Nu Gundam (BG) vs Strike Freedom (BW)

30 Aug 2026, 21:20 · engine 0.1.0 · pilot mode: llm

10 gamesavg 15.6 turnsseed 8100adaptive series (10 games)

9 / 1

AI summary

SERIES RESULT: Nu Gundam (BG) 9-1 Strike Freedom (BW) - two subagent pilots that rewrote their own notebooks after every game (final notebooks: sim/live/series/notes-p1.md, notes-p2.md). THE ONE RULE THAT DECIDED ALL TEN GAMES: attack EVENT count, not damage. Each unblocked face event breaks ~1 shield regardless of AP; at 0 shields the next unblocked face event ends it. Every race is events vs ready [Blocker]s + shields + base. NU (9 wins): curve cheap bodies from T2 and flood events; pair Amuro ST01-010 on an expendable body as pre-combat removal (When Paired rests an enemy <=5HP printed); fresh deploy + matching pilot = attacks the same turn; keep Nu LR + GD05-085 for the kill turn; four small events close faster than one big one. SF (1 win): structural problem - defense curve starts at L3-4 while BG floods 2-cost bodies from T2. Blockers are a counted resource (below 2 total, deploy instead of attack); NEVER rest both blockers at low shields (three of the nine losses were exactly that); Argama base at <=3 shields absorbs an entire flood event; the LR combo turn is total resources = 8 (deploy GD05-002 + pair GD05-081 in one plan); the lone win came from an Amuro-rest tempo start with the LR landing before the flood. DECK VERDICT: to make SF competitive, add early blocker/Argama density, not more top-end. Cross-check: 1200-game heuristic MC (see the validation batch) - the doctrine cannot be expressed as profile knobs (all combos 68-71%); the pilots' edge is per-turn event arithmetic.

▸ Games (10)✦ Strategy

Every game in this series turned on one number: attacks made vs attacks the defender could block or absorb — each unblocked attack breaks one shield regardless of the attacker's power, and the first unblocked attack after the shields are gone ends the game. Both guides below play to that number.

P1 · Nu Gundam (BG)

deck guide · how to play it & where it breaks

Nu Gundam (BG) — deck guide

The plan in one line: out-number the opponent from turn 2 and attack every single turn — this deck wins by making MORE attacks than the opponent can block, not bigger ones.

The one rule that decides every game: each unblocked attack breaks exactly ONE shield, no matter the attacker's power — a 1-power chaff breaks a shield the same as the 7-power Nu Gundam. When shields run out, the next unblocked attack wins outright. So count attacks vs their standing Blockers, and ignore damage.

Early game (turns 1–5)

Deploy a cheap unit every turn — Jegan, Guntank, Zaku II. Yours is the cheapest curve in the matchup; that IS the deck.
Attack the player whenever the lane is open. You lead the shield race by turn 5 almost automatically.
Guntank's deploy effect finishes off weakened enemies for free — remember it when something sits at 1 HP.

Mid game (turns 6–10)

Pair Amuro on a cheap body, never your best one. The moment he's paired he rests one enemy unit — that's a Blocker switched off for your attack wave. Time it with your biggest swing.
Re-GZ ramps an extra resource, which pulls the Nu Gundam turn forward — grab it around turn 6–8.
Three small attacks beat one big one. Every time.

Closing (turn 10+)

Nu Gundam + Amuro attacks the turn it lands, and when it kills a unit its Breach 5 throws 5 damage at the shield area (base first) — the only thing that punishes a base, since bases otherwise swallow all damage.
Build the kill turn: rest their blocker with Amuro, then send 3+ attacks into the gap.

Strengths

Cheapest deployment curve — on the board first, attacking from turn 2
Amuro's pair-trigger deletes a blocker from the defend step
Link (attack-on-deploy) combos create surprise kill turns out of an empty board
Breach closes games through bases and shields alike

Weaknesses

Units are small individually — a wall of tough Blockers trading up blunts you
No hard removal: everything is combat math, so a −3 AP combat trick can flip your clean kill into a mutual trade
Zero plan B if you stop attacking — a stalled board is a lost board

The one series loss

They landed early blocker tempo AND their big card before the flood arrived. If their blockers + shields absorb your attacks two turns running, you lose the race — so never let a blocker live just because it looks small.

