Interpreting Lab Results and Limitations

Read the input summary before the recommendation. A result is useful only when the active profile, current levels, inventory, formation, roles, and selected mode match the intended scenario.

What results can show

Depending on the module, output can include an upgrade order, target checkpoints, resource consumption, remaining inventory, bottlenecks, expected damage, uncertainty range, comparison, sensitivity, or repeated-run distribution. These values compare supported choices under one model.

What results cannot guarantee

The Lab cannot predict player execution, connection conditions, undocumented game changes, every interaction, future prices, or live opponent behavior. Beta modules and data revisions can change an output even when saved inputs did not change. A narrow uncertainty range describes the model, not certainty about the game.

Reading the Cyber interface and visual indicators

The Cyber interface uses standardized visual markers across all optimizers and simulators:

• Semantic troop styling: Results tag troop-specific gains with dedicated color bars: Infantry (cyan/blue), Cavalry (orange/amber), and Archer (green/emerald).
• Combat impact ratings: Steps in Ascension Path classify returns into distinct impact tiers (Very High, High, Medium, Low, and Marginal), allowing quick identification of major milestones.
• Breakpoints: Highlighted icons identify steps that unlock set bonuses, pass mastery gates, or complete red equipment imbuements.
• Comparative trade-offs: Alternative comparison panels explain not only why the winning choice succeeded, but why the competing alternative lost.
• Asset resolution status: Genuine game art displays when mapped. If an image is pending or loading, a semantic SVG category icon takes its place to prevent layout shifts.

When a result looks wrong

1. Confirm the correct profile and save state.
2. Recheck every starting level and owned resource.
3. Verify Infantry, Cavalry, and Archer formation totals and capacity.
4. Confirm captain versus joiner roles, hero skills, widgets, and combat context.
5. Check whether current equipment is already included in the entered stats.
6. Compare one controlled change at a time.
7. Rerun only after the input summary matches the intended scenario.

If profile edits did not appear, save them, reopen the module, and select the profile again. A previously saved run retains its own input snapshot and does not silently become a result for the newest profile state.

Lab result review landmarks

Compare input summary, result, and assumptions as one record.

• Tool name, module availability or beta label, active profile, and saved-input timestamp.
• Complete input summary with mode, objective, formation, roles, equipment, and inventory.
• Main recommendation or modeled result with comparison, range, or resource detail.
• Assumptions, limitations, data version, save or share action, and rerun controls.

When reporting a problem, provide the tool, public input summary, module version information shown, and unexpected output. Remove account credentials and private community data.

Review workflow and worked interpretation

First confirm the active profile and saved state. Then read input assumptions, validation warnings, model and dataset version, ordered steps or calculation stages, before and after values, consumed and leftover resources, evidence labels, and stated limitations. Roles and scope do not change the meaning of a personal scenario; shared screenshots should remove private account or community information.

Worked example: A plan ends with a large resource balance. The user opens the ordered steps and sees another required resource reached zero. Every remaining candidate either needs that missing resource, is locked, is maximum, or has non-positive weighted gain. The correct output interpretation is a constrained stop, not a promise that the leftover resource has no game value. Correct a wrong inventory field and rerun; otherwise preserve the leftovers in the decision.

Modeled output cannot guarantee live damage, future game balance, globally optimal progression, or accuracy of unentered mechanics. Troubleshoot by changing one assumption in a copied scenario. Report module, scenario, versions, input summary, warning, expected branch, observed result, and steps already checked.