76 replies to this topic

[thudo]

Mines could be done (just increase their required cost in the AI script to build em) but again I ask can the AI really use em effectively anyway? Its better to throw the money at units and be present on the field of war.

Now..

I need to be better at producing a TO-DO list for the project to keep the focus on objectives. We've done sooo much and there is just a little more to be done (until Relic steps in but thats here nor there.. we must assume we're on our own). So here is a quick list..

1] Resolve some issues where a squad sticks around at a strat/critical point and sits there even while his base is being destroyed.

2] Tweak jump/melee combat (although its been optimized so much), Excedrin pointed out some things about it which could be better.

3] Determine why we're getting more builders that leave buildings unbuilt right at the start. Larkin, does this happen also in your DoWPro? I don't remember this being such a deal although it is not too often these days it will happen but it does more frequently then back in 1.5 as I recall.

4] Force the Ork AI to build its Fortress Addon. This is getting cronic - each game I play the Orks use all their requisition to feed its military but 9/10 times never get to the Looted Tank/Squigg. Back in 1.5 we had a script that set aside cash which reserved requisition for a serious addon since the AI might be hardpressed to get it done. Problem is: Orks are terrible at making requisition, back before it was power. We need to set aside some requisition for the Ork HQ Addon so they go the distance like all other factions.

5] Commissar bug when using Execute still there? I thought back in 1.5 or 1.6 it was fixed yet its back?

Regardless of the challenges, the team has superseded the goals of the project and with now the prospect of a 2nd expansion coming in the Fall how can this not get any more exciting?!!

[LarkinVB]

I think early weak AI game should be improved as next objective.

The AI is very weak at map control and has no idea how to harass.
I'd like to see a harassplan like this :

Early scout/raptor squad moving to enemy points near their base. They attack capping units and automatically decap afterwards even while under attack.

AI should also be able to counter same tactic be using its first fighting squad/commander to protect their cappers instead if idleing around at gamestart.

@thudo : I didn't noticed unfinished buildings with DoWpro yet but have to admitt that I didn't watch many replays yet.
If you can show me a replay I might see some pattern. Builders code got some heavy changes since 1.5.

Edited by LarkinVB, 24 February 2006 - 08:33 AM.

[ThetaOrion]

Mines could be done (just increase their required cost in the AI script to build em) but again I ask can the AI really use em effectively anyway? Its better to throw the money at units and be present on the field of war.

--

The primary goal is indeed to make the mines the last thing on the list of priorities. The intent is for the AI to lay mines AFTER it has built all that it is going to build, after it has capped out 20/20 or better, and after it's workers or builders are just standing around doing nothing while the resources move up towards infinity -- then let the AI lay mines.

1) Increase their required cost to build the mines, or
2) Is there any way to make mines a Tier 4 object, make it so that all the research has to be done before any mines can be laid? Give the mines the same research prerequisites as the BaneBlade and the other Uber Titan Avatars.

The 1.60 AI was laying mines well enough. Obviously, if mines in uber abundance cause CTD's, then making them expensive and making them available after Tier 4 should cut down on their abundance. But, the AI should be allowed to lay mines after all else has been done. The mines are still a good way to encourage the human player not to be too slow to move out.


--
||
--

but 9/10 times never get to the Looted Tank/Squigg. Back in 1.5 we had a script that set aside cash which reserved requisition for a serious addon since the AI might be hardpressed to get it done.

In the huge 8 player map I played yesterday, the Orks had 16/20 on their tanks, started making a Squig, and then meanwhile they pumped out two or three more looted tanks while the Squig was building. They had at least three of Da Mek Shops. Anyway, the vehicle cap was 20/20 before the Squig got done, one looted tank after another popped out in the mean time. In this case, it wasn't a matter of cash. The Ork resources were moving up towards 10000 and power up above 5000. The gretchin could have started laying mines with all those reserves. The problem was that the AI didn't leave enough space or room in the caps to squeeze in the Squiggoth. It put out Looted Tanks instead and hit 20/20 that way, and the Squig lay dormant in Da Mek Shop.


