How to Use the Conquest Recommender

What the Conquest recommender does

The recommender searches the reviewed dataset for teams that fit both the selected stage and the heroes available on your account. It deliberately returns a limited number of useful teams instead of exposing every recorded lineup.

Enter the correct stage

Use the exact Conquest stage number. When exact-stage data exists, the tool checks known wins first. When the stage is missing, it can use the enemy formation and power to search for similar evidence without publishing the private source rows.

Select the heroes you can actually use

Select every developed hero you would realistically consider, not only the five heroes currently deployed. A broader roster gives the tool room to preserve utility when one recorded hero is unavailable.

• Choose at least five heroes.
• Choose at least six to allow a distinct one-hero experiment.
• Do not select an undeveloped hero merely because it is newer.
• Include older specialists that still have usable gear and skills.

Read the three result types

Trust labels describe the available evidence, not the probability that your account will win.

Use experimental recommendations safely

An experiment is most useful when it changes one hero while preserving the rest of a proven structure. The goal is to test a clear idea—for example adding cleanse against control, anti-heal against sustain, or a backline hunter against a protected carry.

• Run the evidence-based option first when possible.
• Use the experimental option as a separate test.
• Do not assume a higher-generation substitute is automatically better.
• Submit the result even when it loses.

Evaluate your power separately

A formation can be correct while the account remains underpowered. The evaluator compares your player-to-enemy power relationship with trustworthy wins, but it intentionally treats near-equal-power wins as weak formation evidence.

Use the verdict as a range: likely, possible, or unlikely. Gear, skill levels, widget development, and battle variance can still change the result.

Submit the outcome

After testing, submit the exact positions, powers, result, heroes remaining, and any important context. New submissions enter manual review and never alter recommendations automatically.