Mechabellum is like a PDE – A Belated Review from a Veteran Player Who Started Posting Gameplay Videos in 2021
I would call Mechabellum the most brilliant canvas in PVP strategy games. Both its success and its failure should be attributed to its creator, Bearlike.
This part is my personal gaming history. If you're only interested in the game review, please skip to the next section.
My in-game ID is GTBOB. I first learned about Mechabellum (back when it was still called Steel Battle Zone) by watching Bearlike's livestreams on Bilibili. Later, when I had both the time and a device capable of running it, I quickly became a hardcore player. At the time, I was an undergraduate student, single, naturally blessed with plenty of time and energy to pour into gaming.
I often played Mechabellum late into the night until there were no other players left to match with. That was right around when the Community Challenge mode first came out. So, I would spend my nights alone playing those challenges—repeating the tough ones dozens of times, searching for different tactics and strategies.
Back then, the level cap was still three, the supply limit was around 1200, and there were no unlockable active skills—only range and speed upgrades. It was during that period that I honed many of my tactical details, like combining airdropped shields with flanking maneuvers to gain small but meaningful advantages. More importantly, I came to deeply understand that the true win condition of the game is:
Empty the opponent's points before your own reach zero.
Not winning a single round. Not achieving some perfect formation or maxing out a unit's tech.
At that time, unit recycling appeared only as a random option between rounds, and there was no tactical beacon. So I understood deeply that every unit placement should be deliberate—a position reserved in round one might end up playing a critical role in round seven.
As the player base grew, so did my skills. Eventually, I started winning some weekend tournaments, and even took a few tournament championships (some of those matches are still on my Bilibili channel—I revisit them from time to time, and the joy I felt then is unforgettable).
Later, the economy system changed. The supply cap increased significantly, and since the map size stayed the same, high-damage units like the Stormcaller gradually became more viable. I shifted from using Stormcallers occasionally to making them a core part of my game plan.
Then I realized something: the BO1 format of the tournaments introduced randomness that sometimes exceeded my ability to predict. For example, in a match where a mid-sized unit tempo strategy faced off against a giant-unit health-sacrifice strategy, Parasitic Ammo Vulcan could single-handedly dismantle the entire game plan of the tempo player.
As a result, I started playing less and less. By the time ranked matchmaking was introduced, I had already stopped participating in regular tournaments. Out of curiosity, I played two placement matches, losing one. Later, while watching someone else's stream, I had a sudden flash of insight about the current meta—then went on to win my next eight placement matches in a row. Thanks to the high-scoring design of placements, I achieved the highest MMR of my life: around 2400+.
After that, I barely played seriously again. By the end of the Pioneer Season, my MMR had settled at 2327. But because almost all of my matches were from the very early days, my Pioneer Season match count was essentially zero.
Later, in Season 1, I tried to return. Thanks to friends who still remembered me, I even participated in a tournament using a 0-game, 2100-MMR account. I played four matches, winning some and losing some.
Out of curiosity, I joined a QQ group called "Mechabellum Balance Group." I only applied because my MMR happened to meet the requirement—I was just there to hang out. Since the group had a minimum MMR cutoff, there weren't many members. What shocked me was this: even though the group was far from full, I was kicked out simply because my account lacked enough games played (after those four tournament matches, I hadn't queued again). It made sense by their logic, but it still felt unnecessary.
From Seasons 2 through 6, I played a total of 12 ranked 1v1 matches. My MMR had dropped all the way down to 1100.
Now, about the game itself.
Bearlike has said more than once during his livestreams that his goal with Mechabellum was to make a real-time strategy game that required thinking but didn't require hand speed. That idea is genius. The dual-tower design, and the turn-based deduction of HP based on units, gives the game its foundational mechanics. Brilliant gameplay!
Back then, almost every ability (except the starting ones) had to be chosen from the same reinforcement pool available to both players. There were no permanent recycling options, no permanent repositioning beacons. The entire game was filled with the beauty of pure, raw mind games.
I would use PDEs to describe the game during that stage.
Just making players understand that units follow a targeting logic combining both distance and direction, and fight by standing still and firing—that alone is extremely difficult. It's like trying to get started with PDEs.
An entire multi-round match feels like a complete PDE system. The battlefield state is extremely sensitive to initial conditions—just like the nonlinear behavior of PDEs, where a tiny change in initial values can completely alter the solution.
The feedback in Mechabellum is indirect. You can be winning, and then suddenly you're losing, without a clear trigger. That indirectness also reminds me of PDEs.
In most strategy games, replays are clear: you did X in round Y, and it led to result Z. But in Mechabellum:
The impact of decisions accumulates across rounds. A unit you picked in round one might slowly gain experience and become the deciding factor in round seven—but at the time you picked it, you might not have thought that far ahead.
Chains of causality are non-intuitive. The effect of a certain technology choice might only become visible three rounds later, mixed in with dozens of your opponent's decisions, making it hard to isolate.
The "no takebacks" mechanism: once a unit is deployed, it cannot be moved. So each round's decisions are like placing stones in Go—everything that follows is determined by the resulting global state, not any single move.
This is extremely similar to the difficulty structure of PDEs:
A small, localized change, after multiple rounds of iteration and non-linear interactions (unit counters), eventually produces a global, hard-to-trace macroscopic outcome.
There's a mode called Community Challenge—it lets you take over a match in its final round and see if you can turn the tide.
That mode is essentially taking a complex PDE problem and turning it into a simplified version with known initial conditions and fixed boundaries. You don't need to reconstruct the entire causal chain—you just need to find the optimal solution given the current battlefield state.
But a real match has no such simplification. You have to build causality from the chaos. You have to make decisions under uncertainty.
Later, the economy changed. Higher-value units naturally became stronger. More and more abilities became permanent, drastically reducing randomness within matches.
I would call this stage of the game an ODE game.
The same underlying mechanics remain, but every match now shares a large set of fixed conditions. These conditions simplify the game so much that it's gone from a PDE to an ODE.
A game that's like an ODE can still be good. The real problem with Mechabellum lies in its creator's arrogance.
Bearlike is equally arrogant toward every single player.
