How to Read Castle Clash Build Data (and When Not to Trust It)
Every hero page on this site shows percentages next to talents, insignias, enchantments, traits and pets. This article explains what those numbers are, how we turn them into the “How settled is each slot?” ratings, and where they can mislead you.
What a pick rate is
The percentages come from IGG’s official Castle Clash hero data, which reports what top players have equipped on each hero. If Bloodlust shows 73.5% on Bell Elk Maiden, then 73.5% of the top-player Bell Elk Maiden builds in IGG’s data use Bloodlust as the talent.
The numbers in a slot usually don’t add up to 100%. We show the top five options, and the rest is spread across many less-used choices.
What a pick rate is not
It’s not a win rate. A high pick rate means top players choose something. It doesn’t prove it wins more. Popular options are usually popular for a reason, but pick rates also reflect:
- Availability. Players can only equip what they have. A rare talent can be excellent and still show a low pick rate because few people own it.
- Habit. Once a build is “known,” people copy it, sometimes after the game has moved on.
- Cost. Rerolling is random and expensive, so many top players keep a good-enough talent rather than chase the perfect one.
Treat the data as “what experienced players have settled on,” not “the mathematically best build.”
The consensus labels
On each hero page we rate each slot by the top option’s pick rate:
| Label | Top pick rate | What it means for you |
|---|---|---|
| Near-universal | 50% or more | Top players agree. Aim for this option. |
| Clear favourite | 30–50% | Strong preference, but alternatives work. |
| Contested | 15–30% | Several options are fine. Use the runner-up if you have it. |
| No consensus | under 15% | Top players don’t agree. Use what you own. |
We picked these thresholds from the spread of the data. The median hero’s top talent is picked about 25% of the time, so “contested” is roughly what a typical slot looks like.
Common traps
1. Treating a weak #1 as a requirement
If a hero’s top pet is at 6%, it is technically “#1,” but 94% of top players use something else. Pets are the extreme case: the median hero’s top pet is only 5.5%. (More in our pet skills article.)
2. Ignoring split variants
Some options appear twice, for example Bountiful Gift and Bountiful Gift (M), the base and mutant versions of the same pet skill. 87 heroes have both versions of a pet in their top five. For planning, add them together.
3. Copying Epic builds onto Legendaries
Epic and Legendary heroes follow different patterns. Revive and Holy Conviction dominate Epics; Flame Guard, Tyranny and Dodge show up far more on Legendaries. Check the specific hero, not a “universal best build.” (See Epic vs Legendary.)
4. Reading Elite and Ordinary data too closely
For Elite and Ordinary heroes, top picks are often under 10% and there’s no enchantment or trait data at all. The numbers exist, but there’s not much signal in them.
5. Doubling up a talent and insignia
A talent and an insignia with the same name don’t stack; only the higher-level one is active. If a hero’s top talent and top insignia are both Sacred Light, you should pick a different option for one of the two slots.
The other numbers on hero pages
Below the build table, each hero page also shows:
- Stat ranks: where the hero’s base ATK, HP, growth, attack interval and move speed rank within its class and across all heroes. Base stats are IGG’s pre-upgrade values, so they’re useful for comparing heroes, not for predicting exact in-battle numbers.
- Where it’s a top user: items where this hero is one of the five heroes most likely to run it. This helps when a scarce item has to go to one hero.
- Heroes with similar builds: heroes that share the same talent + insignia pair, and an overlap score across all six slots (0 = nothing in common, 600 = identical). If you’re rebuilding a hero, similar heroes’ builds are worth a look.
When the data is out of date
IGG’s data updates as the meta shifts, and we re-sync periodically. New heroes won’t appear until the next sync, and a hero released recently may show little data because few top players have built it yet. For brand-new heroes, community discussion in the weeks after release is often more useful than the early numbers.
See the full methodology on our About page.