76 replies to this topic

[LarkinVB]

But, if you make the HARD setting completely impossible to beat, whether here or over at DoWPro, I and others will fight that, I can pretty much promise you that. Only two or three percent of the population would truly like that. I mean, if HARD setting is completely impossible to beat, then what's the purpose of HARDER and INSANE? They have no purpose, because they too will only be more impossible to beat. There's no logic in that!!

Do you care to listen ? I would like HARD to be hard for good players. We are not talking about very good players here, they will beat the AI at all settings regardless of what we will improve. Currently HARD is hard only for low to average players. Nobody is talking about making it impossible to win. This is not possible due to AI limits even if we try to.

Most DoWpro players (which care to write in the forum) find hard too easy to beat at gamestart so I like to see that improved. Is this so hard to swallow ? You can fight for whatever you want. Good luck with that.

Edited by LarkinVB, 28 February 2006 - 12:50 PM.

[thudo]

This should not be the target audience of the mod. The audience should be the 40% average-to-good players willing to practice if they can't immediatly beat the mod at hard.

Whether you like it or not, the project will be used by EVERYONE in the community for every mod out there. The project is meant to be the defacto Skirmish AI thus 90% of our "clientele", if you will, will be on the lower scale of the playing bracket. Our AI is meant for every mod coming out like your DoWPro AI is very likely the "AI of the future". Hopefully, EASY and STANDARD skills are now taken care of to bring people up and increase their skills from the lower level to higher.

Btw, I am NOT a dictator, Mr. Theta.. I just want to focus people via clean debate. Sometimes coordinating people and making decisions that are "unpopular" is part of the resume. However, we all are MINI-LEADERS here and thus have a stake in how we evolve. This is what it should be BUT we need someone to delegate and concentrate the duties so we accomplish our goals. Also, HARD and up use the same settings. Its the game that grants the AI more resources at HIGHER and up. That should stay the same. HARD should be using the same code as HARDER and up.

Further, you want STANDARD to have only one buildprogram? Whats wrong with it now? Its fine with 4. EASY should stick with 2.

[LarkinVB]

Summary : The AI harass/rush will be not added to HARD as it might alienate the masses because the newbs will loose most of their games. Is this correct ?

Edited by LarkinVB, 28 February 2006 - 04:09 PM.

[thudo]

The AI harass/rush will be not added to HARD as it might alienate the masses because the newbs will loose most of their games. Is this correct ?

Ahhhhh no. On the contrary, all code should be there for HARD and up. Its EASY that should be withheld just to get people into the project and NOT frustrate them. Heck.. even with the improvements, EASY is already "subdued" so we don't need to further retard it anyway. So nope.. no special treatments. Full steam ahead!

[ArkhanTheBlack]

Since my exams are over now, I have some time this week to solve the last problems and improve the early game. We'll see if it's even worth to 'restrict' this improvements to hard+.

BTW, speaking of difficulty levels... In my opinion, the probem of lower and higher difficulty levels are the starting ressources. Regardless of easy or insane, all players/AI's ge 600 starting ressources. This is even for an easy enough to make a successfull rush on a noob, regardless of any further ressource penalties. It's also the reason why insane is still easy to beat in the first minutes. They simply need at least a few minutes to really get a recognizable 'profit' from their ressource boni.
Therefore I'm thinking about including a 10 minutes attack restriction for easy and crippling their economy a bit less. That would probably solve a lot of problems with easy AI 'too weak' or 'too strong' complaints.

[ThetaOrion]

Thudo wrote: Further, you want STANDARD to have only one buildprogram? Whats wrong with it now? Its fine with 4. EASY should stick with 2.

Nope, I thought that you, Thudo, made STANDARD so that it only has one build program. Not a good thing, if that is the case. If STANDARD has all four build programs, then STANDARD is as it should be. But, the last time this topic came up, Thudo, you told me you were removing build programs from the STANDARD setting. I didn't agree with it then, just like I don't agree with the removal of Uber Titans from any difficulty setting.

How many build programs does the current STANDARD setting in the AI Mod B11 really have? If it has them all, then the STANDARD setting is as it should be.

