Expand Your Marvel Rivals Hero Pool: Find Your Next Hero· Season 8

This tool is built around the coach consensus for ranked play: depth in your main role first. A primary plus a same-role secondary for bans and counter-swap, then a third same-role pick at higher ranks. A cross-role flex is a later addition, not a substitute for that depth. The score below ranks every hero against your existing pool so each addition lifts your effective win rate, not just your hero count.

Each candidate is scored on a composite Hero Pool Fit Scoredesigned for players who want to add a hero to an existing pool rather than start fresh. The score combines five signals, each normalised against the site's tier thresholds (0 ≈ F-tier WR, 1 ≈ S-tier WR):

• Win rate (20%). Bayesian-shrunk standalone win rate. Drives the ban-substitute and off-role flex framing: when one of your pool heroes is removed, you want a hero that wins games on their own merit. Boosted 15% when your pool has only one hero in this role (scarcity).
• Pair value (20%). Mean pair win rate across your pool, in one of two flavours (switchable above the recommendations):
  • Best to learn subtracts each hero's standalone strength to isolate how their abilities actually combo with yours, independent of whoever's currently overtuned. The signal stays valuable across balance patches.
  • Best to winuses raw pair win rate this season: what's winning games today. Better for climbing right now, sometimes worse when the balance rotates.
• Counter cover (30%). For each opponent that currently beats your pool, how much does the candidate flip that matchup? Pickrate-weighted so frequent threats matter more than rare ones.
• Map cover (20%). Candidate's mean win rate on the maps where your pool's pooled win rate is below 50%. Missing maps are renormalised out rather than imputed. Falls back to your pool's weakest quartile when no map is below the threshold.
• Confidence (10%). Logistic function of total tracked matches. ~0.1 at 1k, ~0.5 at 10k, ~0.95 at 50k. Keeps small-sample heroes from leading a role.

We apply hard floors before scoring: a candidate needs at least 5,000 tracked matches (1,000 for Stadium). Counter and synergy rows below 200 matches are excluded: shrinkage alone doesn't suppress that noise. Heroes already in your pool are excluded from the candidate set, and honourable mentionssurface heroes that don't take the top spot but answer a specific question: who's your best ban substitute, who handles your worst matchups, and who shines on your weakest maps.

When your pool has all of its recommended slots filled in a role, the recommendation flips to a compact "you're covered" card. When your pool overfills a role beyond the recommended cap, we nudge you toward branching into an under-represented role instead. Practice time is finite and a fourth specialist is usually lower-value than a flex pick.

Data is pulled from thousands of tracked community matches and refreshed daily around 07:00 UTC. Last updated: May 20, 2026.