General Discussion - Page 3971

On September 24 2013 14:04 Kuroeeah wrote:

aui isn't really as stable of a carry either

On September 24 2013 14:31 trinxified wrote:


100% better than EE at this point. Vice versa with support role.

On September 24 2013 14:20 BlitzerSC wrote:
Default korean announcer is the best. Especially the mega-kill part.

On September 24 2013 14:00 trinxified wrote:
Anybody else feel like RS.int should put EE back on support and Aui try to carry? Isn't this more natural? EE as carry hasn't been a factor in any of their games... and Aui as support just seems a bit lost sometimes.

Their games won are mostly carried by either bone7 or Sing.

On September 24 2013 18:15 Zealos wrote:
Dangerously tempted.
Say me and a friend wanted to Bara into high matchmaking. What should he play? : D

On September 24 2013 09:40 Mithhaike wrote:

It has always been that way, what changed is the speed of armlet toggling.

Previously with the macro you can toggle through any DOT in the game, doom ion shell poisons etc, now it should be impossible to do that because they reduce the toggle speed i think.

On September 24 2013 19:05 TheYango wrote:

I did some testing just now with regard to armlet toggling.

The armlet toggling speed has indeed been adjusted. If you mash it really fast, you actually hit the "CD". This happens pretty much guaranteed if you double tap, and you consciously have to wait a split second or just press the key again. Now the thing is, this actually replicates the DotA 1 behavior correctly, it's just somewhat awkward because it replicates the DotA 1 behavior at pretty significant ping.

With the way the Warcraft 3 engine worked, the "speed" of armlet toggling wasn't actually fixed. Rather, it was pretty much dependent on your ping. The way it worked was that your game doesn't register armlet as being deactivated until the server processes your "toggle armlet off" input. So if you double-tapped armlet quickly, it would basically just register two "toggle armlet off" commands rather than sending one off and one on, and just throw the second one out. Your speed of armlet toggling could never be faster than your ping to the host. Practically speaking, this means armlet toggling could rarely be faster than ~0.07s (tested on a local game, not sure why this doesn't go lower), and for most users playing online and not on LAN would probably not be faster than ~0.1-0.2s.

Originally, what DotA 2 did was buffer the toggle armlet on/off commands, which is why things like shift-queueing and aliases even worked. Obviously, the Warcraft 3 functionality is not something that should be replicated (the effectiveness of the item being dependent on your ping is silly), but the effectiveness of aliases, shift-queueing, and sometimes even manual double-tapping far exceeded what was possible on the WC3 engine. The former 2 methods could reliably toggle armlet in ~0.02s which is why you could reliably toggle through even DoTs that tick every 0.1s. The item simply was never balanced around being able to do this. Valve probably picked a delay that they felt was representative of the original Armlet functionality. It just happens that the value they chose is probably closer to the ~0.1-0.2s delay for someone playing online rather than the ~0.07s delay that was theoretically possible on LAN.

On September 24 2013 19:01 Zealos wrote:
Ok, lets go to the logical conclusion. Ultimate pubstomp composition for a AP 5 stack?