--
||
--

5] Commissar bug when using Execute still there? I thought back in 1.5 or 1.6 it was fixed yet its back?

I kind of like your to do list. You have most everything on there that's getting my attention.

One more for the list, though, is that the IG Commissars and Priests, or Ork Warbosses, or SM Commanders, and even the Eldar FarSeer don't ever seem to be attaching up to any squads in Beta 11.

All of the AI Commanders that I see in all the different games are there singlely. None of them are attached to squads. As far as I can tell, none of the Commanders are attaching to any squads ever in Beta 11. Maybe I overlooked it, but every time I remember seeing a Beta 11 Commander of any sort, is is there fighting alone, even though there are squads nearby that it could be attached to.

[thudo]

The AI is very weak at map control and has no idea how to harass.
I'd like to see a harassplan like this :

Early scout/raptor squad moving to enemy points near their base. They attack capping units and automatically decap afterwards even while under attack.

AI should also be able to counter same tactic be using its first fighting squad/commander to protect their cappers instead if idleing around at gamestart.

Hmmm.. wouldn't that make the AI too perfect and thus you'd never win a game? I think for skill levels HARD and up.. perhaps but not anything lower. That early-game harass code would be like the worst way to starve a player right off and might be too good. No?

As for Mine laying.. not a priority for our project. I'll discuss with Arkhan if we need additional scripts for it or if its feasible.

As for that large map you played, Theta..
1) We cannot control build spacing. Thats in the RDGs for the buildings. The AI built the Squigg in the MekShop within his base or had obstacles around it. That happens.
2) Ork AI got to the HQ Addon probably due to the fact it was left alone. This doesn't count. Ork AI *MUST ALWAYS* reach the HQ addon if it gets to meet the prerequisites.

As for attaching.. do you see Commanders and other attachable units (ie. Apothacaries) not attaching or just Commanders? Guy.. I see ALL attachable units including Commanders attaching all the time.

[ThetaOrion]

Hmmm.. wouldn't that make the AI too perfect and thus you'd never win a game?

Even I, after losing ten games in a row, tend to get discouraged and give up trying. Then it's time to try the STANDARD setting on that losing scenario, or it's time to try a different scenario. So, yes, you definitely don't want to do something that guarantees an AI win every time, at least not at the HARD setting. The INSANE setting might be a different story. If a human can never win at the INSANE setting, that just might be what the INSANE setting is all about.

--

Guy.. I see ALL attachable units including Commanders attaching all the time.

Hmm, with a human in the mix, or just AI vs AI? Every attachable unit I see seems to be there fighting alone, but I tried to make it a point to emphasize that it could indeed be my memory playing tricks on me. It's an issue that I think everyone should start keeping an eye on, especially in regards to the games where there is a human in the mix. When I kill those priests, FarSeers, SM Commanders, they are always dying alone and unattached. And, I haven't seen any IG Commissars in Beta 11. They are either not being produced or they are dying in base in Beta 11, or dying where I am not looking. Nothing Commissar attached to Guardsmen squads which makes those squads really easy to kill.

The AI built the Squigg in the MekShop within his base or had obstacles around it. That happens.

In the past Beta 2, or thereabouts, I saw Squiggoth trapped in their base. In this replay that I watched, there was space and place for the Squiggoth to pop out. There was room off to the right of the building where nothing was built. It had space. The problem was that by the time the Squiggoth was finished building, two or three more Looted Leman russ popped out of the other Mek Shops so the 20/20 vehicle cap was reached in the mean time, hence nothing at all came out of that very long production that I watched.

With the Chaos, if their caps are 20/20 on their troops, and you produce a BloodThirster, the caps go temporarily 25/20 to make room for the BloodThirster. Can't something be done like that for the Squiggoth and the Ork Vehicles. Something to let the already produced Squiggoth pop out of Da Mek Shop. If he is stuck in his base after getting out of the Mek Shop, that's another problem, and there's nothing you can do about that one. But, Squigg will be there to defend the base if it is ever attacked.

2) Ork AI got to the HQ Addon probably due to the fact it was left alone. This doesn't count. Ork AI *MUST ALWAYS* reach the HQ addon if it gets to meet the prerequisites.

Agreed.

As for Mine laying.. not a priority for our project. I'll discuss with Arkhan if we need additional scripts for it or if its feasible.

Mine laying is not a priority for any project, and that's the way it should be, imho.

But, if you can make them Tier 4 objects requiring all the research to be done first and make them rather expensive, that should put them into their proper place or the proper perspective -- last on the list of priorities, or last in importance.

LarkinVB said he would look into mines for DoWPro 2.0 -- one team or the other will find a way to work them in but not make them such a prolific annoyance. Mines there, but not overly abundantly so.

[Excedrin]

I'd love to see an AI that plays more human, with a serious fight for early game map control. If it played well (say at the level of a 1400 rated human) then it's probably impossible to beat on Harder or Insane. Along those lines, it would suck to have the Hard (equal res) level play weaker or NOT fight for the map. I don't care if Standard or Easy play weaker, but the resource rate is probably enough to make Standard easy to beat even if it fights early.

It almost seems like the AI should branch off an "expert" version that's intended for play on the Hard setting only, and that tries to slowly accumulate advantages the way a strong human player does. The goals are the same even if it tries to achieve them with early harass or a full out rush or defensive teching play, depending on the AI's evaluation of circumstances (number of SPs, race matchup, current map control / army size / tech level, etc).

Another point, if the AI is too good, that's really a GREAT problem to have!

Edited by Excedrin, 25 February 2006 - 12:06 AM.

[Finaldeath]

Hmmm.. wouldn't that make the AI too perfect and thus you'd never win a game? I think for skill levels HARD and up.. perhaps but not anything lower. That early-game harass code would be like the worst way to starve a player right off and might be too good. No?

The AI is never impossible to beat, except when it might have more resources and a map advantage.

Trust me, it mis-manages troops a lot. I can beat it in most games unless it is a very large, 8 player game, and an AI player has a chance to really build up.

Having commanders idling is just bad. Even bad players won't do that with commanders, they move them up and have them guard...

A plan to specifically harass might be a nice *optional or random* part of the AI too.

The commanders idling is always problem.

They also don't alway attach. I've seen it much too often them running around when they could attach, or should. I recall you adding code to make sure apothecarys always attach (and mad doks) which of course makes a lot of sense

Try adding some code to get commanders to attach to big squads - mekboys, warbosses = nobz, librarians, force commanders = any terminators, chaos lord, sorceror = khorn beserks (unlimited morale!) or possessed marines, farseer = seer council - no ifs! and IG can't attach, of course .

When a commander, late or early game, dies from just not attaching to a big unit and thus running off into CC alone, it's not a good idea to keep the current code, since commanders can make and break games.

Some early units also are very good units to attach too. Mekboys attaching to large Ork mobz to use teleport, for instance. Chaos lords and the comparitivly weak sorcerors should always attach to khorne beserkers, who are melee and immune to morale damage And Farseers should never not attach to seer councils, since they get a bonus recharge rate to their very powerful abilities.

Early early game, with small squads, there should be a good chance of using less melee orientated troops as fodder. Chaos lords can use Cultists, Force commanders can use scouts (who are cheapish), or space marines even, Mekboys of course can use any sluggas, and farseers *can* benfit from being in a ranged squad since they are the only primary/first hero to get very good ranged abilties in the form of spells.

All of the above could and should be taken into account at full, not half, health. Joining a squad at full health gives the squad greater survival - as morale increases - and helps the commander - as it forces the commander to have some cannon fodder around it.

