is that there aren’t really any noteworthy takeaways. I know things are wild for other people and I have no intention to delegitimize their frustrations with the game, but it’s pretty clear from the data that the matchmaker is doing a fine job of providing me with quality matchups, and with the “tanker league” being turned on, even those pests and parasites are becoming increasingly uncommon.

Since I took this snapshot early and am only reporting on it now, you won’t have long to wait before another By the Numbers update. I’m genuinely curious as to how the past couple of weeks have been, and I’d like to take a few more samples. My prediction is that we’ll find a steady course, but in the end there’s really only one way to find out.

Thanks for reading.

4 comments on “By the Numbers: Striking Gold

  1. Where’s the word from a sponsor?
    I’m assuming there is supposed to be a video there?


  2. I’m glad to see that your results show some quality and decent match ups being made overall by the current MM. I feel that it does a fine job when soloing and squadding within your same leagues.
    I think many frustrations with the game which are voiced out publicly happen to be due to the tendency for people to focus on the negative. For every 5 great matches that someone has, they will forget it all and remember that one bad match.


  3. Pingback: This week in community. 21.04.17 – War Robots

  4. Thanks for the data. I would suggest that the prevelance of both energy and physical shields is causing the decrease in damage as damaging a shield does not contribute to damage tallies after-match. I would love to see the data on the weapons you are seeing such as ancile, splash, plasma, projectile, mid range, long range, etc. Keep up the great work Dredd, and see ya on the forums.


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: