Raw data stays private
Stage checks never return verified rows, complete recorded-team lists, exact winning powers, or exact evidence ratios.
Limited public results
Get a usable team without exposing the entire database.
Check whether a stage has data, select the heroes you actually own, and receive up to three recommendations. Raw records, complete team lists, exact winning powers, and exact evidence ratios stay private.
Stage coverage only
Check whether a stage has data
This section intentionally does not show teams or raw records.
Separate recommendation tool
Build a team from your available heroes
Known winning teams are checked first, then tested alternatives. A separate experimental option makes one utility-driven substitution when possible and asks you to submit the result.
Separate review queue
Submit a completed battle result
Battle submissions go only to Website Submissions and never enter recommendations automatically.
Trust and privacy rules
How recommendations are controlled
0.90+ is excluded
Wins at a 0.90 or higher player-to-enemy power ratio are excluded as trustworthy formation evidence.
Overpowered wins are excluded
Rows labeled overpowered, power advantage, or instant—and wins leaving the entire team alive—are treated as untrustworthy evidence.
Evidence levels stay separate
Known wins, tested alternatives, and utility-based experiments are displayed in separate sections with visible source and trust ratings.
Missing stages can still work
When a stage has no data, the tool requests only the enemy lineup and power, stores that separately, and uses those details to search for analogous evidence.
Experiments request results
Experimental teams are never presented as known wins. The page asks players to submit the real result, and nothing affects verified recommendations until manually reviewed.