I'd love to see the AI impliment better, slightly more complex, commander attaching code. Some commanders should always attach, for good reasons. Mainly, this lets them keep with the army when they usually run off unsupported, but in other cases greatly helps the troops they are with, and vice versa (eg: librarians "Will of the Emporer", ie: invincibility, when lib stays alive, will be more effective if he groups with troops before using it!)

Hope this is taken into account, at least for the next version. If not, I'll take a look at the code

[ThetaOrion]

Yes, FinalDeath got it right. Hear! Hear!

Just to add:

The Eldar FarSeer attached to the first squad of Banshees is the game's best early rush combination that there is.

I have taken my FarSeer attached to a squad of Banshees, and I have had that squad standing in the back of a Chaos base the whole game harrassing, and I can replenish the lost units faster than Chaos can destroy them. I can't do that with any other combination of early opening game rush units. With other factions the squad or the Commander dies, but FS attached to Banshees can hold out through the whole game. Pretty amazing.

Of course, once the Eldar get the Seer Council, the FS Commander attached to that is awesome as well, but that combination usually becomes available only later in the game, somewhere around the middle.

--

Anyway, the Beta 11 rushes that the AI does have been easy for me to repel because there is hardly a single commander attached to anything. The rushing squads and the rushing commanders are all easier to kill because they are all unattached. I might have seen a Big Mek attached to Sluggas during Beta 11, but so far that's the only attachment I remember. But, I have seen dozens of unattached Commanders from all the different factions.

[thudo]

The reason for Commanders NOT attaching to squads right away is due to the fact its been always coded that way since 1.3 or earlier. Larkin and us on the team decided that attaching would occur only when the Commander was at a certain health percentage then retreat to find an attachable squad. Yeaaa sometimes I wish I'd see far more attaching going on so the Commander is NOT left idle by itself. I'll chat with Arkhan and Larkin on this. I would like a good script that allows more reliable attachments although we need to discuss this.

[ThetaOrion]

I haven't been playing it every free moment of every day, but my wishlist for Beta 11 seems much smaller than it was for Beta 7.

There just doesn't feel like there as many things that need to go on the to-do list now.

Thudo: I need to be better at producing a TO-DO list for the project to keep the focus on objectives.

Thudo Replies:

6] I'll chat with Arkhan and Larkin on this. I would like a good script that allows more reliable attachments although we need to discuss this.

7] As for Mine laying.. not a priority for our project. I'll discuss with Arkhan if we need additional scripts for it or if its feasible. (Yes, last on any list of priorities. But check to see if mines can be turned into Tier 4 objects.)

--

Anyway, that's about seven items on your to-do list. And, Larkin could probably add seven more if you gave him enough replays to watch. But, it looks like it's starting to close in on the end for a finished 1.70. Thumbs up and keeping fingers crossed.

[LarkinVB]

About attaching commanders. Don't go overboard with this. There are lots of situations were it is good not to attach as the commander can tie another ranged squads nicely. I suggest a good analysis of the situation before attaching.

[Finaldeath]

LarkinVB, you may be right, but since 1.4, or was it 1.3, of the game, if yuo go into melee near two squads, and engage one, the other will usually automatically get attacked too.

And, remember, some heroes are ranged based, or gain good abilties attaching to squads (eg: teleportation, ability boosts or morale immunity, ability to use spells faster).

The AI rarely uses the commander to attack other squads, mearly "whatever comes first" the AI attack code isn't too good, we all know this

I'll look into the code if I have time. I'd not mind checking out this particular "problem".

[LarkinVB]

I'll look into the code if I have time. I'd not mind checking out this particular "problem".

Very good. I'm sure you will find a good allround solution.

[thudo]

Thanks FD, Larkin..

[ThetaOrion]

There's no thanks in it, but in keeping with my current one huge epic game per weekend, I played as 3 Orks vs 5 Space Marines at the HARD setting on Aralez's new DrawBridge_Recreated map. Beta 11 was used.

I was a four hour game. It played out pretty much as I would have expected 1.60 Official Release to play out!!! . . . except there were no mines built by the AI.

My two Ork AI allies were able to keep the Space Marines out of the castle for the whole four hours, while I harrassed the Space Marines outside.

One of my Ork AI allies built a Squiggoth. I, the human, went through about five or six Squiggoth myself. I built a secondary base outside the gates and launched Looted Leman Russ and Squiggs and Nobs from there.

And this time, with Beta 11, the Space Marines weren't the ultra uber gods that they were back in Beta 8. The Space Marines were normal just like they are in 1.60 Official Release of the AI Mod, hard to beat but not impossible to beat at the HARD setting if you keep coming back and keep applying various sorts of cunning or persistence.

Needless to say, but I will say it anyway, I like Beta 11 better than I liked Beta 8. Beta 11 feels more balanced and/or plays out more normally on these huge epic games. Beta 11 feels like a better AI for these huge unapproved maps, and that's a good thing in my not so humble opinion. I was very pleased with the way the Beta 11 AI performed in this particular game. I have played a lot of these 3vs5 games on various versions of these siege maps over the months, and in the couple of huge epics I have played, 1.65 Beta 11 has felt like everything is back to 1.60 Official Release normal on these huge maps, except for the missing mines of course.

[thudo]

Well.. that was a ringering endorsement if I don't say so myself.

Thanks Theta.. Morale Boosting Twinkie for us!

[ThetaOrion]

Yes, very much a ringing endorsement.

Except for the missing mines that I know would have been there in the 1.60 AI Mod Official Release, I would have guessed that I was indeed playing 1.60 when in fact I was actually playing Beta 11.

I know it was maybe only small changes, but they made a big difference. Kudos to you all!!

The missing mines aren't a big deal, the mines are just a distinquishing mark or a sign that you are playing 1.60 and not something subsequent. In fact, a lot of mines in Beta 11 would have made the scrolling even more choppy on that DrawBridge Recreated map, so missing mines or sparse mines is a good thing overall. Still, if you can somehow make the future 1.70 AI Mod mines very expensive and make them Tier 4 objects so that the mines aren't built until after everything else is done, a few AI created mines here and there in this Beta 11 would have indeed made it so that I couldn't really tell the difference between Beta 11 and the 1.60 Official Release.

There is one other sign that lets me know it was Beta 11 and not 1.60 AI Skirmish Mod. This is kind of convoluted or complex indirect reasoning, but here goes. Aralez played the same 3Ork vs 5SM scenario with the 1.60 AI Mod Official Release and a beta version of the map a couple of weeks ago, and he won the map in 40 minutes at the 1.60 HARD setting. Aralez is a better and faster moving player than I am, but that doesn't account for everything or explain why it took me 4 hours to do the same thing. Instead, this gives me possibly another feel for the slight differences between 1.60 and Beta 11. I don't think the Beta 11 Orks are quite as efficient as the 1.60 Orks were, and it all goes back to your TO-DO list item numbers 4 and 6 the HQ Fortress Addon and Commander things and the number of Squiggs produced. 1.60 Official Release AI Orks would have reliably produced more Squiggs and produced them faster or earlier, and I also sincerely believe that you would have seen more Ork Commanders attached to squads in 1.60 of the AI Mod. Not every commander has to be attached up, but certainly more than what we are currently getting in Beta 11.

Slight differences there, but you have them on your TO-DO list now, and the solutions will show up in a future beta. FinalDeath resolves the attaching thing, Thudo gets the Ork HQ working right, Arkhan or Larkin looking at making mines useful again in Tier 4, and then nobody will be able to tell the difference between 1.60 and 1.70, and/or 1.70 will be at least as satsifying for the typicial user on the huge maps as 1.60 is. I keep testing and comparing, because I want 1.70 to handle these large maps just as well as 1.60 did.

