Warlords TBS Series
Spin-off Projects
Home Forum
Welcome,
Guest
|
|
TOPIC: Is the AI cheating in combat?
Is the AI cheating in combat? 13 years, 1 month ago #144
I am one of those players who finds it puzzling how hard the designers made this game. I find it very hard to even begin, since my early troops are usually cut to ribbons at the first castle or ruins I try to take, and the AI replenishes troops much faster than I do (and they are higher level than mine...somehow).
It has been great to see the game running again, but it still seems to me that the AI hits/critical hits at a much higher rate than I do. Just today, I watched a few of my spearmen take on some goblins. Their combat strength was even, and no spells were active. I counted as the goblins hit me seven times in a row, while I missed them seven times in a row. The result was that just one of their units destroyed four of mine. I was playing on Prince difficulty. It's an ongoing grudge of mine: All of my units inflict 1 point of damage all the time, while the AI seems to get big, juicy hits all the time. Are we (you, them, the powers that be, whoever) sure that the AI is not cheating in the background? Is it assumed that the player will lose a few times before building up a retinue and warlord skills? Any solid players out there want to describe their strategies or opening moves? |
|
|
Re: Is the AI cheating in combat? 13 years, 1 month ago #149
Jephrey,
The answer is NO (I know this for sure since I have the code). The AI is not cheating in combat. However it does have some other cheats built in (extra gold, extra hero offers, extra mercenary offers, high levels on it's extra heroes/mercenaries, reduced research times on spells). How much depends on the AI level. At Prince level the overall cheats the AI gets actually slightly favor the player (AI gets less gold, pays more for heroes/mercs) while at Emperor the cheats favor the AI. And I too have seen my armies miss multiple times in a row while my opponent (human, AI, neutral city) have hit me multiple times in a row. The worst is when this happens when you are facing a Troll and he keeps regenerating the lost single points of damage so that you lose an entire group of 4-5 units without killing a single Troll. These are simply cases of bad luck (missing 7 times in a row with equal odds happens 1-128 battles or roughly a few times a game if you have a lot of battles). Early game (turns 1-10) you should focus exclusively on making the 1/2 turn units (the ones you get 2 of a turn) for expansion. Gradually you'll want a couple of cities making your best unit (Pegasi if you are Empire, Archon if you are Knight etc). Usually your Capitol is where you make your first strong unit since you should be getting the building upgrades to allow extra combat/life/faster build time on the Capitol itself. Don't attack any ruin that has more than 2 defenders in turns 1-10. Get them later. Same with neutral cities. Bypass all the ones with more than 4 units in them. The ones with L3 or L4 walls should be avoided as well if they have 4 men as those take a toll on you from the city walls. There are easy cities out there, but you have to search for them so send men out looking following roads obviously where possible. Stronger neutrals (and other players) tend to be toward the middle of the map so when possible look towards the edges for easier cities. Once your retinue arrives (around turn 10) those units can be used to take the tougher neutrals/ruins. KGB |
|
Last Edit: 13 years, 1 month ago by KGB.
|
Re: Is the AI cheating in combat? 13 years, 1 month ago #150
Don't know if I could really be counted as a solid player but...
It isn't supposed to play out like you described it, there are no such AI cheats in battles (as far as I know). Are you sure it's, um, for real? I have a tendency to get upset and angry when it goes bad and not react as much when it goes well, thereby remembering when it goes bad to a greater extent than the opposite. If you want to be absolutely sure you should take statistics on many battles. To really check it out statistically you should save before battle, then play out the same battle over and over again (reload it at least 20 times) and see what happens. Remember there are also other effects like Morale and warlord skills like Crusader and Defender that can affect the battle. But if the in-battle screen says their combat strength is even then they should (statistically) hit and miss equally. Some tactics, then... Taking cities and ruins: I think one of the very fundamentals of W4 is to not attack neutral cities of level 3 or 4 in the initial stages of the game (the first 8-10 rounds or so) if you can avoid it, because precious troops will be lost in the initial stages. Level 3 cities can, depending on their race and the number of troops in them, be taken on with the initial hero stack. But what you focus on is attacking level 1 and 2 cities, produce many weak units (also known as "fodder") that can take on other weak cities, and simply swarm the map. You should also perhaps avoid searching ruins initially unless they have just one unit in them (although even a single Fire Dragon can effectively cripple your army). Then, as you've established yourself a decent territory and start getting into territorial disputes with your opponents, you can slowly start shifting toward more powerful units that take longer to build - but only gradually so. Then you can also use single fodder units to search ruins so you can decide which ones can be worth attacking. Know your enemy: I always start with checking in the "View opponents" menu to see what my opponents are (race, skills, retinues) to know what to expect and to be able to prepare accordingly (research counter spells, build units with slayer abilities). I also keep track through the "Reports" meny. First thing is to check out Morale just once to see whose troops will be tougher than others. During the first 10 round or so the "Cities" screen is most important, because it'll tell you if you're keeping up with the expansion. After that, it's "Upkeep" that becomes the most important because it'll tell you about the relative fire power, especially when compared to the "Units" screen. The number of units in itself is not as interesting. Two examples: If an opponent has many units but not so high upkeep it means his army consists mostly of fodder (and heroes). If an opponent has about the same number of units but much higher Upkeep then you're probably in deep trouble because it means he has a larger portion of power units that will eat your fodder troops for breakfast. Cities: Don't ever keep cities from races that are more than 2 steps away from your own on the race wheel, because they will suffer from slower production times. Raze and rebuild. If you don't have enough money to rebuild, then just raze and rebuild later, as soon as you have enough gold (>400). Some of this also depends on the race and class of your warlord to a certain extent, as there are some races that should be played with differently. What kind of warlord do you have? |
|
"Negate does not negate Negate."
