Hi!

Also ich hab das Spiel jetzt mal ein paar mal mit meinen Nachbarn gezockt und muss sagen Kompliment, echt cool und vom Spielfeeling alleine könnte ich nicht sagen, dass es nachgemacht ist.
So nun ist ja derzeit noch keine KI verfügbar, und wie es der Zufall so will ist das einer meiner beiden Schwerpunkte im Studium (neben Computergraphics) und mein Nachbar macht derzeit erste Schritte mit eigenen neuronalen Netzen.

Da ja Open Source erst ab der nächsten Version kommt, können wir nicht direkt starten, wollen wir aber. Wenn ihr Interesse habt, nehmt doch mal Kontakt mit mir auf: silverspike at web.de. Ggf. bitte möglichst bald wir würden am liebsten heute noch anfangen.

Das Ganze natürlich ohne Gewährleistung, kann nicht versprechen, dass da was draus wird.

Gruß

---

Ich finde es grossartig dass sich Leute wie ihr, die bestimmt mit dem Studium genug zu tun hätten auch noch diese ehrenwerte Aufgabe übernehmen wollen. Ich hoffe, dass ihr bald anfangen könnt/dürft.

Mfg

---

I think that the focus of this program should not involve AI. In my opinion the only thing that will make this game good, is people playing it, and playing against each other. The AI in every incarnation of Settlers has been terrible, because the AI has been designed to beat 10 year olds. And I bet it does a good job at it.

The focus has to be on competitive play between players. Eventually some type of league? The last time I looked in the lobby, nobody was there.

---

Zitat von brijones:

The KI is very good.
I think it should be the same as this one in the original.
Why is it terrible? Because it´s old? Thats not good argumentation (is this the right word?).

---

The AI sucks because it doesn't have a clue. I can stick 3 AI opponents teamed up against me and I still win.

---

Maybe it was built just this way that a 'normal' player can win even under these (seemingly) difficult conditions. In this case a pref to select the difficulty would be fine, if the AI of II.5 will be better (what I hope, too).

---
SPQR

---

Zitat von griflet:

Good idea (=
--
SENATUSPOPULUSQUEROMANUM

---

If players want a less steep learning curve, they can go dick around in the ADK lobby for awhile. I'll play S2.5 for one reason, and one reason alone, to be challenged by players that are of equal skill to me. It's time for Settlers / Siedlers to become more then just that tiny game you used to play around in your spare time, but moved onto other games because it just wasn't that appealing.

Let's be honest, if you are reading what I am saying, it's because you love this game and it runs through your veins like a dual stacked worker and donkey freeway between your HQ and your frontline. Are we truly wanting entry level players? If they love the other settler games, they'll come and play this when they get good.

If you ask me, this game doesn't need any players but the ones that want to duke it out multiplayer style. Every second you waste on AI that nobody here will probably even use, is a second that you take away from doing something useful, like balancing out the maps a bit more, or adding military evacuations. Or getting the damned farms to work correctly, or allowing two 2nd tier buildings to be placed within 2 squares of each other correctly. Or taking away the terrain penalties to building.

If this game ends up being good, it will because someone made some valuable contributions, not because it had AI that could beat a 10 year old. The day I played a real person, was the day, I learnt that I could play much better. So the AI is a side attraction. You'll never get it perfect, because not even the game studios have ever been able to come up with a solution that didn't involve using inside information. Remember Star Craft?

If you want this game to be good, it will do so, but I think that it's about time to chuck the source code up. There's a few things I reckon will work wonders.

---

Zitat:

That's right, I think so too but, I think an other problem is, that if you ever play against the same opponents, you will now their different kinds of strategie, that sayed the good old Napoleon too.So, if the AI is very good, it would maybe be a (little)challenge, to play on a map not just 1 on 1, if you are 2 human players.The AI should be so dangerous, that you have to take it serious.