Theta Wrote: Thudo cut the build programs out of the STANDARD setting. Each faction only has one build program in the STANDARD setting, if I remember correctly.

This is what I said. I should have added a big NOT GOOD to the end of it. Of course, Thudo, you don't always fully read what is said, even if it is numbered and listed with numbers. But, I also was just accused of not listening. So, nobody is fully listening or reading. We are all just talking past each other hoping that something will stick.

--
||
--

Larkin Wrote: Do you care to listen ? I would like HARD to be hard for good players. We are not talking about very good players here, they will beat the AI at all settings regardless of what we will improve. Currently HARD is hard only for low to average players. Nobody is talking about making it impossible to win. This is not possible due to AI limits even if we try to.

Yes, I care to listen. That's why I kept pursuing this particular topic instead of just going away.

I pursued this topic so heavily because it is LarkinVB. I felt that he was reading and trying to understand. Of course, much of the problem is that I'm not a coder, so my vocabulary is different than theirs. I, too, had had to try to figure out what they were talking about.

But it's important, because Larkin will do the AI for DoWPro, so I felt it was very important to help him come to an understanding of my point of view. A couple years ago, I worked as a beta-test for a Rebalance Mod, where they were also adding in new weapons and new user created units. So, DoWPro is based upon design concepts that are near and dear to my heart. The design philosophy of DoWPro is closer to my own. So, there is a good chance that if Larkin does DoWPro right, then I might actually end up liking it better than the AI Skirmish Mod, and anything that I like better than the AI Skirmish Mod is going to be very special indeed.

But, if Larkin purposefully or accidentally makes the HARD setting of DoWPro impossible to beat, he's going to lose me, especially if the STANDARD setting of DoWPro is a true standard setting, because STANDARD no longer satisfies me when it comes to team play.

I too doubt that one single change will make the AI completely impossible to win, though, small changes can sometimes have a big impact as Arkhan and Larkin have tried to teach me in the past. But, if Larkin were to somehow find five or ten improvements to the actual AI that makes the HARD setting completely impossible to win, and implements them all, he'll lose me, and he should know it in advance. That's the point I'm trying to get across. I think he got the point when he wrote, "Nobody is talking about making it impossible to win." Although, Thudo is definitely talking about not wanting to make it impossible to win.

--
||
--

I think I'm listening and reading what's here:

Larkin wrote: Nobody is talking about making it impossible to win.

When Larkin said that, I cheered!! My point got a across. There is now hope that the HARD setting in DowPro will be what I want it to be.

Thudo has made up his mind, and I support Thudo's ideas of keeping the HARD setting of the AI Skirmish Mod where it has historically been. The HARD setting of the AI Mod has been popular in the past and is popular right now in B11. It should stay there.

Larkin wrote: Most DoWpro players (which care to write in the forum) find hard too easy to beat at gamestart so I like to see that improved. Is this so hard to swallow ?

Then, Larkin, you should make your changes over there in DoWPro, and make it the way you want it over there. Is that so hard to swallow? If the DoWPro players who care to write in the forum actually don't like your HARD setting, then yes, you should do something to change it or improve it.

Here, most of us like the HARD setting in the AI Mod as it was in 1.5 and 1.6, and as it is now in B11. So, Thudo shouldn't change it here in the AI Skirmish Mod, because we like it. Is that so hard to swallow?

Definitely, though, if people over at DoWPro hate your HARD setting, then you should indeed do something to change it. It's best to listen to the people, the people who are actually using or playing the thing.

Edited by ThetaOrion, 28 February 2006 - 11:39 PM.

[ThetaOrion]

Thudo wrote: Also, HARD and up use the same settings. Its the game that grants the AI more resources at HIGHER and up. That should stay the same. HARD should be using the same code as HARDER and up.

Also, for point of thoroughness and to show that I am really listening, I wanted to restate my opinion regarding these sentences, just in case I get a vote. The info is elsewhere, but this is the proper place to reiterate it.

