This guide assumes you already know the basics - that attacks have wind-ups, that you punish recoveries, that holding a guard does not work - and is about what comes after. It is a progression, roughly in the order the skills become available, with a drill for each one.
Stage one: reliable punishing
The first genuine competence threshold is punishing recoveries consistently rather than occasionally. Most players can describe the idea long before they execute it reliably, because in the moment the urge to attack early is strong.
The specific failure is attacking during the wind-up rather than after the active window. It feels identical from the inside - you saw movement, you pressed - and the outcome is completely different.
Drill: play a run where you attack only after an attack has visibly missed. Not after the enemy moves; after it has swung and connected with nothing. Your damage will be low and your survival will be much higher than usual. Do this until the waiting stops feeling like restraint.
Stage two: range control
Once you can punish reliably, the next gain is causing the openings rather than waiting for them.
Most machine opponents attack when you enter their range. That makes their range a switch you control: step in to trigger the attack, step out so it misses, then punish. You are no longer reacting to their schedule; you are setting it.
This is more reliable than reaction because you know roughly when the attack is coming. It also means you spend most of the fight just outside their reach, which is the safest place on the field and where most good players actually live.
Drill: fight at maximum range for a whole run. Do not close for a better shot, do not chase a kill. Get used to how far each enemy type's reach actually extends, which is usually shorter than it looks.
Ad space reserved. See README.md to connect an ad unit.
Stage three: reading openings by type
Not all openings are equal, and treating them as though they are is what limits most intermediate players.
A light attack that missed leaves a short recovery - enough for one quick hit. A heavy attack that missed leaves a long one - enough for a heavy of your own. A deflect-induced stagger leaves the longest opening in most games and typically doubles your damage as well.
Using a heavy attack into a short opening is the expensive mistake here, because you are still recovering when the opponent is not. The opening you earned becomes an opening you gave.
Drill: for one run, name the opening before you take it. Literally say 'light' or 'heavy' to yourself before pressing. It sounds absurd and it forces the distinction into conscious processing, after which it starts happening automatically.
Stage four: crowd shape
Everything above is one-on-one. Against multiple opponents the dominant factor is not which enemy you fight but what shape the group is in.
A ring around you is unwinnable regardless of execution, because a front-facing defence covers one direction out of several. A line is a series of one-on-one fights, which you now know how to win.
Producing a line is a movement problem. Enemies move at different speeds, so a wide arc naturally sorts them by pace; straight lines and stationary fighting let them converge. The open space in the arena should be behind you rather than in front of you, and corners should be treated as hazards.
Drill: play a run where you never move in a straight line for more than a moment and never attack anything that is not at the front of the group. Your kill rate will drop and your survival time will usually rise substantially.
Stage five: economy
The last thing to develop is treating your resources as a budget across the whole run rather than as something that refills.
The practical rules are short. Never spend the last point of power - keep a floor at one so an escape is always available. Prefer deflects to boosts when both would work, because deflects cost nothing and produce an opening. Do not spend a boost to reach a kill; spend it to leave a bad position.
In games with a combo multiplier, this connects to scoring: a hit taken usually costs more than a hit landed gains, and that asymmetry increases as the multiplier climbs. Good play in these games therefore becomes more conservative as a run progresses, which is the opposite of how most people play.
Drill: play a run where you never let your power meter reach zero. Nothing else changes. Most players find they attack noticeably less and last noticeably longer.
What mastery actually looks like
It is worth describing the destination, because it is not what beginners expect.
A player who is genuinely good at robot combat games does not look fast. They look unhurried and slightly passive - spending most of the fight just out of range, attacking less often than seems reasonable, and moving almost constantly in shallow arcs. The hits they land look easy, because they are: every one is against something that cannot respond.
The visible skill is in what does not happen. They are rarely surrounded, rarely out of power at a bad moment and rarely hit by anything they did not choose to accept.
A note on plateaus
Most players plateau twice. The first plateau is at reliable punishing, and it breaks when you start controlling range rather than reacting. The second is at one-on-one competence, and it breaks when you start thinking about crowd shape instead of targets.
If you are stuck and cannot tell which, the diagnostic is where runs end. Dying to single enemies means the problem is timing or range. Dying while surrounded means the problem is movement, and no amount of aiming practice will fix it.
Stage six: resource denial
One stage beyond the progression above, and the point at which players stop thinking about their own resources and start thinking about the opponent's.
Most machine opponents have something they spend. A shielded machine's plate has a power pool. A rival with a defensive stance has a meter behind it. In both cases, attacks that appear to achieve nothing are in fact draining something.
That reframes a blocked hit. It is not a wasted attack; it is progress toward an opening you are creating rather than waiting for. Four blocked hits against a defensive opponent produce a broken guard and a second of complete vulnerability, which is a larger opening than anything the opponent will hand you voluntarily.