---

---

Zitat von brijones:

Not quite right: One goal of SII.5 is to develop a playable version of The Settlers. Playable on modern systems like Linux or (what I can't wait for) Mac OS (no, Windows is not modern.  ;) ). If you want a playable Settler game, you need an AI, too: You don't have internet everywhere! I, for example, want to play The Settlers when I'm traveling by train, too.   It's not only an online version of The Settlers where no beginners were accepted.

---
SPQR

---

AI is very hard to design, especially getting it good. But, I think you really need some good players to assist you in helping to design the AI's tactics. Maybe get a string of players who play differently to add their own input into it.

Some of the worst things I see the AI doing:
* Roads longer then 2, sometimes over 5 squares distance.
* Stupid road designs that do not maximise the building efficiency.
* Random and crap placement of buildings.
* Terrible growth tactics, building unlimited amounts of mines, yet failing to secure enough food to make them at all effective.
* Placing buildings on the front line that do not belong there, such as Storehouse, positioned next to the border.
* Terrible placement of catapults.
* Terrible use of coins, sending them to small buildings to be used inefficently.
* Not refreshing troops in a building, meaning they stop upgrading troops.
* Ending up stuck behind a wall of tree's unable to expand the border because they used way too many foresters.
* Building military buildings larger then barracks in area's which will never ever see any type of battle.

The AI shows absolutely no sign of anything close to intelligence. You may as well sit a 10 year old down at the computer and it would probably play better. My 3 year old could probably beat it. It's terrible.

---

I don't think they really stop upgrading troops. If you focus on what AI does, then you will see, that he will in certain time differences withdraw all troops and send new ones. One reason is to occupy unocuppied buildings (if he has too few soldats) or to upgrade some weak units. But as he doesn't change options in the military menu such as weak defence to have weak soldiers in the building again to upgrade them. At a certain point he has that many generals, that this won't have effect anymore.

I agree with all the other points you mentioned. You could say that building some bigger military buildings at unreachable places is not too bad if there is an enemy who will probably overrun you and then AI can start over from those "backyard" position. You can expect that AI will lose anyway but it's not that senseless as you call it.

To the miserable use of catapults I can add that when I'm under heavy attack I begin to build up a catapult defence line in a certain distance to the border that they can be finished and filled with stones before the enemy arrives... This would be at least intelligent for an AI who should represent Intelligence.

---

To be honest, I cannot stand playing maps where there is way too much stone. Green Mile is a perfect map for that above theme. If the map is fairly unrestrictive, it's boring and less competition. The thing about the game is it's about becoming more efficient and finding the fastest way to do what you want. If the map isn't restrictive, then what happens is you become less efficient, and especially if you only ever play the AI, I mean, you may as well be playing a 10 year old. At least a 10 year old is smart enough to make roads that aren't a mile long between flags.

I found that S2:TNG has barely any good maps. Sure the big ones are there for everyone to see, but who the heck wants to spend 5 hours playing the game? My idea on a good map is about taking the focus away from the map, and pointing the focus at a players skill. When I play, I'm not sitting around waiting for stuff to happen, when I play, I am constantly doing something at normal speed, and there's always something to do. Making slight changes, adjustments to the economy.

Changing tactics on the battlefront. I like maps where the gold is right in the middle and all players have to tussle it out for the gold. It lends to the tactics of rushing to the gold, basically first player to get to the gold and setup stronger military buildings wins. Then you have the standoff mode where you watch your coins and military strength bypass your enemies. And do you attack? God no. When I play a game, we like to play the game like backgammon, you get to a set point and you offer to double the stakes, and they have the choice to accept the double, or resign.

My games are won without anyone even sending a single soldier. Sometimes it's not about the big battles, it's about using your strength to leverage your opponents out of the game.

If I was going to create a league or ladder, it would work like this:
* First player to a set level of military strength and coins wins.
* Game stakes can be doubled, or resigned if not accepted.
* A match requires the map played multiple times, with both players switching starting positions.
* No speed increase allowed.
* A time limit in which case set statistics are added up and the winner is handed the points.
* Maps start at 1 point, to win the match the player needs 3 points.
* If one player is 1 point away from winning, the other player is not allowed to double.

---