List of AI Stupidities
- Thread starter Thunderfall
- Start date
TimTheEnchanter
I...am...an Enchanter!
I can understand you taking an interest in defending the honor of a fellow Elephant 
But if I'm not mistaken, the elephant would still have use a full movement turn to attack the Tank, and another full movement to attack the rifle, whether they were on rails or not, meaning there would be no way it could attack yet another rifle in the same turn.
But if I'm not mistaken, the elephant would still have use a full movement turn to attack the Tank, and another full movement to attack the rifle, whether they were on rails or not, meaning there would be no way it could attack yet another rifle in the same turn.
Dark_Wolf
Chieftain
Now that I got all your attention with that and some are questioning and then others are trying to debunk the truth... Yes I admit I stated it as if the elephant was Stampedeing and was completely out of the computer's control...LOL sounds kinda funny though...Let me tell you the surroundings and the way the events unraveled just before the computer decided to make a mockery of me. I had what ever it was that makes all units just created into vets and the elephant I doubt was a vet. Now we'll start with the fact that I had just left a city and was going down a road I built to a fortress I think either I built or the computer did... The computer had plans of attack and I had the idea to fortify the fortress... as I ran out of moves and it became the computer's turn I was met some what half way between my city and the fortress by one archer who attacked and died then a second attacked and died. now remember the road breaks one move down to thirds, the elephant came from either inside the fortress or just beyoundand was attacking me on it's fifth step up the roadupon killing me hewould take my spot on the road due to his last 1/3 step... now as the round passed back over to me I attemped to stop this beast by grabbing the stongest ground units I had built which was rifleman. They go and cut off his path following a road the second riflemen end his move right behind the first, so no not stacked. because he was only one step away he could take both in his one move but the collection of damage was utterly wrong, he attacks the first and survives with a scratch taken off and he take the spot and continues ontothe next riflemenand wins that battle this time collecting enough damage to put him in the yellow. The Round circles back to me and because I didn't have any more troops I can't do anything but watch as my destroyer finishes and pulls out into the bay, again the elephant makes it's six moves planting it right on the shore within the destroyer's reach of four moves left and the elephant is dead... the point being that these really obsolete units are totally destroying advanced units either two or three times their strength and defence stat which should be impossible.
TimTheEnchanter
I...am...an Enchanter!
OK, in the first post you made it sound like it was all one turn. If it was multiple there's nothing fishy. And yes, the units were obsolete, but it's all pseudo-random and while highly unlikely, it's not "impossible" for him to win if he's doing the attacking.Dark_Wolf said:
Edit to add: (And yes, the AI cheats too!)
Humans expect average results. Nature delivers all over the spectrum. A 1 in a million chance is still a chance and sometimes happens.
(Even more interesting is the mathematical chance of zero which is not impossible.)
(Even more interesting is the mathematical chance of zero which is not impossible.)
Any time you do something random number, a damage roll of 0 (zero) is possible. Combat in Civ II is turn-based, yet you can deliver 4 hits in a row without taking damage. How is this possible? The AI rolls a zero for damage. This is far more likely when obsolete, regular units attack a fortified modern unit in a fortress on a mountain. The AI will roll zeros, guaranteed.
Then again, this thread proves that even under those circumstances you can roll zeros enough times for some "Phalanx vs. Armor" to happen when you attack or defend.
Then again, this thread proves that even under those circumstances you can roll zeros enough times for some "Phalanx vs. Armor" to happen when you attack or defend.
Junuxx said:(What do you mean about chances of zero not being impossible)
It is mathematically possible for the chance of an event to be zero (not infinitesimally small, but zero) yet the event is not impossible.scloopy said:
This happens only when you are dealing with infinities so in practical matters such as a game it is irrelevant.
For example, let say you were to pick a real number between 0 and 10. The chance of picking any particular number (1.7, 2, 2.6, pi, ...) is zero.
I think we all wonder whether the game processes combat results to give some slightly higher likelyhood of success when older units controlled by AI are battling more modern human-controlled units. Everyone who has played for a while has had some experiences with unusual combat results (mine include a vet sub destroyed by a damaged AI trireme); what we all need to remember is that unlikely things still have a chance of happening. Good generals/admirals always keep a reserve force, even if they are confident that the main force should win the battle.
Hello, I'm visiting from the GameFAQs web-boards (http://boards.gamefaqs.com/gfaqs9/gentopic.php?board=196926) and wondered if there was any mention of the "Surprise Nuclear" the AI can use.
Unfortunately GameFAQs deletes old topics and at the behest of new moderators, so I can't well link to the discussion that was had over there, but the premise is similar to an already listed "cheat" the AI can use, and that is to send a nuclear warhead into the heart of your capital city. (I do not know if it was SDI protected)
In one of my World-map games, my Japanese army conquered most of the known world except for the eastern United States. The time was mid-20th century and both nations were quite developed in the tech race. The US was resisting bravely, slowing down my advancement past the Mississippi River and protecting it's coastline with several fleets of warships in the Atlantic. One tactic the AI used I found disheartening was the use of sending a Nuke straight into the heart of Kyoto. I tried several ways of trying to stop this, including surrounding the city of origin. But that's where the story gets strange.
