Rally Joiner Selection and Garrison Management — Companion to Ton's Ratio Sheet

Ton's spreadsheet already tells you which joiners are good. What this video adds is everything around it — how those joiners actually get into the rally, and why 20% skills and Mia are worth something different from what most people calculate.

1. Seeing what is actually applied

Where	What it shows
Live rally → yellow flag icon	Rally stats, rally capacity, active hero skills (up to 9), active widgets
Battle report → Details	The leader's 3 heroes + which 4 joiners were actually picked

3 leader heroes × 3 expedition skills = 9 slots. Purple (4★) heroes only have 2 expedition skills, so the count drops. Slots fill in real time as people join; "4 rally joiner skills active" means all four are locked in.

For alliance training, the report details screen is the fastest tool — a duplicated Mia shows up immediately.

2. Who becomes one of the four joiners

	Offensive rally	Garrison (defense)
Selection	First four to press join — not the first to arrive	Top four in the list. The garrison leader is excluded, so it is effectively 1–4 or 2–5
Skill level	If the first hero's first expedition skill is not maxed (Lv.5), that player is skipped entirely and the game moves to the next one
Fixing it after	Impossible → the only lever is pre-assigning who joins first	Possible → promotion cycling (below)

There is exactly one way to control an offensive rally: assign four named players per rally leader ("you join leader A's rally first, with Mia"), and let everyone else pile in behind them.

3. Bonus source — the garrison leader is not the buff owner

	Offensive rally	Garrison (defense)
Buff owner	Always the rally leader (cannot change)	The player with the highest stats, chosen automatically
What is used	Leader's heroes, experts, chief gear, charms, research, pets, ministers, skins, widgets	The bonus source's heroes, experts, stats and gear
Exception	—	Only deployment capacity comes from the garrison leader

"Strongest" flips depending on whether pet combat buffs are on. A normally stronger player loses the bonus source if their pets are off.

Practical rule — the top player in a garrison must turn pets on. And since only capacity is taken from the garrison leader, that seat is best held by whoever has the biggest command/research capacity.

4. Garrison joiner cycling

The mechanic: promote someone to garrison leader, then promote someone else to leader — the first player drops to the bottom of the list.

Example — keep 2/3/4, drop 5, pull in 7:

1. Promote player 5 to garrison leader
2. Promote player 6 (also unwanted) to garrison leader → player 5 falls to the bottom
3. Promote the original leader back → player 6 falls to the bottom
4. Player 7 moves up and becomes the fourth joiner

The same works for player 2: promote them, then promote the original leader back. During castle battle this is a continuous chore, not a one-off.

5. Why 20% joiners get cut — the actual math

Common misconception: a 25% buff is ×1.25 and a 20% debuff is ×0.8, so 1.25 × 0.8 = 1.0 and they cancel out.

Reality: reductions divide, they do not multiply.

Attacker: one 25% damage-up joiner   →  ×1.25
Defender: one 20% damage-taken joiner →  ÷1.2

Result = 1.25 ÷ 1.2 = 1.0417  →  attacker is 4.17% ahead

With four of each (same bucket is additive):

4× damage-up   = ×(1 + 0.25×4) = ×2.0
4× reduction   = ÷(1 + 0.20×4) = ÷1.8

Result = 2.0 ÷ 1.8 = 1.111  →  attacker is 11.1% ahead

So a 20% reduction is not "5pp weaker than 25%" — it is structurally weaker. That is why the spreadsheet drops the 20% joiners entirely.

Exception: against an opponent stacking the same hero four times, 20% skills become worthwhile — their additive bucket is already saturated, so your reduction bites harder in relative terms.

6. Why Mia has no substitute

Mia's joiner skill is a 50% chance per troop type, per turn, to curse the target → the cursed target takes +50% damage. That means one independent roll per troop type.

Troop types in the march	Chance to trigger in a turn	Expected value (chance × 50%)
1 type	50%	25%
2 types	1 − 0.5² = 75%	37.5%
3 types	1 − 0.5³ = 87.5%	43.75%

Jessie, Jasser, Seo-yoon, Patrick, Blanchette, Bradley and the rest are all a flat 25%. With two troop types Mia is 1.5× that; with three, 1.75×. No other single joiner swap moves the needle this much.

Never stack her. A second Mia does not raise the 50% damage multiplier — it only raises the trigger chance, and at three troop types you are already at 87.5%, leaving 12.5pp of headroom. The second Mia is worth at most +6.25pp. Put Hendrik, Patrick, Norah or Renee there instead: a different bucket multiplies.

7. Renee and Hendrik

• Renee — in any lancer ratio where the leader team does not already have her, a Renee joiner is mandatory. In later-generation 60/40 lancer setups, Mia + Renee + Hendrik are effectively fixed and only the fourth slot changes.
• Hendrik — enemy defense −25%, and he owns that bucket alone, so he fits almost anywhere. The exception: Ligeia (Gen 12 marksman) shares his bucket, so rows with Ligeia in the leader team drop Hendrik.

