Make every orb count.
Six summon formats. One clear picture.
Compare your odds of getting a banner character.
Chance of at least one banner character
| Format | Orbs spent | Chance ≥ 1 | Expected copies | This step |
|---|
Banner characters.
Without the filler.
“Banner” means the newly introduced characters you are chasing. “Featured” includes those characters plus older fillers. A 6% featured rate is not necessarily a 6% banner rate.
What does the chart measure?
Cumulative chance is the probability of obtaining at least one banner character by a step. Expected copies is the average total across many players, including duplicates. Chance on each step shows that 10-pull’s odds in isolation.
How are guarantees counted?
A guaranteed slot replaces one of the ten ordinary draws: nine ordinary draws plus one guaranteed draw. Tickets and medal exchanges add a copy when enabled, assuming you choose a banner character and redeem rewards immediately.
The math & simulation assumptions
Ordinary step: P(no banner) = (1 − p)¹⁰. Guaranteed step: P(no banner) = (1 − p)⁹ × (1 − q). Cumulative chance = 1 − the product of all step miss probabilities. Expected copies are summed draw probabilities plus redeemed rewards.
Here p is the combined new-character rate and q is the new-character probability in the guaranteed slot. A featured guarantee uses new characters ÷ featured pool size, assuming equal weights. General ★5 guarantees on modern banners are provisionally treated as retaining the ordinary banner probability because their full distribution is not confirmed.
Extensions are hypothetical: Crown repeats 10 steps; Zenith, Glory and Calamity repeat 20 steps, including rewards. Refined repeats its ordinary final step after 10. Classic stays flat. Lines connect completed purchases; they do not imply partial purchases are possible. Comparing by orbs shows each format’s own purchase milestones.
Rates, rules & sources
RESEARCHED 19 OCT 2026Transparent inputs for every line. Banner formats can change; check the in-game notice before using these comparisons.