--- KGB "Moreover, I advise that Daemons and Dark Elves must switch places on the Race Wheel." --- Marcus Porcius Cato
Last Edit: 13 years, 1 month ago by Seppuccu.
|
Re: Is the AI cheating in combat? 13 years, 1 month ago #152
Sounds like great advice, especially since the two of you identify some of the same problems. If KGB says there's no cheating, that's good enough for me. He is, after all, the Grand Keeper of the Game . Reloading the same battle over and over might be educational anyway, just to see the distribution of results. Good idea. Since I've never tried swarming the maps with fodder, I will definitely give that a whirl.
Yeah, it really was seven in a row. By that time, i was frustrated and counting on my fingers, just to confirm my frustration. What I really hate is those dwarven crossbows, who "poke" you gently (that's what it looks like, anyway) and casually blow up your heroes like they were balloons. KGB makes a good statistical point, though, that a disaster like that is likely to happen in each scenario, given the large number of battles. I have this trait of loving games but not necessarily being good at them. I get distracted and wind up watching the pretty dragon animation, instead of watching how it eats my armies. I don't think I ever got past the second scenario in the W4 campaign and have been challenging myself not to resign ever (at least not until later in the campaign). I play mostly Knights, which I want to understand before experimenting with the other races. Currently I'm focusing on a Divine/Combat warlord. When I resign enough to gain a level (haha!) I usually take production acceleration or +2 cbt or maybe +1 mana regen as my first skill. I will send peons to and fro' about the countryside in large groups, terrorizing enemy sheep and chickens, and let you know how it goes! |
|
|
Re: Is the AI cheating in combat? 13 years, 1 month ago #153
Good good, give those chicken what they deserve!
Knights are definitely a race you can play according to the standard formula. It's also (one of) the best race(s) for the campaign, as the campaign clearly was designed to be played with a good (or neutral) race, and because most of the cities close to your capital will tend to be either Knight, Empire or Dwarven cities. I would perhaps have gone for a Priest or a Paladin instead of your Templar, but in any case Divine is good to use with Knights and I suppose the Altar is what you're after. You don't need to let fodder move in large groups; let them spread alone or in pairs and then group them together when it's time to attack. Four Swordmen should be enough to take out any L1 neutral city with only two defenders. Being four will also give them +1 Morale. You should always take the -1 Production Time building as your first upgrade, and then go for at least one (preferably two) +2 Combat before boosting anything else. Oh, and one more important thing: Use vectoring. Your newly produced units will move much quicker to a place where they are needed. |
|
"Negate does not negate Negate."
--- KGB "Moreover, I advise that Daemons and Dark Elves must switch places on the Race Wheel." --- Marcus Porcius Cato
Last Edit: 13 years, 1 month ago by Seppuccu.
|
Re: Is the AI cheating in combat? 13 years, 1 month ago #155
Okay, now I'm seeing many new things happening and new strategies opening up. I even bumped it up to King difficulty and still managed to win the 1st scenario. It took much longer than before - lots and lots of mayhem and death. Good stuff. I'm also seeing that my fodder occasionally kills strong enemy units, so I got some payback.
