Gen 4 → Gen 6 transition · checked 2026-08-07
When should you replace Rosa with Yang?
No controlled test exists for this and probably never will, because nobody can hold gear, troops, buffs, and rally composition constant across accounts. What does exist is a large and unusually consistent body of player reports. This page pools them, states where they agree, and shows how wide the spread is.
What the pooled reports agree on
Direction is unanimous. Across eleven independent accounts, including two video damage tests and several guide-site rankings, not one reports Rosa beating Yang as a Bear rally lead. That level of agreement is unusual in this community and is the strongest part of the evidence.
Magnitude is not. Seven reports give a computable uplift after swapping Rosa for Yang: +23%, +25%, +28%, +33%, +40%, +64%, and +83%. The median sits around +33%. The spread is wide because every account differs in gear, troop ratio, Truegold level, and rally composition, so treat the range as the honest answer and the median as a planning number rather than a prediction.
The breakpoint, which is what actually decides your spending
The useful question is not whether Yang wins eventually, but how little Yang you need before the swap pays. Four reports speak to that directly, and they cluster tightly:
| Reported breakpoint | Source type | What it implies |
|---|---|---|
| Yang at widget 3 roughly equals Rosa at widget 10 | Video damage test | The widget gap you can concede is large, around seven levels. |
| Yang at 4.3 stars with no widget shows no notable drop against a maxed Rosa | Player report | Parity arrives before Yang is finished. |
| Yang at 3 stars and widget 2 already outperforming Rosa | Player report | Some accounts cross over earlier still. |
| Thread rule of thumb: Yang at 4 stars and widget 2, or widget within about 2 levels of your Rosa | Community guidance | A conservative planning figure, not a measured result. |
Practical read: the swap tends to pay somewhere between 3 and 4.3 stars, well before Yang is finished. Waiting for a maxed Yang is the expensive mistake. Targeting roughly 4 stars and widget 2 is the defensible plan.
Why Yang wins, mechanically
The reports agree partly because the kits are not close. Rosa carries a 40% chance of 50% damage plus a 30% archer attack buff. Yang carries a 40% chance of 100% damage, an Ambush effect adding a further per-turn damage chance, and the only direct rally Lethality widget in Generation 6, sitting on the highest base attack at launch. Widget bonuses compound multiplicatively, and only the first expedition skill applies to a rally, which is exactly where Yang's advantage lands.
The two detailed cases, in full
| Yang | Rosa | Reported result | Evidence limit |
|---|---|---|---|
| 3 stars, widget 2 | Not stated | The player reported that Yang was already outperforming Rosa. | The player gave no star level, widget level, or damage figure for Rosa, and no controlled gear, troop, buff, or rally comparison was published. |
| About 4.2 stars, widget 5 | 5 stars, widget 7 | The player reported a Bear Hunt increase from roughly 60–80 million to 100–130 million. | The ranges are self-reported and cannot predict another account's damage. |
Where sources disagree
One community wiki still lists the Gen 6 free-to-play Bear lineup as Zoe, Petra, and Rosa, while putting Yang in the spender lineup. That does not contradict the damage reports. It answers a different question: whether a free-to-play account can realistically reach Yang's breakpoint at all, given that Yang comes from rally and ranking events rather than the wheel. If your shard income cannot get Yang to roughly 4 stars in a reasonable window, a finished Rosa is the better real lead, even though a developed Yang would beat her.
What this evidence still cannot tell you
- It cannot predict your damage. The uplift range spans nearly fourfold across accounts.
- It cannot prove a Bear Hunt gain transfers to Arena, garrison, or every rally type.
- It cannot set a universal star threshold, because no report controls for gear, troop ratio, or Truegold level.
- Every figure here is self-reported. Two come from video tests, which are more repeatable than a forum post but still single-account.
Your replacement checkpoint
- Confirm Yang's live acquisition route in your kingdom, since rotations vary.
- Record Rosa's exact stars, widget, gear, and Bear role.
- Estimate the Yang stars and widget you can actually finish, not merely unlock.
- Compare repeatable Bear runs under the same rally, troop, buff, and gear conditions.
- Keep Rosa if you cannot reach roughly 4 stars and widget 2 on Yang in a reasonable window.
Current decision status
Verdict: replace Rosa with Yang for Bear rally lead once Yang reaches roughly 4 stars and widget 2. Do not wait for a finished Yang, and do not switch to one you cannot develop.
Confidence: medium on direction, low on magnitude. Eleven independent accounts agree Yang wins and none disagree, which is strong for direction. The reported uplift still ranges from +23% to +83%, so any single number would be false precision.
Recheck when: a balance update changes either hero, the Gen 6 event rotation changes, the Gen 7 archer arrives, or a controlled multi-account test is published.