Choose The Right Mode
There are two simulator tabs:
- Gacha Odds Calculator is for a fixed wish budget. Use it when you want the chance of reaching a limited 5-star constellation target with the wishes you have.
- Outcome Distribution is for target planning. Use it when you want to see how many wishes a target usually takes, including mean, median, percentiles, and the outcome chart.
Basic Banner Inputs
- Starting Wishes is only used by the Gacha Odds Calculator. Enter the pull budget you plan to spend.
- Current Constellation is your current limited character state before pulling. Use Not Owned when starting from no copy.
- Target Constellation is the goal you want to reach. The simulator converts this into the number of new limited copies required.
- 5-Star Pity is your current character banner pity before the next pull. Use 0 if you are at a fresh pity counter.
- 4-Star Pity is your current 4-star pity before the next pull. This is mostly optional and won't affect the outcome too much.
- Guaranteed? should be Yes if your previous 5-star on the character event banner was not the promotional character.
- Radiance Step is your current Capturing Radiance counter, from 0 to 3.
What does the result mean?
The Gacha Odds Calculator is useful to calculate how many pulls it will take you to get a character to a specific constellation in Genshin Impact. Specifically:
- Estimated Chance is the share of simulated budgets that reached the target.
- Avg New Limited Copies shows how many limited copies the budget gained on average.
- Avg Pulls Used is how many pulls were spent before stopping or reaching the budget.
- Avg Starglitter Pulls shows the average extra pulls from starglitter when a starglitter mode is enabled.
- The Limited 5-Star Goal Curve shows the cumulative chance to reach each constellation by a given pull count.
Running Outcome Distribution
The Outcome Distribution gives you a better visualization as to how many pulls it takes to achieve your pull target.
- Open Settings to choose the target. Graph Target can be Featured 5-Star or Specific On-banner 4-Star.
- For a 4-star target, choose which rate-up 4-star to graph and how many copies you want. This is especially useful to see how many pulls on average is needed to get a C6 4-star.
- Click Run Simulation to plot the outcomes. Histogram bins group results for readability, while Every dot plots individual runs (but will demand more from your PC)
- Mean Wishes is the average number of wishes spent. Median is the middle result.
- P90, P95, and P99 are better for planning bad luck. For example, P95 means 95% of simulated runs finished by that number of wishes.
- Radiance Used is the share of runs where Capturing Radiance affected at least one result.
- Refund Rate and Avg Starglitter estimate how much starglitter the run generates.
Custom Settings
There are a lot of custom settings and it can vastly affect the outcome.
- Runs controls sample size. More runs make the result more reliable but take longer.
- 5-Star Model controls base rate, soft pity start, soft pity curve, and hard pity.
- Featured Outcome controls the base 50/50 chance and Capturing Radiance behavior.
- 4-Star Model controls base 4-star rate, hard pity, featured chance, and current constellation for each rate-up 4-star.
- Refund Rate Mode changes how starglitter are used in the simulation.
- Chart Display controls whether the distribution uses histogram bins or every-dot mode, plus bin size.
Starglitter Modes
- Do Not Use means no starglitter will be used.
- Use Once spends only amount generated by the starting budget.
- Use Recursively essentially means to full send and keep spending all starglitters until there's not enough left for a pull.
The simulator uses 10 starglitter per 5-star, 5 for a banner 4-star copy that is already C6, and 2 for other 4-stars. You can change the constellation of the 4 star line up for even more refunds.
Reading Results Correctly
- Treat the output as an estimate, not a guarantee. A 90% chance still means 10% simulated outcomes fail.
- If you change pity, guarantee, Radiance, starglitter, or model settings, rerun the simulation before using the chart. Note that if you change anything related to the odds then the result might not reflect reality.
When To Reset
- Click Reset Defaults if you changed model settings and want to return to the default event-wish assumptions.
- Reset before comparing scenarios if you are unsure which settings are still active.
- Use Stop if a large run is taking too long, then lower Runs or switch back to histogram mode.