Templar apparently doesn't have the -1 production turn improvement. I didn't see it on the list anywhere, so I took +2 combat. I had a total of +4 cbt as I moved to the 2nd scenario. Everything played much differently and better. It isn't easy, but it makes more sense now. At King level, they're throwing absurdly strongly units at me, and I'm doing okay against them. I resigned once but didn't mind too much. My first targets were the two cities north of the starting point that both have windmills attached. Costly to take but it allowed me to build many knights. I kept my capital building +4 cbt swordsmen and sent them wherever needed. A merchant offered me 3 units of hydra (hydrae?), which I bought and used to take a quarter of the map. Very satisfying to attack a castle with 16 units inside and watch them all get chewed up by multiattack. With a pair of heroes in one army (one with +5 heal & leadership, the other with +5 armor and uber combat strength) I eventually defeated Yellow, though I made some mistakes that cost me all three hydra. We buried each head separately, so the funerals took a very long time. I also burned down some cities just for the money and spent that money building my other cities to Citadel level. I reasoned that one doesn't actually need every city on the map - the ready cash provides its own advantage? Even with +4 cbt coming out of my capital, there's far too much land to protect with the capital units. I need to find a way to keep my captured cities safe using their own wimpy troops. The AI was throwing very advanced units at me, constantly. I was too busy defending to do the quest for the sword of Etheria, which began deep in enemy territory. But thinking back, maybe having the sword would have helped more than the extra cities? Not sure. At this point, I may be too weak to win. My defenses are very thin all over my territory, due to Orange sending very powerful armies from the west and northwest. But it was a blast, for the most part. Spreading out the swordsmen makes me think you're allowing plenty of time before taking cities. I've been assuming that I want to grab at least a couple cities right away. I owned most of the southern section of the map (almost all cities on 'my side' of the rivers) before enemies starting flowing in, but maybe that was expanding too quickly? Questions: Beyond the morale bonus for 4 similar or 8 identical, does unit alignment have any impact on the stack? Shall I freely mix opposite-aligned units, if it makes for a powerful stack? |
|
|
Re: Is the AI cheating in combat? 13 years, 1 month ago #157
First of all, a lot of the Campaign maps play out differently than your average map because they in general are smaller and because there are more L3 and L4 cities. This means you get into serious confrontation earlier in the game. Personally I think the first campaign map is one of the tougher ones I've played.
I don't allow time to pass before taking cities, you should take as many cities as possible as fast as possible. It's just that when you play with hidden map and fog of war you'lll uncover the map much faster if you spread out your units, thereby finding out where the points of interest (weak cities) are so you can go directly for them with the rest of your troops. You should definitely have gone for the Sword of Etheria, it's one of the best items in the game if you play with a good or neutral race. You get -2 Morale for oppositely aligned units in the same stack, so in general it should be avoided. |
|
"Negate does not negate Negate."
--- KGB "Moreover, I advise that Daemons and Dark Elves must switch places on the Race Wheel." --- Marcus Porcius Cato |
Re: Is the AI cheating in combat? 13 years, 1 month ago #158
Jephrey wrote:
Templar apparently doesn't have the -1 production turn improvement. Boo for Templar then. Jephrey wrote: With a pair of heroes in one army (one with +5 heal & leadership Oh yes, a Hero with Heal is a key unit. Make sure you don't lose it. Jephrey wrote: I also burned down some cities just for the money and spent that money building my other cities to Citadel level. I reasoned that one doesn't actually need every city on the map - the ready cash provides its own advantage? I rarely build cities all the way to L4. If it's a key city that's going to see a lot of battle then yes (but more often L3), otherwise not simply because it costs too much. Plus, remember that two L4 cities can still just produce 4 Swordsmen per turn, while four L1 cities can produce 8. That's a relevant difference. You might not need every city but you'll need most of them, especially on small maps. Jephrey wrote: At this point, I may be too weak to win. My defenses are very thin all over my territory, due to Orange sending very powerful armies from the west and northwest. But it was a blast, for the most part. What map would that be? Jephrey wrote: I owned most of the southern section of the map (almost all cities on 'my side' of the rivers) before enemies starting flowing in, but maybe that was expanding too quickly? No, it's perfect. Maybe the one staight north from your capital can be left out from the start though. I think it's best to first take the two cities in the south - first the city straight to the south, then you head northwest, and last to the north - and then you vector the produced units from the cities in the south toward the north so you can make your capital and the northwestern city strongholds. Leave the cities in the south completely empty, instead you use the troops in the north to wipe out anything that tries to get past them. Then I'd make two seige units and attack the Dwarven city to the east. And since you're playing with Knights you can initially just raze the two Orcish cities in the northwest until you've built up a stronger army. |
|
"Negate does not negate Negate."