1) "HARD and up use the same settings," and I believe likewise HARD and lower should also use the same AI. I believe that the AI should stay the same at all difficulty settings, even the EASY setting. Removing units and build programs creates confusion and questions, even at the EASY setting. People notice and start to wonder where they went. There's nothing wrong with limiting undesirable units, but it is wrong to completely eliminate them. Titan Uber Avatars should not be removed from any difficulty setting, even the EASY setting. It creates dead ends in the research and questions in the user. It's not a good training foundation to have unbuildable units or dead ends in the research, at any difficulty setting. The IG does the research, builds the Mars Production Facility, and wastes the resources, but then the IG can't build a BaneBlade. That's confusing, even at the EASY setting. The Titan Uber Avatars shouldn't be removed from any difficulty setting.

2) EASY and STANDARD should be using the same AI code as HARD, and HARDER and up.

3) Only two things should change between the different difficulty settings:

. . . a) The AI should be granted more resources at HIGHER and up, and the AI should be granted less resources at STANDARD and lower. This is as it is and it shouldn't change. The "game grants this" as Thudo just said.

. . . b) The TechBreak Modifier thing should be different at each of the different difficulty settings, as I believe it currently is in B11. That TechBreak Modifier worked and didn't create any confusion or dead ends or missing items in the AI at any difficulty setting.

4) Those are the only two things that should change from one difficulty setting to the next, resources and the TechBreak Modifier. Thus, the AI should be using the same code at all difficulty settings, and not just the same code at HARD and up.

5) If the AI is going to be the same code at the HARD setting and up, per Thudo's Executive Directive, then I think the AI code should be the same code at the HARD setting and down as well.

Edited by ThetaOrion, 01 March 2006 - 03:03 AM.

[LarkinVB]

Here, most of us like the HARD setting in the AI Mod as it was in 1.5 and 1.6, and as it is now in B11. So, Thudo shouldn't change it here in the AI Skirmish Mod, because we like it. Is that so hard to swallow?

Please define "most of us".
I know at least three people (me, Finaldeath and Excedrin) who would be happy if the early gameplay of HARD can be improved. Not sure about Arkhan as he cleverly stays clear of this discussion. So there must be at least 4 others constituting the majority.
Who are they ?

Whatever. I hope Arkhan can do some improvments and I'm sure you will like them. The AI will be stronger, it will be more fun and you will beat it after adapting your playstyle.

Edited by LarkinVB, 01 March 2006 - 09:16 AM.

[ThetaOrion]

Please define "most of us".

Ah, trust Larkin to think that Larkin, Arkhan, Excedrin, and FinalDeath are the only 'we' or 'us' in the universe, and thus the majority.

A coders only club.

My definition of 'most of us' per my messages above is all the thousands of people who downloaded 1.5 and 1.60 of the AI Skirmish Mod and played it. Most of us liked 1.5 and 1.60 just the way they were.

I hope Arkhan can do some improvments and I'm sure you will like them.

If you had a hand in B12, I do indeed like them. The AI does something different every time now. In B7 and B8, they gathered or pooled, and then came all at once into the human base and camped.

Now in B12, I have no idea what the AI is going to do. When they come, they seem to attack my ally as much as they attack me, and you never know what they are building or making, so you can't really prepare for it. There's no way to memorize a set strategy, and the SM are no longer way over powered. B12 and B11 are much more balanced feeling or pleasing feeling. You have to hustle and you have to tier up fast and you have to keep pressing, or you are going to lose at the HARD setting, but I don't feel like the B12 SM AI is cheating like I believed or felt the B8 SM was doing.

With B11 and B12, I have to adapt my play style every game, which is a good thing. I just never know what's coming next with a B12 game. I used to could predict how a game was going to play out after I got to know the version of the AI that I was playing, but not any more.

I think your goal of making the AI more fun is turning out just fine, if B12 is any indication.

[LarkinVB]

Ah, trust Larkin to think that Larkin, Arkhan, Excedrin, and FinalDeath are the only 'we' or 'us' in the universe, and thus the majority. lol.gif

A coders only club.

My definition of 'most of us' per my messages above is all the thousands of people who downloaded 1.5 and 1.60 of the AI Skirmish Mod and played it. Most of us liked 1.5 and 1.60 just the way they were.

I was counting only those who made a statement here in the forum. I don't think that me or you can speak for the masses.
No 'coders club' statement at all.

