Strategy interface research and review rubric

Research pass: 2026-09-08. This is interface guidance for the local prototype branch, not a change to game rules. Sources were checked on this date. Recommendations below are applications and design judgments, not claims that another game's rules fit this game.

Evidence and reference patterns

• Civilization VI's official manual presents an active-unit combat preview containing modified strengths, probable outcome and expected damage. Apply that pattern as a persistent, explicit pre-commit card for the selected action and target. Our stochastic system needs exact hit probability distinguished from estimated damage; melee must show successful opportunities and legal responses instead of promising a deterministic outcome. Civilization VI manual, Combat Preview.
• Subset Games describes telegraphed enemy attacks as central to Into the Breach's planning. Apply this by making known movement paths, reaction locations and threatened areas readable before committing. Do not invent enemy intent or expose private roster choices. Into the Breach official overview.
• Xbox's motion guidance recommends control over scrolling, blinking, auto-updating content and camera motion. Apply it through brief optional movement playback, a stable camera, reduced motion and persistent text results. Do not require players to read a disappearing toast or wait for a cosmetic effect to learn a rule outcome. XAG 117: Visual distractions and motion settings.
• Xbox's text guidance applies readability settings across HUD, prompts and secondary windows. Apply existing enlarged text consistently to combat numbers, dice labels and inspector explanations. XAG 101: Text display.
• Xbox's contrast guidance calls for sufficient text and visual contrast, and alternatives to color alone. Team A/B labels, active/inspected outlines, watch hatching, reaction symbols and destroyed-part text should reinforce colors. Disabled action text still needs to remain readable. XAG 102: Contrast.
• Xbox's input guidance calls for alternative input methods and avoidance of unnecessary timing or simultaneous-input demands. Action explanations and public unit information must work by click/tap and keyboard, not only native title tooltips. XAG 107: Input.

Observed baseline

Inspected current source and Chromium at 1440 × 1000, opening the adjacent two-handed melee trial. The board now has a stable sized container, which should be retained. The current interface still places nearly every action in one narrow scrolling column. Disabled actions rely heavily on mouse-only title explanations. Score sits below those actions. Site navigation consumes 216 horizontal pixels during combat. The active identity repeats the pilot name. On the full-map view, the two adjacent units are small and six floating body labels can clip against the map edge. These observations motivate the following order of work.

Priority 1: orientation and action clarity

1. Give battle its own focused workspace, with a clear route back to packet navigation and tools. Keep the map's position and size stable when a preview, reaction, melee stage or error appears.
2. Keep a compact score/round/actor strip visible. Say whose decision it is now, including temporary reaction and melee control. Show remaining movement and Quick/Full budget without implying that all three are independently available.
3. Separate Active mech from Inspecting. Clicking any public unit should produce a pinned unit dossier; inspecting an enemy must never suggest its commands are being issued. Show a clear return-to-active control. During activation selection, the explicit eligible-unit choice remains authoritative.
4. Group commands into a small number of readable categories. Always expose action cost, distinguish an armed action, and provide a tap-accessible explanation for unavailable actions. Keep End activation in a predictable position. Do not make the player hunt through tools or reports to finish a turn.
5. Reserve one stable decision area for intent, target, cost, risk, confirmation and cancellation. Replace its content when the interaction changes. Do not append a succession of panels that pushes the next required button offscreen.
6. Preserve essential public state: damaged/disabled body parts, Heat/cap, statuses, reaction availability and activation spent/available. Compact board markers plus a richer pinned dossier are easier to read than six full labels on every unit.

Priority 2: feedback that explains the rules

• Preview the actual planned path with numbered or otherwise distinct waypoints, total movement cost, climb Heat and deliberate-fall warning. Distinguish legal movement from a geometric range overlay.
• After acceptance, animate along committed per-hex movement events. Do not interpolate through obstacles along a straight line from origin to destination. If reactions stop movement, stop at the actual triggering hex. Cosmetic playback must not mutate state, consume RNG or reorder events.
• Preserve camera position during normal choices and animation. An explicit focus/fit control can frame combatants. Do not recenter on hover. Provide a user preference to skip motion and honor reduced motion.
• Show a persistent last-resolution card: actor/action/target, hit or miss, actual part durability lost, disabled part, Heat gained and resulting status. Detailed dice/modifier/armor steps belong in an expandable explanation linked to the accepted event.
• Fire previews must distinguish exact hit chance from estimated damage. Aimed location, cover, range, weapon/load and Overdrive should be visible without opening raw JSON.
• In melee, show both rolled pools, success threshold, available versus spent opportunities, source weapon, selected target part and current decision owner. Parry selection must visibly explain the opportunity count required; successful attack-capable dice stay eligible for parry. Do not introduce a second roll or consume the defender's normal activation.
• In reaction windows, show the watching/prepared reactor, target, reason, firing point, weapon and available reaction budget. Passing should explain that the current offered movement is being declined while preparation/budget remain available under the existing rules.

Priority 3: analyst depth without constant clutter

A normal play view should answer what to do next, what can happen and what just happened. A separate accessible analysis view can expose the event log, raw dice and modifier accounting, complete equipment/pilot effects, terrain/LOS details, accepted decision index, seed and rules identity, bookmarks, replay, reports and saves. Keep these available in one or two deliberate interactions rather than removing them for cosmetic simplicity. Opening a dossier or log must not move the board or discard the pending action.

Independent review rubric

Review honestly against actual interaction evidence. Target at least 90/100 with no major interaction defect; a passing number is not a reason to waive a defect. Record viewport, scenario and evidence. A defect causing an accidental action, inaccessible legal response, incorrect rule explanation, lost state or unstable board is a release blocker regardless of total.

Minimum review flows: ordinary move/fire/end; pending melee strike/parry; movement reaction choose/pass; inspect an inactive ally and an enemy while an action is armed; save/reload pending choice; reduced motion; desktop 1440 × 1000 and tablet landscape 1180 × 820. Re-run board alignment and layout stability regressions after animation or workspace changes. Add automated assertions for actual path playback and keyboard/touch access where they protect a real behavior rather than implementation structure.