I certainly didn't want to destroy morale around here. I just wanted 1.70 to be as satisfying on all the maps as 1.60 was. I'm not expecting anything special or anything custom scripted for these huge maps. I'm just wanting 1.70 to play them as well as 1.60 did. And, Beta 11 is pretty much there!!

Edited by ThetaOrion, 26 February 2006 - 12:22 AM.

[ThetaOrion]

BTW, Beta 11 Space Marines will build a LandSpeeder. With Beta 11, I have more than once had to quit what I was doing and wipe out the LandSpeeder because the pesky thing was dancing around and doing too much damage to my squads. Wonderful!!

Also, that Ork WarTruk is now showing up in every Beta 11 game with the Ork AI, and that thing is practically indestructible, aka hard to kill, and it seems to have some teeth against infantry. Wonderful!!

A single Tier 1 IG Chimera is often very brutal against Tier 1 enemy infantry, but not of much use thereafter.

Over there, we got off on a tangent and got to discussing this topic over at Relic forums in the DoWPro topic.

This is what Aralez recently had to say about Eldar Rangers and the removal of questionable units:
http://forums.relicn...0&postcount=128

And, here's the overall discussion about removed items in case it proves useful or referential to anyone:

http://forums.relicn...68&page=7&pp=15

In one of the games I played recently, the Inquistor's Fortress Map with 1.60 of the AI Mod, you got about four of those worthless SM Rhinos given to you for free at the start of the game. I actually ended up placing them there as cannon fodder, and the enemy attacked them, while my ranged Space Marines attacked the enemy. It was like having invulnerable Space Marines for as long as the Rhinos lasted, and they all but one lasted until the end of the game, iirc. Anyway, a Rhino as tough as a Squiggoth might even have its uses, but then again, this is with a human in the mix, running the Rhinos and the Space Marines. I definitely wasn't using the Rhinos as intended. I never loaded anything into them. I used them to draw fire. But, I don't know how you would get the AI to put its Rhino between its troops and the attacking enemy so the Rhino can absorb all the hits or damage while the ranged SM standing outside of the Rhino actually attack the enemy while simultaneously being invulnerable thanks to the Rhino.

As concluded in the DoWPro thread up above, the AI can't make proper use of some things like the Rhino, so the AI shouldn't even build the Rhino, but there are other undesirable 1.41 Plain Vanilla units that a Super AI can be made to use properly, such as the LandSpeeder, the Ork WarTruk, the IG Chimera, and Mines, and the Eldar Rangers. When it comes to worthless or undesirable items, as Aralez said in his post, it's better to limit their numbers than to completely eliminate them.

Edited by ThetaOrion, 26 February 2006 - 03:03 AM.

[LarkinVB]

Reading your posts I get the impression that v1.6 played still better (at least not worse) than b11 on these big maps. Is this true. ?
Arkhan made lots of changes for b11, I thought some of them would make b11 stronger than 1.6.

I answered Aralaz post in the relicnews thread. His arguments were not that strong and convincing.
And yes, I know you are missing mines

Edited by LarkinVB, 26 February 2006 - 09:14 AM.

[ThetaOrion]

Reading your posts I get the impression that v1.6 played still better (at least not worse) than b11 on these big maps. Is this true. ?

Well, we are back again to the highly subjective, and in this case, there are two different humans in the mix, which makes the guessing even worse. It's hearsay of hearsay. It was Aralez playing the v1.6 and I playing the Beta 11, with me speculating on the results. The only real similarity was the map. Yes, I am working my past extensive experience with these siege maps into the mix or the feel as well, but it's still speculation. Nevertheless, Beta 11 played out as I expected or hoped that it would play out. Yay!! Kudos!! Good stuff!! I went in with no expectations, just wanting to play a game, but I came out very satisfied. Beta 4 and Beta 8 wouldn't have played out as I expected or hoped -- those diverged too much from the feel and the expectations of v1.6, but again, that is only my opinion, though I believe I'm right.

