Colosseum commentary notes

INTRO
- What is Colosseum? What is the goal of the game?
	> Like Stadium, but has a story mode
	> Similar to handheld (RPG), but no wild Pokemon, no overworld and all double battles
	> Shadow Pokemon (more on that later)
- Explain Espeon manip briefly
- Quilava as secondary Pokemon + history of Colosseum
	> Croc used initially because it "seemed" better + Torrent Mt. Battle was very fast
	> Quilava was always considered to be close, but unappealing because of rough early game and Fire Blast accuracy
	> In general, Pokemon speedrunners undervalued late-game time saves during the early years (2012-2016)
	> Croc progression: Unoptimized (4+ hours), optimized safe route (sub-4), Master Ball skip (3:50), X item glitch (3:37), Espeon manip (3:35)
	> Ryziken (Fire lover) initially routed Quilava and it fared better than expected. Manip for HP Grass helped
	> I looked at route and found big Ferma time save, and then Quilava seemed barely faster
	> Runs showed Quilava was much faster (Mt. Battle is actually really good!)
	> Deathwarp discovery for No Snags --> great for Croc, questionable for Quilava, but Quilava still faster
- Snag Makuhita is a route change, more on that later (Maku is a 100% catch at full HP with any ball)

QUILAVA/CAIL
- Ball glitch
- Why we YOLO ball turn 1
- Stats (just need Sp. Atk because you mostly use Fire moves and you have high base Speed)
- Natures (effect on purification; will elaborate in a minute; my Quilava has the best nature -- Sassy -- because it purifies
quickly by battling and walking, both of which you frequently at this point in the game with no additional time loss)
- Late shop (walking cycle)
- During Cail fight, explain that Shadow Pokemon have random stats, but non-Shadow Pokemon have 0 IVs -- so Shadows are quite a bit stronger on average!
- More Shadow Pokemon mechanics: Shadow Rush (90 BP, no type effectiveness), Hyper Mode (crit rate + can't use items + can't reliably use other moves), experience delayed until post-purification (and none until 2 bars are cleared), moves acquired after 1st/3rd/4th bars are cleared)

PYRITE COLOSSEUM
- Experience route (20614 -- much more than Croc route! Requires calls on most natures)
- Early game comparison -- good Attack Quilava loses almost no time to average Croc!
- MIREZ: X item glitch
	> Similar to ball glitch
	> Saves turns and animation time

PYRITE UPCLIMB
- Makuhita strats + general strategy (Hyper Mode the early fights, Flame Wheel the late ones, manage HP intelligently)
- BEFORE FERMA: Fixed AI --> well-routed fights! Different than any other Pokemon game (handhelds have random moves in most cases; XD has random target)
- Ferma: Walk through normal fight, and how this one differed (bulky Quilava)
	> Also discuss Shadow Pokemon damage calculator (used to figure out Mantine 2HKO% -- but very complex with Shadow vs. Shadow!)

MIROR B.
- X item glitch starts to put in a lot of work
- Value of extra experience
- Ether skip
- Why we catch Meditite
- YOLO Ball Meditite -- same reason as Quilava

SKRUB
- Exp. Share to get L47 (Psychic), but don't put it on yet because Typhlosion's levels on Mt. Battle are more important (maintain Blaze FW + OHKO Geodude with Selfdestruct)
- Calcium skip
- Skrub: First fight with random combos (here there are 5). This adds tons of routing! No other Pokemon game has this
- Skrub: Biggest mistake (misread damage chart)

MT. BATTLE / DAKIM
- Blaze saves a lot of time! Similar to Croc
- Lucky and important first fight -- 30% to get this outcome
- Both punished and rewarded for Calcium skip
- Early Psybeam
- 5th battle: Fixed AI begins here! Helps Typhlosion a lot because it can safely stay in Blaze with Dig
- Defensive damage ranges are very important to know -- spreadsheet made by me and Ryziken
- 8th battle: X Special to save a lot of animation time
- Geodude fight: Other options
- Dakim: This fight, and the next one, show why Colosseum is the best Pokemon speedrun (IMO). The best way to think of this is to compare the speedrun strat to the typical fight that a casual player gets. A casual player is going to have a very difficult time with this fight because Dakim's Pokemon are powerful, have a type advantage against you, and use an effective strategy (Protect + EQ). You're probably going to get hit with an Earthquake or two and die. But a speedrunner can predict with 100% certainty which moves each of his Pokemon will use, which makes the fight a solvable puzzle.

VENUS
- Combo frequency (uneven on fights like Dakim)
- Turning point in the run for Feraligatr vs. Typhlosion
- Delcatty/Vileplume new strat fails because FB miss -- but can recover from this miss, unlike misses on other combos

NELIS
- Low Speed finally matters
- Fire Blast PP
- Shifted AI -- wasn't expecting it, but randomness of Stantler stats causes shift

EIN
- Lab deathwarp in depth:
	> 6 fights can be lost, but 1 is faster to win since it's quite late in the lab
	> If you win the fights, you use enough Psychic and Fire Blast PP that you have to get an extra chest of Ethers (15 seconds)
	> Still need to keep Psybeam to skip Ethers (requires time loss to animation and having Psychic in slot 2)
	> Deathwarp is faster than fights even if the fights go perfectly
	> Deathwarp loses about 9000 experience on Espeon and Typhlosion. This is about 1 level on Espeon and 2 levels on Typhlosion now (slightly less later since you require more experience to reach higher levels)
	> Plusle dies so quickly to Lare that it's worth a 37-second trip to withdraw it
- DNA sample patterns

REALGAM TOWER
- Miror B: Makuhita (safer and slower than old strats) -- needs to be sacrificed at some point, will explain more later
- Dakim 2:
	> Combo frequency affects routing!
	> Usually I try to get L50, but that would lose L45 on Typhlosion for Bellossom (guaranteed FW), and Claydol is 15/16 at L49
	> Can also heal here, but Blaze is faster if it works, and avoids crit risk trying to get back into Blaze
- Sunflora range
- Venus 2: Safe combo, usually have 1-2 risks here

GONZAP
- Klest: AI mechanics + value of "case 1"
- Gonzap: Worst combo, still better than any Croc fight!

FINAL 6
- Going against really strong WR splits!
- Houndoom was very lucky, and needed!
- Now you see why Makuhita in slot 1 is needed -- having it as a lead on Nascour/Evice baits attacks away from Espeon and allows Espeon to set up to +2 with X item glitch
- Evice: Three general strategies:
	> Most common: Set up Espeon to +3, FW Scizor, leave Slowking for last, HP Grass Tyranitar (2 turns of sand)
	> Second-most common (this one): Set up Espeon +3, FW Scizor, kill Slowking early, EQ+EQ Tyranitar (1 turn of sand)
	> Third-most common: Save Scizor for last (usually requires X Attacks), isolate Tyranitar (1 turn of sand)