BUMP
Test 1
I think the tests havent been repeated enough and to me its not as clear as to some of you. You seem to end the tests when you got the result you wanted instead of continuing. this kind of testing would need at least 100 repetitions each to say anything.
For test 2 with shortest route, pika have one of the best vertical recoveries in the game as well as very easy understood for a computer(goes in straight lines)and pikachu is agile so this dont prove that much to me
Test 3. Well I can agree that they seem to have some sort of locking system, nothing else is really that visible in the vid, its too short IMO.
AllIsBrawl thread. My link plays the same as that one, my CPU link havent learned anything from LINKS FULL STRATEGIC GAME(lol at calling it that). They dont DAC, dont Arrowcancel, dont FJ spam, dont zair, overuses Up B,grab and Usmash uses rang at wrong situations, charges arrows, pick bombs standing, uses wrong recovery options and probs more.
The falcon pawnching seemed very convincing though, but I am wondering if it isnt just cause the spacing is the same all the time. I know the melee comps were very dependant on spacing, it was almost the most important part in their algorithm for choosing a move. Repeat this test a few more times with a recording of before and after Falcon pawnch FFA, Same stage same two characters. Maybe should be done with CPUs only(not in the FAlcon pawnch FFA of course).