Japan was hit by a missile that came from the eastern coastline of the United States. That is about half-way around the world. Am I to believe that a nuclear missile, fresh of the assembly in Chicago, seemingly vanishes out of thin air and reappears right next to Kyoto?
Abe Lincoln had no Carriers, and the closest air field is smack-dab in the middle of China. How did a nuclear fly that far in only ONE turn? I turned cheats on to see if the missle "moves" over any spaces, and it doesn't. The turn any US city finishes building it (even if I change their production, at least one city has enough shields), it magically teleports to Kyoto. The only tactic I found I could use to protect myself was to surround my city with 8 spies (a unit each city had an abundance of). Only until I surrounded the city, was it safe from nukes. To make sure of this, I reloaded and only protected 7 sides. Boom. Reload. However, the AI was wise (relatively), seeing as it could not hit my oldest city, it would send the nuclear went to hit another city. Reload.
Only once I had surrounded all my Asian cities (my oldest and most profitable) with units did the radioactive attacks stop. Of course, the US still had them sitting in their cities, but they couldn't let them loose unless they wanted to go through their Atlantic fleet to reach Europe/Africa, or they wanted to bomb their own continent. (Australia was untouched and South America had just recently been discovered by Abe Lincoln)
So, to make a long story short, SDI everything and surround all your cities with units. It may be the only chance (at least reducing the possibility of) to protect yourself against a random nuclear strike.
-=PakoPako=-
---
"This was brought to you by FRUNGY, the Sport of Kings!"
Unfortunately GameFAQs deletes old topics and at the behest of new moderators, so I can't well link to the discussion that was had over there, but the premise is similar to an already listed "cheat" the AI can use, and that is to send a nuclear warhead into the heart of your capital city. (I do not know if it was SDI protected)
In one of my World-map games, my Japanese army conquered most of the known world except for the eastern United States. The time was mid-20th century and both nations were quite developed in the tech race. The US was resisting bravely, slowing down my advancement past the Mississippi River and protecting it's coastline with several fleets of warships in the Atlantic. One tactic the AI used I found disheartening was the use of sending a Nuke straight into the heart of Kyoto. I tried several ways of trying to stop this, including surrounding the city of origin. But that's where the story gets strange.
Japan was hit by a missile that came from the eastern coastline of the United States. That is about half-way around the world. Am I to believe that a nuclear missile, fresh of the assembly in Chicago, seemingly vanishes out of thin air and reappears right next to Kyoto?
Abe Lincoln had no Carriers, and the closest air field is smack-dab in the middle of China. How did a nuclear fly that far in only ONE turn? I turned cheats on to see if the missle "moves" over any spaces, and it doesn't. The turn any US city finishes building it (even if I change their production, at least one city has enough shields), it magically teleports to Kyoto. The only tactic I found I could use to protect myself was to surround my city with 8 spies (a unit each city had an abundance of). Only until I surrounded the city, was it safe from nukes. To make sure of this, I reloaded and only protected 7 sides. Boom. Reload. However, the AI was wise (relatively), seeing as it could not hit my oldest city, it would send the nuclear went to hit another city. Reload.
Only once I had surrounded all my Asian cities (my oldest and most profitable) with units did the radioactive attacks stop. Of course, the US still had them sitting in their cities, but they couldn't let them loose unless they wanted to go through their Atlantic fleet to reach Europe/Africa, or they wanted to bomb their own continent. (Australia was untouched and South America had just recently been discovered by Abe Lincoln)
So, to make a long story short, SDI everything and surround all your cities with units. It may be the only chance (at least reducing the possibility of) to protect yourself against a random nuclear strike.
-=PakoPako=-
---
"This was brought to you by FRUNGY, the Sport of Kings!"
Ace
Emperor
Just build the SDIs, surrounding your cities with units doesn't really add to the SDI defense.
TimTheEnchanter
I...am...an Enchanter!
I don't know how much of this has been definitively proven, but here's my understanding on some of those issues...
1) Yes, the AI's nukes (and cruise missiles too, I think) to not truly travel to their destination. They just appear next to it. Putting a barricade of Units or Airfields around the AI cities to try to trap an AI missile will fail. It sure seems like cheating, but that's how the game works. The programmers probably had too many missles fly off into nowhereland whenever one encountered a unit blocking its path (judging by how ships handle adversity in their path) so they just took the easy way out.
2) The AI could have theoretically had a sub with missles on it bomb you.