The judgement is knowing when to switch modes. Against an aggressive opponent, patience wins because they will create openings themselves. Against a defensive one, patience achieves nothing - they are perfectly happy to wait - and pressure is the only thing that progresses.
Common intermediate errors
Distinct from beginner errors, and harder to notice because each one is a correct habit applied in the wrong place.
The pattern is that every rule in this guide is a default rather than a law. The step from competent to good is mostly learning which default each situation suspends, and the way to learn that is to notice the runs where following a rule produced a bad outcome rather than dismissing them as variance.
| Error | Why it looks right | Why it is not |
|---|---|---|
| Always waiting for openings | Patience wins fights | Defensive opponents never create one |
| Always keeping maximum distance | Range is free defence | Turrets are stronger at range than up close |
| Always deflecting rather than boosting | Deflects cost nothing | Unblockable attacks ignore the plate entirely |
| Always attacking the staggered target | Open targets pay double | Not if reaching it walks you into the crowd |
| Always conserving the last power point | It is your escape | Not when the alternative is a guaranteed hit now |
Practising against a specific enemy
The most efficient drill available once you are past the basics, and almost nobody does it.
Identify the enemy type that is on screen most often when your runs end. In a game with a wave composition display this takes about three runs to establish, and the answer is usually the same type far more often than chance would explain.
Then play specifically to learn it. Ignore score. When that type appears, engage it deliberately and observe: how long is its telegraph relative to others, what is its exact range, does it keep distance or close, can it be baited, and what does its recovery look like.
Two or three runs of this usually removes the enemy from your list of problems permanently, which is a better return than any amount of general practice. It also tends to transfer: enemy types are built from shared components, so understanding one deeply makes reading the others noticeably faster.
Knowing when you have actually mastered something
Mastery is a vague word, so here is a concrete test for each stage in the progression above.
Punishing: you can play a whole run without a single attack that was not into a confirmed opening. Not mostly - entirely.
Range control: you can describe each enemy type's reach in terms of your own, and you position by it without thinking about it.
Opening types: you have not used a heavy attack into a short opening in several runs, and you notice immediately when you nearly do.
Crowd shape: you can say, at any moment, where the open space is and which direction you would leave in. If you have to look, you are not there yet.
Economy: your resource meter never reaches zero, and you stopped having to think about it some time ago.
The reason for concrete tests is that the alternative - feeling like you have got it - is unreliable in exactly the wrong direction. Most players feel they have mastered a stage several sessions before they have.
Key takeaways
- Punish only after an attack has visibly missed, not when the enemy moves - the difference feels identical and is not.
- Control range to cause openings rather than waiting for them; most of a good fight is spent just out of reach.
- Match the punish to the opening. A heavy attack into a short opening hands the opening back.
- Against crowds, shape matters more than targets - move in arcs so the group strings into a line.
- Keep a floor of one power point so an escape is always available, and prefer deflects to boosts.
- Where runs end tells you what to fix: single enemies mean timing, being surrounded means movement.
- Against defensive opponents, blocked hits drain their resource - pressure creates the opening patience never will.
- Every rule in combat is a default; the step to good play is learning which situation suspends which.
- Identify the enemy type present when your runs end, and spend two runs learning only that.
- Understanding one enemy deeply transfers, because the types share components.
- Each stage has a concrete test: no unconfirmed attacks, reach described in your own terms, no heavies into short openings, open space known without looking, meter never at zero.
- Feeling like you have mastered a stage reliably precedes actually having done so.
Frequently asked questions
How do I stop attacking too early?
Practise attacking only after an attack has visibly missed - not when the enemy moves. Do it for whole runs until waiting stops feeling like restraint.
What is range control?
Deliberately stepping into and back out of an enemy's range to trigger an attack you know is coming, then punishing the recovery. It is more reliable than reacting.
When should I use a heavy attack?
Only into a long opening - a missed heavy, a stagger, or a completed charge. Into a short opening it leaves you recovering while the opponent is not.
Why do I die when several enemies are on screen?
Because the group has closed into a ring. Move in wide arcs so faster enemies pull ahead, which turns the ring into a line you can fight one at a time.
How much power should I keep in reserve?
One point, always. A boost you cannot afford is the most common cause of taking a hit you could see coming.
Why does good play look passive?
Because the hits are all against targets that cannot respond, which means most of the fight is spent waiting just outside range rather than exchanging.
Is a blocked hit wasted?
Not against an opponent with a defensive resource. Each blocked hit drains it, and a broken guard is a larger opening than anything they will give you voluntarily.
When should I stop waiting for an opening?
Against an opponent content to defend. Patience only wins against something that will eventually commit to an attack of its own.
How do I know when I have mastered a stage?
Use a concrete test rather than a feeling - for punishing, a whole run with no attack that was not into a confirmed opening.