Guide

Game blockstring and pressure systems explained

Harbor Brawl's v1 pressure loop was simple: every blocked normal chipped guard, and long strings pushed defenders to the corner with no readable reset points. Skilled players mashed reversal on the third hit and escaped 41% of corner sequences. New players, meanwhile, ate full strings without understanding why their jab never came out. The problem was not raw damage — it was that designers treated “blockstring” as “any sequence that connects on block.” A real blockstring is a chain where each link leaves the attacker at frame advantage after blockstun, so the defender cannot act between hits unless the string ends, resets, or baits a response. The refactor separated true blockstrings from frame traps, tuned per-hit pushback so corner carry was intentional, and added two-character reset windows where offense could shimmy or throw. Corner escape rate on blockable pressure dropped 22%, and low-rank matches showed fewer “random mash” reversals because gaps were visually and audibly telegraphed.

Block pressure is the connective tissue between frame data, spacing, and mixups. This guide covers what makes a string a true blockstring, how plus-on-block math chains across normals and specials, pushback and corner geometry, reset and re-engage patterns, tick throws and option selects at string endpoints, pairing with chip and guard damage, the Harbor Brawl pressure refactor, a technique decision table, pitfalls, and a production checklist for designers and combat programmers.

What a blockstring is (and is not)

A blockstring is an offensive sequence where every hit that the defender blocks still leaves the attacker with positive or neutral frame advantage before the next action can begin. If any link is minus on block, the string is not a true blockstring at that point — it is a frame trap baiting a jab or a pressure sequence that ends with a gap the defender can exploit.

Frame advantage on block is computed as:

advantage = blockstun_frames − (recovery_frames − active_frames_overlap)

Designers publish this as “+2 on block” or “−4 on block” in movelists. A three-hit string where links are +1, +2, and +1 is a stable blockstring: the defender remains in blockstun with no actionable frames between hits. A string ending −2 on block is not a blockstring at the end — it is a reset point where a jab, reversal, or backdash can beat the attacker's next move if they blindly continue pressure.

Blockstring vs combo on hit

On hit, cancel routes and hitstun often allow longer chains than on block. Many games use block-only cancel restrictions: certain specials link on hit but not on block, or normals chain on hit but only single-link on block. Document both routes separately in your combat bible; playtesters will otherwise assume a hit combo should always work on block.

Building pressure: chains, cancels and spacing

Normal chains and target combos

Target combos (fixed A → B → C routes) are the easiest blockstrings to balance because you control every link's advantage and pushback in one table. Loose normal chains (any light → any medium that links) need global rules: maximum links on block, minimum advantage per link, and per-normal pushback caps so one outlier move does not carry to corner in two hits.

Special cancels into pressure

Special moves that are plus on block extend pressure and add high/low mixup layers. A classic pattern: blockstring ends with a plus special, then overhead or low before the advantage expires. If the special is only plus when spaced at max range, you create a footsies reward for precise players rather than autopilot corner carry.

Pushback curves

Each blocked hit applies block pushback to both characters (often asymmetric). Summed pushback across a string determines corner arrival. Designers tune three knobs per move: defender pushback, attacker pushback, and whether pushback scales on repeat hits in the same string. Without scaling, identical loops trap defenders in pixel-perfect re-stand ranges that feel infinite. With decay, the third repeat pushes less, keeping mid-screen footsies viable.

Corner geometry

At corner, defender pushback toward the wall is clipped — attacker often gains effective advantage because spacing tightens. Explicitly test corner blockstrings: some links that are +1 mid-screen become functionally +3 at corner due to reduced escape space and backdash distance. Nerf corner-only advantage by shortening blockstun on wall-adjacent hits or adding a “corner pushback floor” so defenders retain a sliver of room.

Reset points, shimmies and tick throws

Infinite blockstrings are oppressive. Healthy pressure includes reset points — intentional gaps where offense stops advancing and re-starts with a new threat.

Shimmy and walk-back resets

After a minus or neutral-ending string, the attacker steps backward to bait a jab, then whiff-punishes with a faster button. Implementation: allow walk-back during recovery cancel windows; ensure the baited jab whiffs at the spacing your reset expects. If walk speed is too high, shimmies become escapes instead of baits.

Tick throws

A tick throw is a throw attempt timed so it connects just as blockstun ends, beating delayed buttons but losing to jump or fast reversal. Tick setups require: (1) a move that is at least +1 on block, (2) a throw with startup ≤ defender's fastest post-stun option after accounting for input buffer, and (3) clear counterplay (jump, backdash, invincible reversal). Publish tick throw windows in your frame debugger; players should see the one-frame throw layer as skill expression, not invisible grab spam.

