Combat power rose, damage did not: the formula omits damage
✍️ Alice・愛麗沙
Combat power is a display number, not a damage input — and that is provable from the expression, not a matter of taste. The version the Korean community records reads: CP = attack + defence/2 + HP/4 + 0.3 × sum of sub-stats + rune grade bonus. Look at what is missing. No skill level. No final damage. No gem tag damage. No crit multiplier. Not one multiplicative bucket. Two consequences follow directly. +1,000 HP gives you 250 combat power and zero damage. Pushing skill enhancement to 30 doubles the skill coefficient and moves combat power not at all. Use combat power to get through a dungeon’s door; use the damage chain to decide what to buy. They are two different rulers.
The weighted sum
| Number on the character sheet | Combat power per point | What the source gives | Strength of the basis |
|---|---|---|---|
| Attack | 1 | A single number | Medium |
| Defence | 0.5 | A single number | Medium |
| HP | 0.25 | A single number | High |
| Primary stat (the one your class scales on) | 0.6 ～ 1 | A band | Medium |
| Secondary stat (the one ranked next) | 0.2 ～ 0.4 | A band | Medium |
| The other three primary stats | 0 ～ 0.26 | A band | Low |
| Sub-stat values (crit hit, pursuit, and so on) | 0.3 ～ 0.6 | A band | Medium |
The combat-power display formula is a KR community measurement, not official; two measured substat weights (0.3 and 0.6) are shown side by side, never averaged.How far behind am I
How much combat power each point of a stat converts into is listed row by row above. Two sources disagree outright on the sub-stat and primary-stat weights — one writes 0.3, the other 0.6 — and this site prints both as a band and does not average them. Averaging to a midpoint nobody measured would only replace two honest numbers with one invented one.
What matters most about this expression is its list of absentees.
Four things that give combat power and no damage
HP. +1,000 HP is +250 combat power and +0 damage. It is useful — you have to survive to deal damage — but it appears in no damage multiplier anywhere.
Defence. +1,000 is +500 combat power, +0 damage.
The rune grade bonus. This is the insidious one. Combat power grants a flat score based on the rune’s grade and star count — Legendary 6-star 250, Epic 5-star 150, Elite 4-star 80, Rare 3-star 60, Rare 2-star 40, Advanced 1-star 20, Advanced 0-star 10 — entirely regardless of what that rune actually does. Swap a well-chosen Elite 4-star for a useless Legendary 6-star and combat power jumps +170 on the spot while your real damage may go down.
Sub-stats you cannot trigger. The 0.3-to-0.6 combat power per point arrives all the same. But if the stat feeds a judgement type your class’s rotation never produces — stacking area enhancement with no area attack in the cycle — the actual trigger rate is zero.
Four things that give damage and no combat power
Skill level. The skill coefficient is base × (1 + 0.03×level + 0.02×n), where n gains one at levels 2, 10, 15, 20 and 30. Pushed to 30: 1 + 0.9 + 0.10 = 2.000.
The skill’s damage exactly doubles, and combat power moves by zero. That 2.000 is neither a coincidence nor an estimate — it was checked against measured skills from two different classes and came out with no error. Skill enhancement is the most deeply buried second progression track in this game.
Final damage %. It sits at the outermost layer of the damage chain, and the combat-power expression contains no such term. On the Korean server, accessory transcendence pushed across all three pieces is +9.27% final damage and zero combat power — one and a half million gold spent on something the character sheet never shows.
Gem tag damage. Its own multiplicative bucket, and likewise absent from combat power.
Crit multiplier. The crit-hit number counts as a sub-stat and carries a weight. But the multiplier it buys — 1.4 + stat/5000 — is multiplicative behaviour, and combat power only counts the number in front, never the curve behind it.
So what does “30,000 combat power” mean?
It means you passed a threshold check. The recommended-combat-power figures exist for exactly that: they are the doorman at a dungeon entrance, not a guarantee of output.
Picture two characters at the same 30,000. One banked the score in HP and defence, picked sub-stats their class cannot trigger, and props the total up with rune star counts. The other banked it in attack and sub-stats that actually fire, pushed skill enhancement to the cap, and matched gem tags to their rotation.
At the door they look identical. Inside, their damage can differ by more than a factor of two.
Which is why “how much combat power do I need” is always downstream of “which door are you trying to open” — thresholds are discrete, damage is continuous, and the optimal strategy for each is different. For how the multiplicative chain actually composes, and for why some of its buckets only pay inside a window, see armour break against the defenceless bucket; for the crit half of the same chain, the crit curve.
What to do with this
- When you need to clear a threshold, read combat power. If you are short, buy it from the cheapest source — HP and defence have low cost per point.
- When you need to actually kill things, read the damage chain. Skill level, multiplier-bucket options, tag coverage: combat power will not tell you about any of them.
- Before swapping a rune, compare effects before star counts. Combat power will argue for the higher-star rune, and its argument has nothing to do with your output.
- When somebody boasts about combat power, ask what their skill levels are. No single question separates score-stackers from damage-stackers more cleanly.
What could overturn this. The combat-power expression itself is only medium confidence, and two sources conflict head-on over the sub-stat and primary-stat weights (0.3 against 0.6; whether primary stats get an independent term at all). We keep both. The operator has never published the expression — only the recommended thresholds per piece of content. If the real formula is ever published, or if someone regresses stable weights out of a large character dataset, every conversion on this page needs recomputing. But overturning the structural conclusion — that the expression contains no multiplicative term — would require an example of a purely multiplicative source raising combat power, and no such example currently exists.