3) I recall reading somewhere that once the AI has a nuke, it will only attack within 16 squares of that AI's cities, but that they can essentially launch it from ANY city they own - it doesn't have to be the one with the missle in it. So if the AI builds a nuke in New york and also has Seattle, or Alaska, they could launch the nuke from the west coast and travel 16 squares from there without having to transport the nuke to the launch city first. Again, I don't know how well verified this is, but it's what I read. I definitely don't know if the same applies for automatically transporting and launching from a sub.
4) Learn to create Nuke traps. The AI will not nuke a city with SDI, and they tend to target large cities. SDI covers all areas within 3 squares of its location - so place a small city near your big cities and put an SDI in the small city, but leave the big cities without. They will be protected by the smaller cities SDI umbrella, but the AI can't figure this out so they will send missle after missile at the big city and fail every time because of the nearby SDI.
1) Yes, the AI's nukes (and cruise missiles too, I think) to not truly travel to their destination. They just appear next to it. Putting a barricade of Units or Airfields around the AI cities to try to trap an AI missile will fail. It sure seems like cheating, but that's how the game works. The programmers probably had too many missles fly off into nowhereland whenever one encountered a unit blocking its path (judging by how ships handle adversity in their path) so they just took the easy way out.
2) The AI could have theoretically had a sub with missles on it bomb you.
3) I recall reading somewhere that once the AI has a nuke, it will only attack within 16 squares of that AI's cities, but that they can essentially launch it from ANY city they own - it doesn't have to be the one with the missle in it. So if the AI builds a nuke in New york and also has Seattle, or Alaska, they could launch the nuke from the west coast and travel 16 squares from there without having to transport the nuke to the launch city first. Again, I don't know how well verified this is, but it's what I read. I definitely don't know if the same applies for automatically transporting and launching from a sub.
4) Learn to create Nuke traps. The AI will not nuke a city with SDI, and they tend to target large cities. SDI covers all areas within 3 squares of its location - so place a small city near your big cities and put an SDI in the small city, but leave the big cities without. They will be protected by the smaller cities SDI umbrella, but the AI can't figure this out so they will send missle after missile at the big city and fail every time because of the nearby SDI.
Personally, I say that if you need to cheat half as much as the AI to win Civ II, you may want to re-evaluate your strategy. I solve this problem efficiently by ensuring that nobody is alive but me by the Modern Age. (Then again, I never did learn to play II above Prince level, so I may just sound like an idiot in this regard. )
)It worked for me. By "impeding" the path, so the rocket can't reach its destination (a city), the AI won't fire the rocket off.TimTheEnchanter said:
I turned on cheats to view everything. I continued the turn with and without units surrounding my city. The nuke would just "appear". The AI did not have subs, and its carrier (singular) was half-way across the world.
I actually thought subs couldn't let anything "board" them (missiles, nukes, etc.).
Yes, is true and frightening.
If you're going to send a settler/engineer to do that, why not just build an SDI in your big city? If SDIs really do work as they chaim...
-=PakoPako=-
TimTheEnchanter
I...am...an Enchanter!
You also can do this with existing cities if they're close enough, and the one with the SDI doesn't have to be small. It's just that to get maximum coverage, a well placed new town often does better than your existing cities due to spacingpakopako said:
The value of the nuke trap (over SDIs in every town) is twofold - First with a well placed little town, you can often cover multiple big cities with one nuke trap, saving the cost of multiple 200 sheild SDIs with the cost of a 40s engineer. Second, because the big fat cities do not have SDIs in them the AI will send nukes at them, wasting their sheilds (160s at a time) on a doomed attack. If you put the SDIs in all those big cities, the AI will either direct its nuke attack on some other city you haven't protected yet, or will devote those shields to some effort that has a greater-than-zero chance of hurting you. Big cities attract nukes like flies, and your little Nuke trap city just swats them away every time.
Junuxx
Emperor
- Joined
- Sep 6, 2005
- Messages
- 1,154
Yeah, nuke traps work like a charm. And like Tim says, the point is that the AI will shoot all its missile and hit nothing. With the SDI in the big city,it will either shoot at something else or keep the missiles.
Anyway, another AI stupidity I remember having seen a couple of times is when a computer player starts on a pretty small island (say, 50-100 tiles) and it builds 3 cities, none of which is coastal
In one game I just couldn't find where the Germans were hiding. I had contact with them because of Marco Polo, but I didn't bother exploring all the minor islands in that game. I finally found them when I completed the Apollo Program, still with 3 cities and loads of obsolete units
Anyway, another AI stupidity I remember having seen a couple of times is when a computer player starts on a pretty small island (say, 50-100 tiles) and it builds 3 cities, none of which is coastal
In one game I just couldn't find where the Germans were hiding. I had contact with them because of Marco Polo, but I didn't bother exploring all the minor islands in that game. I finally found them when I completed the Apollo Program, still with 3 cities and loads of obsolete units
