Godlike
Deamoned (Breath of Fire: Dragon Quarter) vs Katrina Liane Elesius (Wild ARMs XF): Not... actually that bad a match, because Katrina wins with lots of weak (for Godlike) attacks, the type of thing that struggles against Absolute Defence. But... she's still probably not going to need to get more than 5-6 hits, Ned doesn't need many less (4-5), and Katrina is significantly faster to start with and it only gets worse from there.
Defender X (Final Fantasy X) vs Jecht (Final Fantasy X): Physicals aren't the way to go against Defender X. I think Jecht can win this if you both restrict the use of Haymaker and treat Jecht's attacks as piercing but yeah no way I am doing both of those, either is a stretch to me.
Lady (Shadow Hearts: From the New World) vs Sanctuary Keeper (Final Fantasy X): Sanctuary Keeper is faster and can two-shot with Mana Breath! Unfortunately I don't think it survives the full wrath of a Lady turn + the death counter. He might if Lady's Umbral turn comes before her own turn.
Jade (Breath of Fire) vs Tir McDohl (Suikoden Series): I wouldn't be surprised if Jade only 3HKOed here (Tir MDef + element halving), in which case Tir does an amazing amount of damage before he loses. Unfortunately, the key word there is "loses".
Heavy
Garnet til Alexandros (Final Fantasy IX) vs Ominas Crowe (Bravely Default): I was ready to vote for Garnet, but apparently she can only half fire. As such some combination of status and fire magic should overwhelm Garnet before she can slug Ominas down.
Tony Eisler (Mana Khemia: Alchemists of Al-Revis) vs Artea (Lufia II: Rise of the Sinistrals): Tony 2HKOs and does break damage while he's at it. To win this, Artea needs both to 3HKO and to have a 4-3 (so he can heal on turn 2 and attack otherwise). By my math he just misses the second, and the first is debatable.
Lawfer (Valkyrie Profile) vs Geddoe (Suikoden III): I intuively think that Lawfer should be able to defend lightning, both because hey he resists lightning physicals and because there's a bunch of lightning-resistant armour in VP. But none of it's DL-legal, so Geddoe's got this.
Say’ri (Fire Emblem: Awakening) vs Oerba dia Vanille (Final Fantasy XIII): Say'ri wins if she gets a turn, and Vanille's only way to stop that is with Pain. Between Say'ri's evade (which I'm certainly inclined to use for damage-with-added-status) and the fact that Vanille pretty much needs to short-stock her ATB to go first, I don't think Pain lands in time.
Middle
Typhlosion (Pokémon) vs Juno (Soul Nomad & The World Eaters): I think so. Both 2HKO each other (Juno hitting a weakness), Typhlosion is faster (even after Juno buffs).
Barret Wallace (Final Fantasy VII) vs Yang Fang Leiden (Final Fantasy IV): Block sleep, get off Hammberblow.
Meganium (Pokémon) vs Stocke (Radiant Historia): Meganium is faster, uses Light Screen, but Stocke hits weakness with G-Fire so he still 3HKOs, and isn't dying first. Mega can turtle with Synthesis somewhat but since the speed gap is small and Light Screen wears off this is useless.
Light
Kahn Marley (Suikoden II) vs Swordmaster (Bravely Default): Healing ends what small glimmer of an argument Swordmaster may have had.
White Wizard (Final Fantasy) vs Mao (Shadow Hearts: From the New World): WW absolutely must cast Wall, so the gameplan is Wall then get lots of evade going. Surged Mao can still do 35% with a finisher even then, and 15% otherwise (so forces a heal every four attacks). WW will probably need to heal multiple times just setting up evade buffs (since Mao 3HKOs her with physicals then, and gets the first before any evade buffing). Cure 3... doesn't really offset Mao's spammable damage, so yeah, I think WW runs out of resources before one Fade and a short chain of mighty 19HKO physicals can seal the deal. EDIT: Sanity could be an issue here for Mao, and the low-SP powerups Mao gets could be an issue for WW. This match gives me a headache, but I'll hold with my initial analysis for now.
Umaro (Final Fantasy VI) vs Spar (Breath of Fire II): Nulling ice and slow means Umaro just has too much durability, most likely. Could be argued.