▸ raw pilot notebook (technical)

Nu Gundam (BG) — FINAL series notebook (seat P1)

SERIES RESULT: 9-1. Game 10 was the cleanest kill: T10, their shields 2, their board only SF+Kira → paired GD05-085 Amuro onto a Re-GZ (+2/+2 → 5AP), four face events vs one possible block = 2 shield breaks + 1 lethal at 0. shield_empty T9. The win condition is arithmetic, not luck.

The five doctrines that decided the series

1. Count attack EVENTS, not damage. ~1 shield breaks per attack event

regardless of AP; at 0 shields the next unblocked face EVENT wins. Four 1-3AP bodies close faster than one 6AP body. Build every endgame turn to maximize event count.

2. Run the leak check every turn: my attacks − their READY [Blocker]

units vs their shields. Only [Blocker]s block (Lfrith IS one — check the kw line). Rested or non-Blocker actives never intercept. ≤0 spare = don't overextend.

3. Fresh deploy + matching pilot = attack on deploy. The deck's tempo

engine (BWS+ST01-010 4AP T8; every kill turn used it). Fresh alone never attacks. Never send a plan without its attack steps — a plan without attacks still ends the main phase.

4. Amuro ST01-010's rest trigger (printed HP ≤ 5) is a pre-combat removal.

Resting their one blocker converts "probable win" into "forced win" — and it filters by PRINTED HP, not effective. Pair it on the expendable body first.

5. Blocked attacks still damage the blocker. 1-effective-HP chaff dies to

any hit; 5AP kills SF+Kira. Send the big hit first to make every block lethal for them.

What beats SF (BW)

SF+Kira lands ~T9-10 as a 5AP/5HP wall and they hoard cards (8-card hands) and double-block with chaff. It never mattered: event-count pressure + one rest trigger outpaces their hand. Their -3AP combat trick (Unforeseen Incident) can't stop leaks — leak math is event-based. Watch it on attack declarations, but don't let it deter the swing.

For the next Nu pilot (one page)

Curve Jegan/Guntank → Kayra Re-GZs → BWS+Amuro → Nu turn. Ramp matters:

BWS's EX Resource compounds; the Nu turn lands a turn earlier.

Keep GD05-085 for Nu (7AP/7HP heal-on-kill) unless the kill turn needs the

+2/+2 NOW — game 10 spent it and won; that's the right greed when 3+ leaks are available.

Guntank's deploy ping is not dead: scan effective HP for 1HP targets.
Driver: plans resume after menus; re-`--next` after every answer; `end`

LAST. Duplicate numbers on board + hand cause stale steps — sequence attacks before deploys.

Mirror-check trust: first-player advantage ~63%; but this deck's edge is

systematic pressure — count events, remove blockers, win on the math.

P2 · Strike Freedom (BW)

deck guide · how to play it & where it breaks

Strike Freedom (BW) — deck guide

The plan in one line: survive the early swarm behind Blockers and a base, then land Strike Freedom and take over — this is the control deck of the pairing, and the whole skill is living long enough.

The trap that loses this deck games: attacking too much. A unit that attacks rests, and rested units can't block. Three of the nine series losses were exactly this: swung with everything, got swung back at, died with an 8-HP monster standing there useless. Below 3 shields with no base up, attack with ONE unit at most and keep a Blocker standing.

Early game (turns 1–5)

Deploy a Blocker every turn — Rick Dias, Gundam Lfrith, Strike Rouge. Blocker count is money: track it, spend it deliberately.
You will not out-tempo them; don't try. Block to KILL their attackers (a block also damages the attacker — choose the blocker whose power kills what's coming).
Argama (base) is your best anti-swarm card. Once you have a spare resource and shields are ≤3, put it down: a base absorbs the ENTIRE damage of one attack into its own HP — even a huge hit takes only the base, never spills.

Mid game (turns 6–10)

Unforeseen Incident (−3 AP) is a combat trick — cast it DURING an enemy attack to turn their swing into a suicide. Casting it with no battle happening is a wasted card.
Count their attacks vs your Blockers + shields EVERY turn. If their number is bigger, deploy instead of attack.

