gayed again by the battle tower and co.
i'm relatively ok with superpowering CHing +6 lax there, but flare blitz? why would she use flare blitz? i mean i understand the whole "weakest attack that has a chance to KO regardless of accuracy" thing virtually always being true, but since when does that literally apply to CHes? CH flare blitz only does 55-65% to my lax, and i was at 59%. a flare blitz CH kills lax less than 60% of the time there. i don't get it.
sure it's "can" and not "will always" but that doesn't make any sense either, because dahlia needed a CH to beat me. there's tons of times the AI has been in position to kill a poke of mine with "the weakest move that
can KO,
if there's a CH". Thunderpunch does 54% max with a CH there (night slash does 51% max), so ok...but seriously?
i mean there've been times where the AI has used the "weakest attack that has a chance to KO regardless of accuracy" on things like my tyranitar and failed to do so, but i'm pretty sure that if "weakest attack that has a chance to KO regardless of accuracy
if there's a CH" is always true, there was another move it could have used. remember this:
http://www.smogon.com/forums/showpost.php?p=1574137&postcount=1023
Peterko said:
Your 156HP/81def latios faces Aerodactyl 4 (CB, stone edge, iron head, earthquake, dragon claw) turn one, what attack will Aerodactyl use and why?
if aerodactyl adhered to "weakest attack that has a chance to KO regardless of accuracy
if there's a CH", it would have used iron head, which does 103% with a CH min even if the aero has a hindering attack nature. i guess if it knows everyting kills with a CH it has to just use the "weakest attack that has a chance to KO regardless of accuracy". however, it had to be a hindering attack nature like peterko said it was, because neutral attack (162) does 106% min with stone edge, and obviously more with dragon claw. 145 attack SE does 95% min, our "weakest attack that has a chance to KO regardless of accuracy", which i guess works out.
i later noted that i "distinctly remember it being Dragon Clawed at least once" with my latios behind a sub, but that i didnt think that i was at 100% HP. my own latios has 154HP/98Def stats at Lv50 (either 98Def or 99), and it's interesting that 145 atk aero never does 100% with stone edge, doing 82-96%. i reeeeally dont think i was at 100% when it hit my sub though. 162 neutral attack CB SE does 91-108% to my latios, so that is the "weakest attack that has a chance to KO regardless of accuracy" ignoring CH chance (it has never used GI or IH to my memory). this all means that it had to have been at 145 attack the time it used DC on me, but more importantly: CB DC does 87.01-102.60%. CB CH IH does the same exact damage, which means...nothing since they randomly decide which moves to use if they're the same power, so even if CB IH were preprogrammed to CH i wouldn't have evidence either way.
peterko did note that he was paying attention to cb aero and "IIRC it always dragon clawed me hence the 3-0 skarmory sweep", which is odd because stone edge is the weaker move. stone edge literally is always the "weakest attack that has a chance to KO regardless of accuracy", as even a CH CB IH on peterko's latios does 103% min as stated above. so something doesn't check out...it should never, ever have used dragon claw on peterko's latios because stone edge is always the "weakest attack that has a chance to KO regardless of accuracy", even if aero has 145 attack and not 162, because 145 atk means 95% min. so something is still a little weird.
if my snorlax had 103 defense or more (instead of 92) it never dies to CH Flare Blitz becaue of Thick Fat, but there's already a lot of evidence to support the suspicion that the AI doesn't know your ability, and if I didn't have thick fat it does 109% min. however, all of this is ignoring the fact that there's no guarantee that i'm smart enough to use Return there with 151/256HP lax facing blaziken and not rest, and superpower is a 3HKO on lax even after attack drops.
in sum (i hope you didnt read all that shit), there is no reason for the AI not to have used SE on peterko's latios every time no matter its attack, unless it accounts for CHing with stone edge. but that means it indeed can pick and choose when it wants to CH, because it should then use Iron Head if it knows it is going to CH, because that does 103% min even from 145atk, where every other move does more
if the move it's going to use is going to CH, which i point out because while CB DC and a CH CB IH do the same, a CH CB DC obviously does double and is actually the most powerful attack any CB Aero can use on any Latios no matter what.
while my +6 Def lax vs blaziken case doesn't give conclusive evidence to the AI CHing you when it wants/needs to, when coupled with peterko's CB aero vs latios case it seems to prove beyond reasonable doubt that
the AI can pick and choose which attacks will CH. again, superpower is a 3HKO on my lax without CHes, and there's no reason to believe I'm using Return (unless the AI utilizes input-stealing like some arcade fighting games as well and doesn't decide what move to use until you select an attack, which I wouldn't rule out either). If the AI does not know it's going to CH or isn't input-stealing, there's no reason not to use Superpower. i think ive rambled on enough though, someone feel free to point out a flaw in that logic