Clover RPG reroll odds: probability and budget guide

Define the outcome before entering a rate

Write exactly what you count as success: one named result, or any result from an acceptable set. Then identify the pool and the rate displayed for that outcome. A rarity-category rate is not necessarily the chance of one item inside that category. Do not treat a number copied from a conflicting community table as an official rate.

If several acceptable outcomes are mutually exclusive results of the same roll, their verified individual probabilities can be added. Do not add rates from different pools, or from overlapping categories. Getting a particular race and a particular magic through separate processes requires a separate model; the chance of either result is not the chance of obtaining both.

Calculate success by calculating no success first

Let p be the chance per attempt as a decimal and n the number of attempts. With independent trials and an unchanged rate, the chance of no success is (1 - p)^n. Subtract that from one to obtain the chance of at least one success: 1 - (1 - p)^n. Multiply by 100 to express the answer as a percentage.

The table uses a hypothetical 1% rate throughout. It is generated from the formula, not collected from live Clover RPG rolls. The last column is just as important as the success column: a budget can have a meaningful chance of producing nothing you wanted.

Scroll sideways to compare all columns →

A target chance is not a guaranteed result

For a target probability t between zero and one, and a per-roll rate p strictly between zero and one, calculate n = ceil(log(1 - t) / log(1 - p)). Rounding upward matters because an attempt is indivisible. With the same hypothetical 1% rate, reaching at least 50% requires 69 attempts; 90% requires 230; and 95% requires 299.

These are thresholds for a probability, not promises about an individual player. At a hypothetical 1%, the expected waiting time is 100 attempts, but a run of 100 still has a 36.60% chance of no success. The average waiting time and a near-guaranteed budget are different quantities.

Zero attempts give no chance of a success. A zero-rate outcome cannot be reached through repeated attempts. A 100% rate succeeds on the first attempt. With a rate strictly between those endpoints, no finite number of independent attempts guarantees 100% success.

Turn the calculation into a stopping rule

Start with the resources you are willing to use, not with the number a desired confidence level demands. Check the current displayed cost and reset consequences, then determine how many attempts that allowance covers. Enter those attempts and a verified rate into the probability tool to see the risk within your limit.

Suppose your limit is 50 attempts and your verified input were 1%. The model would give a 39.50% chance of success, with 60.50% risk of no success. Raising the target box to 90% does not change the odds of those 50 attempts; it reveals that the required 230 attempts exceed the limit. You can accept the smaller chance, choose a broader acceptable outcome set or keep your current setup.

Under the independent model, earlier misses do not raise the chance of the next attempt. Do not increase a preset limit solely because you feel a result is due. Stop at the limit or when your acceptable result appears, unless you deliberately decide on a new plan using current information.

Know when this calculator is the wrong model

Pity, guaranteed milestones, a rate that rises after failures, outcomes removed from the pool, event modifiers and conditional eligibility can all change the calculation. The independent model does not infer those rules and this guide does not confirm that Clover RPG uses any of them. If the interface documents such a rule, calculate from that rule instead.

Record the pool, exact desired outcome, source and date of the rate, accepted alternatives, attempt limit and any stated guarantee. The tool uses your input; it cannot validate a community probability table, predict your next result or simulate a character build.

The model follows the fixed-probability binomial framework. Public game references were reviewed October 25, 2026 without a live reroll experiment. Keep an explicit resource limit and use the output as a risk estimate; a hypothetical example is not a reason to reroll.