Report #2 for 1.8 Beta 3Match: 3 Vs 3 (Me, as Space Marines, allied with Eldar & IG, Vs 2 x Orks & Chaos)
Map: Testing Grounds
A.I Difficulty: Hard
Resources Rate: Normal/Standard
Resources Sharing: Off/No
Winner: Me & Allies
Comments: That test had a
lot of things going on in it (a 45+ minutes match). And the most prominent Beta 3 feature in it was, I believe, the
Allied Support. That test
very clearly showed how better Allied Support really is, compared to previous builds. And the demonstration of it especially came from one of the Ork factions, namely Computer #5. But, briefly said, I still think there is room for improvements concerning that feature.
As I said, that test alone could be used to revise and analyse Allied Support, but I noticed some "problems" here and there. And, believe me, I'm sure I missed more of such issues, since that game was so long, and with five A.I players in it battling it out, I couldn't really do that replay justice with just two views at it. Technically speaking, I would logically have to watch it five times, from the point of view of each of the A.I factions. And, as of now I don't really have time to do that.
The first of the issues I noticed concerned the Chaos and Eldar.
First, here's a picture I took, and then, I will explain:
http://img413.images...26/eldar2hj.jpgBy around 15:00 minutes, up to about 16:30 minutes, the Chaos faction attacked the main Eldar base (earlier, harassing from the Eldar against the Chaos must have pissed them off, hehe). And, at one point, the Eldar's Webway Assembly was to the mercy of a single Chaos Defiler. It was about to be destroyed, basically wiping out the Eldar as well. But,
the Chaos forces present within the main Eldar base suddenly stopped attacking the Eldar structures, including the Webway Assembly, which only had just a bit more than 1000 hit points remaining. Those Chaos forces turned south to face a reinforcing Imperial Guard army coming to the help the Eldar.
Technically, it is a good response, from Chaos. But, considering that the Webway Assembly was getting owned, that Defiler should not have left its "priority", and resumed the destruction of it.
That single mistake, I believe, allowed the Eldar to resume building up their forces again, and maintain the forces present in the match even. It stayed 3 Vs 3, but it could have suddenly turned in a 2 Vs 3 match, in favor of the enemy. And it happened indeed. Eventually the Eldar repaired their Assembly, and built a new army over time. So, I'm not 100% certain, but I think we won the game because even though the Imperial Guard did most of the job, they could surely not take on both Orks and Chaos by themselves. The fact that the Eldar survived that attack was nothing short but pure luck, simply due to, as I believe so, an issue regarding priorities against/for targets when they are in the process of being part of history by the hands of a mechanical demon.
Now, a second issue I noticed, which concerns one of the Ork faction, namely Comp. #5.
Although that very specific Ork faction was the one in the whole game helping one of its ally the most out of any other A.I players, they did make a mistake, resulting perhaps in the survival (potentially) of my main base, again,
caused by the sudden needs of reinforcing.
Here's a set of pictures, and then, I will explain: 1)
http://img419.images...ttacking4lj.jpg 2)
http://img247.images...sretreat3it.jpgIn picture #1, as you can see, that Ork faction finally decided to attack my Listening Post position on the west side of my main base, near the Relic. That attack was probably a direct retaliation from the assault I myself ordered at their base (but failed) just prior to that moment. Suffice it to say, they owned that location, and
could have taken it for themselves very easily following the successful strike.
Picture #2 shows what I think was a mistake, or an issue. The whole of the striking Ork forces suddenly turned around, and went north to help their Chaos ally which were under attack by the Imperial Guards.
Technically, the reinforcement itself was awesome, and successfuly pushed the Imperial Guards back in their territory, allowing the Chaos to regenerate their base and forces. That, in itself, when I saw it, really impressed me.
But, the "problem" within that great A.I support feature as it was clearly shown especially during that very moment in the game (it was in my humble opinion the best reinforcing army I ever saw during a test), was that the Ork faction in question
sent everything he had. He could, and in my opinion should have split its forces in two parts. One resuming their assault up to my own base, and the other reinforcing its ally that needs help.
