AI Notes from my first Human/Drengin war scenario
from 
GalCiv2 Forums
GalCiv2 Forums
I've been playing at Normal or below up to this point, but for this match I jumped up to Intelligent to see how I would fare.
The AI had me licked in every category but research, which turned out to be my saving grace. I took a big risk in my limited militarization. I was waiting to push out my first gen ship until I had something that was fast, armed and defended, unlike the AI which came straight at me with an unarmored, slow as molasses missile boat.
I had 2 ships in service, one of which the AI was about to wipe out, until I ran away to meet my second ship and form a fleet, which let me come away victorious. If that hadn't happened, I probably would have fallen terribly far behind militarily and the AI could have sat in orbit clipping my ships as they were produced. So I got a little lucky then I rushed to develop Planetary invasion. This cost me 13 weeks, which the AI was using to reinforce its fleets and develop better missiles, but without transports it couldn't really touch my worlds.
This was where the AI let me down a little. I had time to develop invasion tech, built transports, have 5 failed invasion attempts and built up a fleet about 30% of the size of the drengin fleet before they had even researched invasion.
Then they just started sending transports at my world, which had ships in orbit, and fleets patrolling nearby.
In think what it comes down to is the AI needs to make better use of scouts and survey ships. I learned that trick playing this game, that a small survey ship with a couple survey modules and an engine upgrade is a sure-fire way not to have your transports surprised by a patrol.
I think I probably killed 8 enemy transports that I just stumbled across in space. I couldn't even tell if they were going anywhere. The AI recaptured one of my planets solely out of my stupidity in leaving it undefended briefly, but I retook it quickly. That was good that the AI capitalized on its opportunity, but the first thing it did after it retook the planet was rebuild its starport and buy a single defender ship, the same model that I'd been crushing battle after battle, even though the AI possessed beam technology to which I had no defenses.
It wouldn't have helped it any, mind you, my primary fleet contained two fighters that had seen so much action they were up to 70HPs. (devs: is that a bug, or should a fighter really be able to gain that many hitpoints? Also, the computer never attacked them, they always went after the weaker ships in the fleet first, otherwise they might have been able to whittle those HPs away over time)
All in all this was a challenging game, and I think on a randomly generated galaxy it might have been more likely that I'd have lost, but the AI made a number of critical mistakes:
1) Poor job of adapting to my weaponry
2) Continued to develop armor technology but never armored any of its ships until very late in the game
3) Eventually developed lasers but I only saw 3 ships in combat that had them.
4) Continued producing ships with missiles after it switched to lasers even though I had pretty good ECM defenses
5) Took too long to develop planetary invasion
6) Did not protect transports
I made a few mistakes myself, but recovered from them quickly, and had a couple of gambles pay off bigtime early in the game, such as rapidly developing invasion and focusing on research at the expense of military production (not too big a risk in hindsight as the AI couldn't invade me if it wanted to)
The other thing that I did pretty much the whole game was deficit spending. My treasury was pretty much almost empty, except when I needed to get it positive quickly so that I could upgrade my two 70HP fighters. There was never a real penalty for this as long as it didn't get too low. I played the entire second half of the game with my industrial output at 60% or lower to keep spending under control, and I was still ahead of the AI.
The AI had me licked in every category but research, which turned out to be my saving grace. I took a big risk in my limited militarization. I was waiting to push out my first gen ship until I had something that was fast, armed and defended, unlike the AI which came straight at me with an unarmored, slow as molasses missile boat.
I had 2 ships in service, one of which the AI was about to wipe out, until I ran away to meet my second ship and form a fleet, which let me come away victorious. If that hadn't happened, I probably would have fallen terribly far behind militarily and the AI could have sat in orbit clipping my ships as they were produced. So I got a little lucky then I rushed to develop Planetary invasion. This cost me 13 weeks, which the AI was using to reinforce its fleets and develop better missiles, but without transports it couldn't really touch my worlds.
This was where the AI let me down a little. I had time to develop invasion tech, built transports, have 5 failed invasion attempts and built up a fleet about 30% of the size of the drengin fleet before they had even researched invasion.
Then they just started sending transports at my world, which had ships in orbit, and fleets patrolling nearby.
In think what it comes down to is the AI needs to make better use of scouts and survey ships. I learned that trick playing this game, that a small survey ship with a couple survey modules and an engine upgrade is a sure-fire way not to have your transports surprised by a patrol.
I think I probably killed 8 enemy transports that I just stumbled across in space. I couldn't even tell if they were going anywhere. The AI recaptured one of my planets solely out of my stupidity in leaving it undefended briefly, but I retook it quickly. That was good that the AI capitalized on its opportunity, but the first thing it did after it retook the planet was rebuild its starport and buy a single defender ship, the same model that I'd been crushing battle after battle, even though the AI possessed beam technology to which I had no defenses.
It wouldn't have helped it any, mind you, my primary fleet contained two fighters that had seen so much action they were up to 70HPs. (devs: is that a bug, or should a fighter really be able to gain that many hitpoints? Also, the computer never attacked them, they always went after the weaker ships in the fleet first, otherwise they might have been able to whittle those HPs away over time)
All in all this was a challenging game, and I think on a randomly generated galaxy it might have been more likely that I'd have lost, but the AI made a number of critical mistakes:
1) Poor job of adapting to my weaponry
2) Continued to develop armor technology but never armored any of its ships until very late in the game
3) Eventually developed lasers but I only saw 3 ships in combat that had them.
4) Continued producing ships with missiles after it switched to lasers even though I had pretty good ECM defenses
5) Took too long to develop planetary invasion
6) Did not protect transports
I made a few mistakes myself, but recovered from them quickly, and had a couple of gambles pay off bigtime early in the game, such as rapidly developing invasion and focusing on research at the expense of military production (not too big a risk in hindsight as the AI couldn't invade me if it wanted to)
The other thing that I did pretty much the whole game was deficit spending. My treasury was pretty much almost empty, except when I needed to get it positive quickly so that I could upgrade my two 70HP fighters. There was never a real penalty for this as long as it didn't get too low. I played the entire second half of the game with my industrial output at 60% or lower to keep spending under control, and I was still ahead of the AI.