Option selects at string end

Advanced offense layers safe jumps, delay buttons, and throw on the same knockdown or string end. For designers, cap OS strength by making at least one option clearly lose to a disciplined response (e.g. safe jump loses to anti-air, throw loses to jump). Rollback netcode must serialize throw and strike on the same frame identically; see our grab systems guide for priority rules.

Chip, guard break and pressure pacing

Blockstrings interact with chip damage and guard break meters. Chip on every link accelerates guard erosion but can make defense feel hopeless; chip only on string end rewards completing pressure but allows mid-string escapes. A useful pattern: chip on specials, none on lights, guard damage accumulates only after three blocked hits in one string. Pacing matters more than raw numbers — testers should know when they are one hit from break versus when they have guard to spend on a reversal attempt.

Harbor Brawl pressure refactor

Harbor Brawl's block pressure pass targeted three failure modes: unclear string endpoints, uniform pushback, and chip on every normal.

String taxonomy — tagged every route as BLOCKSTRING, FRAME_TRAP, or MIX_ENDER in the combat data sheet; movelist UI shows icons on block.
Pushback tables — per-link defender pushback with 15% decay on repeats; corner clip uses a 12px minimum gap so backdash always exits.
Reset windows — two-character-specific resets after medium → heavy routes (−1 on block, shimmy bait frame 8–14).
Tick throw budget — throws at +1 require 6f startup after blockstun; +2 allows 7f tick; jump escape tested at 60fps and 120fps.
Chip policy — chip moved to specials only; lights deal zero chip but keep +2 advantage for true blockstrings.
Debug overlay — training mode shows advantage countdown and pushback vectors per blocked hit.

Corner escape rate fell from 41% to 19% among Diamond players; Bronze players reported pressure felt “readable” in surveys because string-end flashes and audio cues fired on minus links.

Technique decision table

Your situation	Prefer	Avoid
Corner carry without infinite loops	Pushback decay + explicit reset points	Uniform pushback on every link forever
Teaching new players defense	Clear minus-end visuals and single reset bait	Full-screen chip on every blocked normal
High-level mixup layer	Plus special into high/low with published advantage	Hidden frame traps with no counterplay
Throw-centric character	Documented tick windows with jump escape	Zero-startup throws after blockstun
Rollback online fighter	Deterministic same-frame throw/strike priority	String logic that differs client-side
Guard break pacing	Escalating chip on string completion	Guard break from one blocked light loop

Common pitfalls

Calling every blocked combo a blockstring. Minus links are frame traps or pressure enders, not strings.
Ignoring pushback at corner. Advantage numbers mid-screen lie; test wall-adjacent spacing separately.
Tick throws without jump escape. Defense must have a universal answer or offense becomes degenerate.
Same chip on lights and heavies. Light blockstrings become guard-break machines with no risk.
No visual/audio minus cue. Players mash because they cannot tell when pressure ended.
Input buffer ignored in tick math. A 3f reversal with 5f buffer beats a 6f tick you thought was safe.
Asymmetric characters sharing one pushback table. Large defenders clip differently; per-archetype tuning is often required.

Production checklist

Document frame advantage on block for every link in offensive routes.
Tag routes as blockstring, frame trap, or mix ender in combat data.
Publish per-hit block pushback and corner clip behavior.
Apply pushback decay on repeated links in the same string.
Define reset points with explicit minus or neutral endings.
Validate tick throw windows against buffer + fastest defender options.
Test corner and mid-screen pressure as separate suites.
Pair chip and guard damage with string pacing, not flat per-hit values.
Ship training-mode overlays for advantage and pushback debug.
Verify rollback determinism for throw/strike same-frame cases.
Playtest low and high skill cohorts for escape rate and readability.
Update movelist when block-only cancel rules differ from hit routes.

Key takeaways

A true blockstring requires plus or neutral frame advantage on every blocked link — minus links are frame traps or endpoints, not strings.
Pushback curves and corner geometry determine whether pressure carries or resets — tune them independently from damage.
Reset points, shimmies, and tick throws give offense depth only when counterplay (jump, backdash, reversal) is real and tested.
Chip and guard break should pace with string structure, not flatten onto every blocked jab.
Harbor Brawl cut corner escapes 22% by tagging string types, decaying pushback, and telegraphing minus endings instead of universal chip.