If even 30% or so of that attacking force coming from that Ork faction alone, decided to resume their way up to my base, believe me, they could have caused some serious damage, and stop me from getting anywhere outside of my base. Technically, they probably could not have managed to actually destroy the whole of my base, since I had some Dreadnoughts ready (two, I think, and third on its way), but it would have forced me to rebuild one or more destroyed Listening Post(s) in the process, as well as perhaps one or two critical structures.
And that Listening Post location they initially attacked (north of my Relic, as shown in the picture #2)
remained in my control, even though they destoyed the guarding structure protecting it (I believe it was level 1). They could have sent infantry to take control of it, but they never did it.
A third issue would concern a Squiggoth, which I think belonged to Ork faction #5. It happened at around 30:00 minutes or so (maybe 31:00 or 32:00, I'm not sure). An Ork army was reinforcing the Chaos faction, and in the process of reinforcing it, that Ork army continued their way up to the Eldar base, constantly pushing back the Eldar and Imperial Guard forces. And, while doing so, a portion of that Ork army, including the Squiggoth in question, was trying to pass through the Eldar base, without actually attacking anything in it, to eventually resume its way south of the Eldar base, only to go back west, and then north, back to the Chaos base. As if the Squiggoth had some sort of priority or primary target that wasn't on its very path.
So, by trying to reach "point B", within enemy territory, the poor beast eventually died, without giving any fight whatsoever. If I remember well, a few Imperial Guards tanks and a few other units were on its very way, and were happily shooting it.
I'm not sure if it'd be possible to do some kind of a "
constant potential priority/target change" system for the A.I, but it seems to me that it happens often to see one or more units, especially Relic units, having in mind one specific thing to do, and can't do anything else while doing that task, putting it to high risks of being attacked on its way to that task/location/target. That Squiggoth was a great example of that problem. Because I do believe it's a problem. It just seemed too obvious that it had something to do elsewhere, and didn't really mind about dying on its way without doing anything to defend itself.
Well, I think it's all I had to say about that test for the moment.
I know I am forgeting a few things I wanted to say as I type this line, but I can't remember exactly what I saw and wanted to share. That's why I include the replay. Please, take a look at it. It shows a lot of things, and serves as a great reference for Beta 3 testing in my opinion.
I might watch the replay again and edit this post back if I think I should share something important that I missed during this report.
Oh, and Arkhan, I'd like to reply to a part of your own reply concerning my first report, and here I quote:
The critical at the east side was not captured because you left there a scout squad as guards.
Do you think that it'd be a reason enough so that the Chaos never ever tried to go there and kill the mere Scouts and send those weak units back to their mom ? Tecnically I do not think that because one or even more of any infantry "guarding a CL" would be an actual reason good enough to justify the total absence of desire from the enemy to at least try and get it back for them. Unless of course they have much more important priorities elsewhere, which I think it wasn't really the case in that specific test I made the report about. But I guess that's just me. You're the coder, I'm just here to report what I see and experience.
So, as a conclusion, I'd say that:
A) Beta 3's Allied Support, overall, is indeed better than any of the previous builds, in my opinion at least.
B) There might be some room for improvements regarding "decision making" in terms of targets/priorities (taking the "Squiggoth issue" as a reference in that very test).
C) Although "Allied Support" itself is technically better and more responsive, I'd say that a great disadvantage of it remains in the fact that
when a faction decides to send forces to reinforce a specific ally, and that those very forces were previously in the process of destroying a base or a structure, then the reinforcing army should be able to split its forces in at least two parts, one to resume the "previous goal prior to reinforcing", and the other to actually leave its previous goal for the new one, which is to of course help its ally.
You know, not necessarily a 50/50 plit with perfect precision. But perhaps something like "
alright, I got three Tanks and two infantry Squads, my ally is under attack by Tanks, and this base I was destroying offers no resistance, so I will send all my Tanks to help my ally, and leave my infantry here to finish the job I already started".
How hard would it be to do that, I don't know. Would it be actually possible to do it in a technical point of view ? I don't know. But I do suggest such a behavior, because it could make a Win/Loose difference, simply said.
Alright, enough talk.
I think this report shows everything necessary to properly analyse Beta 3.
Hope you guys like it !
Edited by Zenoth, 07 June 2006 - 11:19 PM.
