On September 13 2011 14:12 DustyShelf wrote:Hello! I am the creator of the original test (tpyo / BelligerentWombat) and the videos that are linked to. I just wanted to add my 0.02 cents.
First off I should really clarify that
I 100% agree with the TL;DR on the original postThat is:
"it probably requires unrealistic conditions to take full effect" I apologise for my sensationalist headline and minimalist commentary on my initial posting on reddit, it has resulted in the community wildly overreacting to the change. It
probably wont be an issue for most games. I have performed further testing on more realistic death balls and have clarified that everything
should be okay.
Anyway, I did want to clarify a couple of things. The tests were
not "failed" in any way. While I appreciate that I used a somewhat unrealistic parameters but I really wanted to see if there was any change between the two versions, hence I made a huge ball of closely clumped together zealots.
I verified my results multiple times. Exactly the same micro and setup was used almost every single time. The holes circled in the OP are not real gaps but merely appear to be due to the perspective of the 3d.
Unless 1.4 also has also made changes to the way one can clump a bunch of zealots
those gaps do not existTo clarify my findings: There
is a difference between 1.3 and 1.4 in the way that the drop code is handled. In 1.3 it will permit the dropping of units into a tight cluster of units if it cannot find available space within a specific radius. In 1.4 it will delay the drop command until if finds viable space.
However in 1.4 if it fails to find available space within a specific period of time the drop command will be
cancelled. This is why in my 1.4 test you can see that most of the overlords fail to drop their full contingent of banelings.
Now while this should not be an issue in most games my fear is that cunning players of the future will find mechanisms to exploit this logic and cause baneling drops to fail. I really wanted to bring light to the fact that there is a flaw in the new logic in 1.4 that should be resolved before people find mechanisms to exploit it. I have added some musings to my thread on
battle.net (*shudders*) and I hope the devs take note and consider improving the logic.
One example of such exploits that concern me is how players exploit the "hold" position command on workers in micro fights to make melee troops act idiotically. People seem to champion this as "amazing micro" whereas (as a software engineer) I see it as just a flaw in the system that should be fixed (not that I dislike players that do it but I don't like the fact that the exploit exists at all).
Once more I'd like to apologise for creating the video without the proper commentary alongside it to put into context and for my sensationalist title. I will be more careful with any future postings I make.
Thanks for reading this far.
TL;DR;
"it probably requires unrealistic conditions to take full effect" (but there is a difference between the two versions and the tests were genuine)
---------
EDIT: WHOA! Check the post at the top of this page that states:
Show nested quote +has anyone tested if surrounding your stalkers super tightly with forcefields will make it so banelings cannot even hit the prtosos untis when dropped?
I just did a couple of tests... and by all means we really need to test this further, preferably with two people to make the micro good (as its tough to do it all as one person)..... but from what I can gather
It's true.
I don't know if it is unrealistic to have so many sentries/FFs but it is at least possible.
EPT