[Finaldeath]

You can go and read the several incarnations of the skirmish AI threads at relicnews to see a couple of hundred other people saying the AI still has bugs and should play much more completitivly.

[ThetaOrion]

You can go and read the several incarnations of the skirmish AI threads at relicnews to see a couple of hundred other people saying the AI still has bugs and should play much more completitivly.

Yes, and ThetaOrion was one of them, if you go and read RelicNews.

--

And, if you take the time to go and read up above, you will see that the point I was trying to make up above was that we don't want the HARD setting to be made impossible to win. Competetive and bug-free and fun AI are just fine for the HARD setting, but not impossible to win. They are completely different things. If HARD is made impossible to win, then what's the purpose of HARDER and INSANE . . . more impossible to win?

The INSANE setting is the setting that should be impossible to win. HARD setting should be basically possible to win ninety-five percent of the time if a person works hard or hustles to win and has the skills, just like HARD setting was in 1.60 and 1.50. And so far, B12 is still on track for the HARD setting just like 1.60 and 1.50 were in their day. None of them are impossible to win at HARD setting. HARD setting in B12 is possible to win. I have won it a bunch of times already, so all is on track. I also lost to B12 once so far. I believe that Thudo gets this, and I trust that Thudo won't ever let it get off track.

Again, everyone is just talking past each other trying to make their own point, and nobody is really reading or listening.

[ThetaOrion]

Whenever Thudo says that the HARD setting is the bread and butter of the AI Mod, the heart and soul of the AI Mod, or the core of the AI mod (something he has told us many times), and whenever Thudo says that the HARD setting should never be made impossible to win (as he suggested up above and elsewhere as well), I come away from the encounter thinking that Thudo is wise and that Thudo really knows what he is doing.

I also trust LarkinVB and Thudo when they tell me up above that the AI in the HARD setting has a very long ways to go before it becomes impossible to win. I also trusted and appreciated Larkin's input when he said that he's not talking about making the HARD setting impossible to win.

The B12 HARD setting is right where it should be. It's possible to win. But the AI in B12 is different, because I can't see a common theme, the games all play out differently, and I like that. I could tell you if I was playing B4 or B8. They had themes, recurrent flaws or the same routines. But, B12 and even B11 seemed to be a completely different AI Skirmish Mod each time I play them, and I really like that. I really like the diversity. So the AI has gotten better, funner, and more diverse in B12, but it's still possible to win at the HARD setting if you have the mad skills, just like 1.60 and 1.50 were possible to win at the HARD setting.

ALL is on track and where it should be!!

[LarkinVB]

Amen and topic closed.

Ooops, I'm no mod.

[ThetaOrion]

BTW, to fully understand my point of view on this subject, you need to understand the source of my concern.

In Beta 7 and Beta 8, there were scenarios at the HARD setting that were completely impossible to win, if you chose the wrong factions for you allies and enemies.

Whenever I chose SM as enemies and Orks as allies on the Mountain Trails Map, those scenarios were completely impossible to win, and it didn't change no matter how many times I restarted or how many times I replayed the scenario. And, it wasn't just because I was being a wimp.

A typical HARD setting on the Mountain Trails lasts 20 to 30 minutes before I win or lose.

But in Beta 7 and Beta 8, with Ork allies and SM enemies, the games were going an hour or longer before I would lose, and the best I did was a stalemate after two hours of play. Something was seriously wrong there. I mean, the Space Marines were showing up in my base with a LandRaider, fully tiered, and fully capped with Terminators Dreads and Commanders and everything, while I was just producing my second tank, had two squads in the field, and was trying to save up for Tier 3 research. It was messed up, and it was impossible to win. It played out the same way ten times in a row, and every time was impossible to win. The HARD setting shouldn't be impossible to win, and it shouldn't play out the same way ten times in a row. The only time it played differently was with -dev. B5, B6, B7, B8, they were all not right. There were scenarios that were completely impossible to win at the HARD setting, but only if I chose the wrong factions. I mean games going on an hour and a half just trying to survive. It felt like HARDER or INSANE, but it was the HARD setting. The thing was seriously out of balance, or at least I thought it was.