8. There is no such thing as an "offensive" or "defensive" joiner

The damage formula is identical whether you attack or defend. There is no defense bonus or offense bonus. Sorting joiners into "Sergey/Patrick for defense, Jessie/Seo-yoon for offense" is simply wrong.

Defensive skills tend to be 20% and offensive ones 25%. So stacking more 25% offensive joiners wins even in defensive scenarios — Ton has tested this repeatedly.

The only real offense/defense split = widgets	Applies to
"Squad attack", "Rally attack +15%"	Offense only
"Defender troops' attack" etc.	Defense only

Widgets are a separate multiplier applied on top of both troop types, so they matter a lot. That is why offense and defense use different leader heroes → different occupied slots → different leftover slots → different joiners. The joiners change because the leader's slots changed, not because some joiners are "defensive".

9. Ratios are almost never countered

If you built for it, nearly every ratio in the sheet works. Reading the enemy's ratio and answering with a counter-ratio is not realistic in practice.

Rarely, a setup is close to unbeatable — Gen 6 META: 45/5/50 or 40/0/60 · Logan + Philly + Wayne (defense). Stronger players keep failing to break it. The answer is not a ratio: it is more rallies, better timing (tight simultaneous landings) and strategy.

Priority: joiners > leader heroes > strategy and timing > troop ratio

10. The Gen 10+ T12 rule

T12 lancer skills buff both infantry and marksmen. So even a marksman-heavy ratio should carry at least one lancer's worth (up to 8 players) to trigger the T12 lancer skill. Anyone still running a pure 0%-lancer marksman ratio needs to fix it.

11. Anything from before the December 2025 patch is suspect

• That patch reworked a lot of Gen 1–6 hero skills. Guides and videos made before it are unreliable for those generations.
• The same goes for advice from friends on older, higher-generation servers — they do not track the low-gen meta.
• Molly and Philly were re-evaluated as marksman-ratio defense leaders.
• Ahmose got worse, so Logan stays in garrison duty up to around Gen 6.

12. Spreadsheet changes (checked 2026-09-01)

① New Gen 5 defense row (added 2026-07-19 from viewer reports)

Ratio	Leader	Joiners 1–4	Alternative
60/40/0 (defense)	Logan + Norah + Greg	Mia · Norah · Jessie* · Philly	Norah or 25%

② New META tag — Gen 6 defense 45/5/50 · 40/0/60 (Logan + Philly + Wayne), plus two Gen 12 rows.

③ Gen 12 joiner columns are now filled in — shared note on every row: to test — Wu Ming (double effect with Hervor to counter Ligeia)

Type	Ratio	Leader	Joiners 1–4	Alternative
Lancer defense	60/40/0	Hervor + Lloyd + Bradley	Mia · Renee · Patrick · Hendrik	Gatot
Mix defense	60/20/20	Hervor + Lloyd + Ligeia	Mia · Patrick · Norah · Seo-yoon	Gatot · Norah
META mix defense	60/20/20	Hervor + Lloyd + Ligeia	Mia · Gatot · Patrick · Seo-yoon	Norah
Marksman defense	50/2/48	Hervor + Molly/Philly + Ligeia	Mia · Patrick · Norah · Gatot	Norah
Mix defense	50/10/40	Hervor + Lloyd + Ligeia	Mia · Patrick · Norah · Gatot	Seo-yoon
Marksman offense	48/4/48	Hervor + Mia + Rufus	Patrick · Hendrik · Norah · Gatot	Norah
META mix offense	50/10/40	Hervor + Karol + Rufus	Mia · Patrick · Hendrik · Norah	Norah
Lancer offense	60/40/0	Hervor + Karol + Bradley/Greg	Mia · Renee · Gatot · Hendrik	Patrick

How to read it — rows with Ligeia in the leader team have no Hendrik (same bucket, additive, wasted). Rows led by Rufus or Bradley do take Hendrik. Mia appears in 7 of the 8 rows; the one exception already has Mia in the leader team (marksman offense 48/4/48). Rows with any lancer weight take Renee. Gen 13–16 are still empty templates.

Field checklist

Rally leader

• Full buffs (consumables, pets, chief gear, charms, ministers) — only the leader's apply
• Check the yellow flag for the intended 9 slots + widgets
• Name the four joiners in advance, including their join order

Joiners

• Slot 1 = your assigned hero, first expedition skill at Lv.5 (below that you are skipped entirely)
• Send your highest-tier troops (troop stats do carry over)
• Consumables, chief gear and pets are pointless — none of them apply

Garrison manager

• The strongest player turns pets ON (it decides the bonus source)
• Garrison leader = whoever has the largest capacity
• Verify the top four; fix them with promotion cycling
• Confirm exactly one Mia, plus one Renee in lancer setups

After the fight — open report details and check whether the four joiners were the intended ones. Duplicates (two Mias) feed straight back into the next rally.

Source — TontonYouTube, (Gen 1 to 12) Best Rallies to USE to win every battle!
Public sheet: Heroes, Rallies, Ratios and Joiners (re-checked 2026-09-01) · Discord