Dustbowl, right now

Discussion in 'Tutorials & Resources' started by Pinsplash, Dec 21, 2019.

  1. Pinsplash

    Pinsplash bad bad bad

    Positive Ratings:
    I played Dustbowl recently. 100 matches, 250 stages, and 490 contested points later, I'm not surprised to find that most people I've seen make statements on Dustbowl are wrong. I'm sure everyone's heard somebody say it's too easy to defend, or it's impossible to defend the second point of a stage after losing the first (rather contradictory statements, huh?) Toss those thoughts in the trash! You're about to see what's, to my knowledge, the only organized set of data on an official map to be seen in over five years, and the most detailed data we've ever had on one, aside from Mangycarface's paper on Frontier.

    Blu has the advantage on every point, as they should. Here's how often they cap each point: (94%, 90%, 97%, 78%, 96%, 74%)
    Now, if you're a less-experienced mapper you may be thinking "clearly Dustbowl is blu-sided, then!" Nope! Dustbowl is balanced, with a 47-53 split in Red's favor. How does this work when Blu seemingly always has the advantage? You see Billy, the chances for Red add up over time. Here's the same data, presented in a more... progressive manner (6%, 15%, 17%, 35%, 37%, 53%):
    To put it simply, this chart shows how often red wins or has already won the game, at the time a given point is contested. This chart also tells us that the notion of points being won together (due to red having little recovery time after losing the first point in a stage) is false. Blu nearly always gets the first point in a stage, then struggles on the second. (Those chokes really do matter.) I can't help but theorize that this is what Valve aimed for, the first point acting as some sort of buffer. Speaking of stages, here's how often Blu wins an entire stage (shown in dark red along with first graph's data, 85%, 76%, 72%):
    Other stuff:
    • Average percent of map experienced: 89%
    • Average cap rate: 88%
    • All matches were played on NA Valve casual servers. I really doubt results are very different elsewhere.
    • I did not record matches with less than 20 players (A/D gameplay is too different in low numbers), ones with an obvious imbalance in numbers or skill, or ones where a too-large number of the player population changed over the course of the match. Back caps, though surprisingly rare, are included.
    • As I said, this data is a mere 100 rounds; point 1 has 100 samples, but point 6 only has 47. The margin of error (uhhhh idk how much) increases a bit for each point down the line. No I will not play 200 rounds of that gosh darn map.
    • The Dragon's Fury and Thermal Thruster work too dang well on this map! The open areas are quite easy to soar over, and the closed areas make the DF scary to fight.
    • Depending on Red's setup, a Blu pyro can make a big difference on Stage 2 Point 1 by hugging the right wall.
    • I did plan to record clock times and class compositions, but it was too difficult to.
    Here's my original spreadsheet.
    • Thanks Thanks x 8
    • Like Like x 3