--- KGB "Moreover, I advise that Daemons and Dark Elves must switch places on the Race Wheel." --- Marcus Porcius Cato |
Re: Is the AI cheating in combat? 13 years, 1 month ago #161
This was actually the 2nd campaign map, with two cities north of the starting city that have -1 production time (each).
Holding cities is a real problem. It seems like (at King level) no army can defend against the AI successfully, unless it has some bonuses from spells or expensive units. It frequently wipes out 8 of my swordsmen, perhaps weakening it enough so my hero can take the city back, when he arrives. Which means he can't go looking for the sword of Etheria. grrr. I agree about the healing hero. Even with high-hit-point units (golems and elementals) I don't see how playing the dwarves can work without a healing unit thrown into the stack. Fighting against an 8-unit enemy, it doesn't seem possible to survive without healing, even if the attacking unit is exceedingly strong and the enemy units are weak. Or I should say, it is possible, but the odds will catch up with you. Since a main fighting unit (like your main hero's unit) must engage in scores of battles, it seems like you either have to have the odds locked up tight (with an undefeatable unit) or constantly replace units and keep a couple of "fodder slots" in your army - units that you assume will die quickly. In the latter case, it's very hard to keep supplied with fodder, when you are going deep into enemy territory. Even if you build an army of fodder to follow along, they'll get ambushed. And if you're going several moves into enemy land to pursue a quest item, it becomes nearly impossible to send reinforcement. So that only leaves the former case, where you've got a combination of units that can win at least 99% of battles in the field. Is my logic flawed here? My hero armies frequently get stalled when replacing the fodder becomes too difficult, and then they turn into defensive armies. What is your strategy for keeping your main attack armies moving forward? |
|
|
Re: Is the AI cheating in combat? 13 years, 1 month ago #162
Jephrey wrote:
This was actually the 2nd campaign map, with two cities north of the starting city that have -1 production time (each). I'm guessing we're talking about the Dwarves map then, I thought it was the 2nd storyline map (Giantridge Mountains). I remember that map, it was the first time I played Emperor + Uber Retinue. It was painful. Jephrey wrote: Even with high-hit-point units (golems and elementals) I don't see how playing the dwarves can work without a healing unit thrown into the stack. Fighting against an 8-unit enemy, it doesn't seem possible to survive without healing, even if the attacking unit is exceedingly strong and the enemy units are weak. Or I should say, it is possible, but the odds will catch up with you. I'm not extremely fond of neither Golems nor Elementals. Golems don't really have so much combat even though they can take a lot of punishment. If I build them I usually do it with Armour and not Crushing Blow. Elementals are ok but it takes too many turns to build them. 4 turns (or even 3) for a unit with only 7 combat - that's just not worth it. I'd rather have 6-8 Axemen. I usually use Dwarves for fodder, since their Axemen are better than Swordsmen - they are 5/13 which means they on average can survive one extra hit, and thereby on average get one extra swing. It also means they won't die from a single critical hit. If there aren't Orcs around I build Axemen with Build so that I get a discount on city upgrades. Jephrey wrote: Since a main fighting unit (like your main hero's unit) must engage in scores of battles, it seems like you either have to have the odds locked up tight (with an undefeatable unit) or constantly replace units and keep a couple of "fodder slots" in your army - units that you assume will die quickly. In the latter case, it's very hard to keep supplied with fodder, when you are going deep into enemy territory. Even if you build an army of fodder to follow along, they'll get ambushed. And if you're going several moves into enemy land to pursue a quest item, it becomes nearly impossible to send reinforcement. So that only leaves the former case, where you've got a combination of units that can win at least 99% of battles in the field. There's another way also, which is the one I almost always use before I've built a proper retinue unit with nice abilities and items: Forget about the quest until you're sure you'll win, and surround the last enemy capital while you complete the quest. The AI won't try completing the quest. This means you won't have the advantage of the Sword of Etheria on this map but on the next one you will. It'll take a few more rounds to complete the map but hey, that only means your Warlord will get one or two extra XP. Another thing: You want to have one evil (or Ogre) hero in the same stack as your Leadership hero, because then your stack will enjoy the benefits of both Leadership and Fear. Leadership +5 and Fear +5 together makes for a total power difference of up to 10 combat. |
|
"Negate does not negate Negate."
--- KGB "Moreover, I advise that Daemons and Dark Elves must switch places on the Race Wheel." --- Marcus Porcius Cato
Last Edit: 13 years, 1 month ago by Seppuccu.
|
|
Time to create page: 0.74 seconds