Nevertheless, I too was of the opinion that -dev made a real difference, and we all know where that went.

The thing I can say for sure with some great confidence after my Beta 11 tests is that Thudo's TO-DO list, including item 6 about the attaching commanders that I added, is indeed where their attention should be focussed for the next betas.

--

Arkhan made lots of changes for b11, I thought some of them would make b11 stronger than 1.6.

My guess is that in some places B11 is indeed stronger than 1.6, and there are still probably places where 1.6 is stronger than B11. But point in fact, all this speculative guessing could be taken in lots of directions. It's possible that the B11 Orks are weaker than 1.6 Orks, or that B11 SM are slightly stronger than 1.6 SM, or vice versa if the Mountain Trails tests I made are taken into consideration.

The only way to know for sure is to Cage Match B11 against v1.6, AI vs AI. Then you completely remove the subjective human speculation from the thing and only the pure data comes through. The only problem with the Cage Matches is the fact that the AI often seems to play totally different than when a human is in the mix. Humans do silly or ingenius random stuff at random times that the AI just isn't programmed to handle, and thus things can easily go off onto unforeseen tangents. I guess there's also the possibility that the interpretation of the Cage Match data could also be subjective, due to the fact that there will be a human interpreting the data, when the thing is studied. But, the only way to know for sure if B11 is stronger on average than v1.6 is to Cage Match them against each other and let the data fall out as it will.

I can understand the fear of the Cage Test, though, because the human reaction is to want to demand that Beta 11 always win. But with randomization in effect, Beta 11 won't always win, even if it is the better AI. But if they can get past the human need for wanting Beta 11 to always win, then they can use the comparisons between two AI to further enhance the TO-DO list or the list of priorities that they want to work on. They'll see things that they like, and things that they don't. And, some things will go higher on the list of priorities or higher on the TO-DO list than they currently are, and some things will be added to the lists. And, other things might be removed from the list, because an AI vs AI comparison will make them appear not as critical as the humans think them to be.

And, there's a whole other aspect to this specific issue. In many respects, I don't care if B11 is stronger or better than v1.60. What is more important to me, the subjective human in the mix, is that B11 feels and plays out more balanced and more satisfying on average than Beta 8 and Beta 4 were doing. With Beta 4 and Beta 8, I was feeling frustration and dissatisfaction after playing a game, so much so that I deleted Beta 4 and went back to Beta 3. So much so, that I made a big stink about Beta 8. They just didn't feel right. Beta 4 felt broken. And, there was too much imbalance and too much dissatisfaction on my part with Beta 8. I wasn't enjoying the games that I played with Beta 8, especially if I chose the wrong faction as my AI opponent or my AI ally. The 1.60 Official Release provides many satisfying games to a subjective feeling human, and so far Beta 11 is doing the same. There's a feeling of harmony or a feeling of balance with Beta 11 and v1.6, that are very important for the times when there is a human in the mix. The B11 games are pretty much consistently satisfying with a human in the mix. It's not perfect or finished, of course, but it's on the right track. With Beta 11 and v1.6, you do not come away feeling that the AI is broken or feeling that the AI has an unfair advantage at the HARD setting. With Beta 11 and v1.6, you want to play them. But with some of the previous betas, I wanted to stop playing them.

--

And yes, I know you are missing mines.

Not any more than I'm missing the Eldar Ranger or was missing the LandSpeeder, WarTruk, Vyper, Chaos Horror Demons, IG Chimera, or Titan Uber Avatars when they were missing.

LarkinVB, you promised me you would look into mines for DoWPro 2.0, and I will satisfy myself with that. If you succeed there in creating an impressive acceptable workaround to the overabundance of mines once the AI gets going on them, Arkhan and Thud will be wanting to borrow it, and the mines will then find their way back into the AI Mod.

Edited by ThetaOrion, 26 February 2006 - 10:26 PM.