There shouldn't be any long streaks of unwinnable scenarios on a balanced map with the HARD setting -- if you have the mad skills you should be able to hit Restart after losing the HARD setting and be able to win the next one if you lost the previous one -- that's the way it is with 1.60 and 1.50 and B12. If you have the skills and a balanced map, you should be able to win 9 times out of 10 playing the exact same scenario over and over again using Restart -- not lose ten times in a row and never be able to win no matter what you try.

Anyway, I did have a reason for putting in my plug for not wanting the HARD setting ever to be impossible to win.

B12 is on the right track, in my opinion, and so was B11.

[ThetaOrion]

Back in Beta 2, I stumbled upon a favorite Mountain Trails Scenario, I called it my revelatory scenario, because I made all of my important discoveries regarding the AI with that scenario. It's I as Chaos 1, Ork ally 2, against IG in 3, and SM in 4.

In Beta 2, I discovered that the IG were locked into a Tier 1 rush, and never really got past Sentinels. It was a constant recurring theme, over and over again, with the TechPriests and single Guardsmen squads rushing my turrets one after another like lemmings. My revelatory scenario was really easy to win at the HARD setting and Beta 2, because the IG were broken or locked into Tier 1 rush mode. It happened every time I played the Revelatory Scenario. It was a recurring theme.

In Beta 4 with my Revelatory Scenario, I discovered that the SM were broke. They refused to recapture disputed listening posts, and they would pool or picnic in the center of the map and never move. My Ork AI ally was taking down both the SM and the IG, and the game was really easy to win, even though it was the HARD setting. It was a recurrent theme, over and over again. HARD was really easy to win with my revelatory scenario.

Then I get to Beta 7 and Beta 8, and suddenly my Revelatory Scenario was completely impossible to win at the HARD setting no matter what I tried. It was a recurrent theme, over and over again. The Orks would skip the SM and go straight for the IG. I too would be up there with a couple of armies and a tank trying to help the Orks kill the IG, and then the SM would suddenly come down fully capped and fully researched and destroy my base. One time the SM were lured into the Ork base and destroyed it, but the result was always the same. My Revelatory Scenario was completely impossible to win, ten times in a row, throughout Beta 7 and the 'Final' Beta 8. I could play other scenarios and win easily enough, but my Revelatory Scenario was completely impossible to win. So, what good were the HARDER and INSANE settings if my Revelatory Scenario was completely impossible to win at the HARD setting?

I'm persistent. Eventually with B7 and B8 and my Revelatory Scenario, I found that I as the Chaos in position 1 had to go up into IG territory and build a second base up there in the northern most part of position 3, a place where the SM AI would never go, and then I was able to achieve stalemate with my Revelatory Scenario, because I had a safe place on the Mountain Trails map where the fully capped and fully researched SM would never go. That's a massive amount of work and 'cheating' just to hit stalemate on an impossible to win scenario at the HARD setting -- finding a place to hide and building a second base and then watching the play time move toward 3 hours of time without breaking the stalemate. I never had to find a place to hide and build a second base at the HARD setting with 1.60, 1.50, B12, or B2 through B4, because I had never run into a HARD setting Revelatory Scenario that was impossible to win with 1.60 and B2 through B4. B7 and B8 were something new, and it was a recurring theme. My Revelatory Scenario was completely impossible for me to win at the HARD setting. I only won my Revelatory Scenario with B8 at the HARD setting with -dev turned on. With -dev, I finally got a different game and it played out semi normally at the HARD setting.

In B12, my Revelatory Scenario is possible to win most of the time, and it's possible to lose because I lost it one time, and it plays out differently each time. The recurrent theme with B12 is that there is no recurrent theme, it plays out differently each time. Wonderful stuff!! Thus my recurrent claim that B12 HARD setting seems to be on track and seems to be as it should be.

And, there's a summary of my history and my contribution as a beta-tester for the AI Skirmish Mod.

[thudo]

Alrighty gentleman.. lets use the OTHER thread.. B12..

and.. please keep the paragraphs to a minimum! Scripters/coders like point-form and precise discussions to root out the problems. Thanks all!