The Strike Freedom turn

At 8 total resources: deploy Strike Freedom and pair Kira on it in the SAME turn — it attacks immediately at 7 power, removes their biggest threat and refills your hand. This is the win condition; everything before it is survival.

Strengths

The most Blockers in the pairing, plus base absorption — the best late-game defence
Strike Freedom + Kira is the strongest single turn either deck has
Combat tricks turn bad blocks into even trades

Weaknesses (structural — why it went 1–9)

The defence curve starts at level 3–4; the swarm starts at 2. Behind on board from turn 2, often for good.
The win condition arrives around turn 10–15 — usually too slow against the flood.
Card advantage doesn't help: an 8-card hand blocks nothing.

If you want to fix the deck

More early Blockers and more Argama copies; cut expensive top-end. L1–L3 Blocker density is the single change that would move this matchup.

▸ raw pilot notebook (technical)

Strike Freedom (BW) — FINAL series notebook (seat P2, series 1-9)

Game 10: LOST T10 "shield_empty". The cleanest loss of the series: I read the lethal turn exactly (their 4 bodies, my 2 shields, 1 blocker = death on their 4th event) and died to it anyway, because I sequenced my T9 wrong — the dig found the answer (Lfrith) on the fourth card and I had already spent the mana to deploy it.

THE TEN GAMES IN ONE SENTENCE

Every loss traced to the same counter: **enemy attack events > (blockers + shields + base) on one enemy turn.** Not AP, not card advantage, not the LR race — event count. SF lost the tempo war to BG's cheap flood every game because this deck's defense engine starts at L3-L4 (Dias/Lfrith/Gundam) while BG deploys 2-cost bodies from T2.

What an SF pilot must know (hard-won, all ten games)

1. Blockers are a counted resource — hand + board. Below 2 total, deploy instead of

attack. Three games died with both blockers rested; game 10 died with zero spare.

2. 1 shield per attack EVENT, both directions, always. AP only matters for killing

units and bases. Count races in events, never in damage.

3. **A base absorbs ALL face damage of one event into its HP; excess never spills — even

a 1HP Argama eats an 8AP turn's first swing entirely.** Deploy it the moment you have spare mana at ≤3 shields; it dies defending but the shields behind it don't.

4. 0 shields = sudden death, not game over. The loss triggers on the first event that

LANDS at 0 with no base. Count "events that will land at 0," not "shields hit 0."

5. LR combo turn = the turn total resources hit 8 (level gate is TOTAL, cost is READY):

deploy GD05-002 + pair GD05-081 in ONE plan. Never earlier. In this engine's pacing that is ~T15 — the reason SF must survive to T15 first, and the reason defense spending always outranks attack spending before then.

6. Amuro ST01-010 When Paired is a double-edged rest. Mine: pair → rest their best

≤5HP body → kill it same plan (best tempo line in the deck). Theirs: they rested MY blocker to open the attack lane (game 10, T8). Treat enemy Amuro pairing as "one of your units is about to be rested" — don't rely on a ≤5HP blocker holding a lane against a linked Amuro attack.

7. Exact-kill damaged units with chaff AP (HP − damage). Free kills win games 8 and 9.
8. Blocker choice: pick the blocker whose AP exactly kills the attacker — chump blocks

are also chaff-clearers.

9. Repair X heals at END of owner's turn — kill Repair units with one overkill battle

or don't bother chipping.

10. **When lethal is on the board for NEXT enemy turn, pay for defense FIRST and dig only

with spare mana, one command at a time.** Game 10's fatal sequence: batched Affection + Resolve before seeing draws — drew Lfrith (the out) with 1 ready vs its C2. Resolve-first would have deployed Lfrith + held Gundam + 2 shields = survival at 0 shields. The read was right; the batching killed me.

Series-level verdict

Final 1-9. The one win came from a double Amuro-rest tempo start plus the LR combo

landing before the flood. Every other loss was event-math attrition from T8-T12.

First-player advantage (my edge this game) bought one extra blocker turn and changed

nothing — the deficit is structural: BG answers every early BW body with a cheaper one.

If the deck is revised: more L1-L3 blockers/Argama copies (defense density before the

L8 LR turn), not more top-end. The SF LR never resolved once in games where I fell behind the blocker count by T8.