Upgrades convert account currency into permanent stats. Community system pages have documented categories including Luck, Roll Speed, Potion Duration, Border Chance, Boss Chance, and Prestige-related progress. Exact names, caps, prices, and reset options are version-sensitive; inspect the current Upgrades interface.
Check the live menu
Record each available stat, current level, next effect, next cost, cap, currency, and any reset information. Do not copy an old maximum or price. Check whether the displayed increase solves a current problem.
Preserve enough currency for required unlocks or recipes before buying.
Prioritize by bottleneck
Choose Luck when future rolling quality is the limiting factor. Choose Roll Speed when active time, not supply, is the problem. Potion Duration matters only when timed buffs are used deliberately. Border or boss-related stats matter when that specific outcome is your goal.
Combat walls may require cards, items, traits, or mode progress rather than a rolling stat. Do not buy an unrelated upgrade merely because it is cheap.
Compare marginal value
Look at what the next level changes, not the total name of the stat. Compare cost with a Story unlock, tower checkpoint, craft, or another upgrade. Early affordable levels can be efficient, while later expensive levels may wait.
Buy one planned level and verify that the displayed stat updates before purchasing more.
Coordinate with rolling sessions
Claim reachable Luck Index progress and planned upgrades before activating potions or spending Instant Rolls. Select the pack and budget afterward. Permanent improvements affect more future sessions than a single timed effect.
Do not delay every roll for a far-away cap. Use a practical threshold based on current access.
Handle reset options carefully
If the live menu offers a reset, read cost, refund behavior, and confirmation. Community documentation can be outdated. Take a record of the current allocation before an irreversible change.
Reset only for a defined new plan, not to experiment with no benchmark.
Recheck after updates
Open the menu after a major patch and compare category names, prices, caps, and effects. Update personal notes with a date. If a stat behaves differently, establish a fresh baseline before spending more.
Use Luck Index for content-linked permanent gains, Potions for temporary effects, and Farming to earn the exact upgrade currency.
A purchase checklist
Name the bottleneck, choose the matching stat, confirm the next effect and price, protect other required currency, buy once, and verify the result. This keeps permanent spending intentional and makes later progression decisions easier to explain.
Use an upgrade budget
Divide the current currency into the planned purchase, a protected reserve, and any amount committed to another shop or unlock. Confirm that the upgrade menu actually uses the same currency before counting it. Buy the smallest meaningful step first and observe the new value.
If several cheap levels are available, recalculate after each rather than pressing through the entire balance.
Benchmark practical effects
For Roll Speed, time a comparable normal batch before and after. For Potion Duration, read the active timer during a prepared session. For boss- or border-related effects, use the display and enough observations to guide personal planning without claiming an official rate. For Luck, confirm the stat changes and combine it with Card Index or Luck Index progress.
Keep potions, Moon state, and selected pack constant when comparing a rolling stat.
Avoid false efficiency
A low price is not valuable when the stat does not affect the current goal. An expensive level may be justified if it unlocks a broad permanent improvement, but compare it with a card, craft, or first clear that solves the wall sooner. Write the expected outcome before purchasing.
If no outcome can be named, hold the currency until the next unlock or update clarifies the choice.
Recover from an outdated build
When an update changes categories or resets a value, record the current allocation and live reset terms. Rebuild around the present bottleneck. Do not copy a pre-update distribution with different caps or costs.
After changing the allocation, repeat one known benchmark. If the result does not improve, preserve the evidence and avoid another reset until the source of the problem is identified.