AI's Lack of Interest in Sensors Places it at a Huge Disadvantage!

When the sensor limit of 15 was imposed Brad said it was for performance reasons.
With "Eyes of the Universe", all your ships and starbases get 15 sensor range. However, planets remain limited to a range of around 4, even those with ships in orbit! This makes no sense!

A much bigger issue is the AI's lack of interest in senors! The AI will not research sensors, not even the inexpensive first one which gives all ships a +1 boost to their sensor range! This places the AI at a large disadvantage, as a skilled human player will quickly research this branch and build "Eyes of the Universe".

In v1.0 the AI did not value speed and that was corrected. Brad has said the AI continues to ignore sensors, for performance reasons.

Paul D.
28,643 views 28 replies
Reply #1 Top
However, planets remain limited to a range of around 4, even those with ships in orbit! This makes no sense!

I agree. The "ability value" in sensors should be added to planet sight range, including any boosts occurring.

And the AI seems to be ignoring the sensor techs indeed. I wonder if this is fully intentional.
Reply #2 Top
They should remove Eye of the Universe. It's plainly unbalancing. We human player always rush for this tech, and rush to build this wonder. AI doesn't.

We players really know how to use 15-range sensor, AI don't.

Reply #4 Top
Perhaps the AI already has the ability to see the entire map, and thus doesn't need sensors.


No. The AI gets no 'cheats' at intelligent and below and (at all levels) has to scout manually. It does get a hefty sensor bonus on the top-end difficulty levels though along with all the other bonuses it gets there.
Reply #5 Top
They should remove Eye of the Universe. It's plainly unbalancing. We human player always rush for this tech, and rush to build this wonder. AI doesn't.


Yes most of us do, but is it unbalancing? If I didn't have eyes of the universe, I'd just build ships with 15 sensor range. They would travel alongside my fleets, not in them, so there's no logistics penalty, you wouldn't need all that many of them and they're not all *that* expensive to build.

I don't think it makes much difference to be honest. Apart from removing the need for some boring micromanagement (there's enough of that in the game already, thanks!)

If the AI doesn't know how to use sensors properly, and I suspect it doesn't, then it would be a waste of time having it research them just to begrudge the human player of something which is easily, albeit tediously, recreated by building ships.
Reply #6 Top
Even if the AI knows how to get sensors it still doesn't know how to translate this into superior tactics. The AI learning what to do with sensors would only help if it knew to attack transports, sneak past fleets to invade lightly defended planets, hit military starbases first and launch Blitzkreig attacks. Without this the AI simply has to smash the player with greatly overwhelming force over the course of a few dozen turns.
Reply #7 Top
Yes most of us do, but is it unbalancing?



IMO, yes. Those ships with 15 sensor range cannot pack as many weapons or armour, so they need to be defended, and there will be far fewer of them. The advantage of having every ship have a huge sensor range is tremendous, espcially early on in the game. The strategic advantage it gives is immense - being able to keep an eye on the AIs means that I can maintain my defences with a much smaller fleet.

I think a good fix for EoTU would be for it to double sensor range up to a maximum of 15, so that it only works on ships which have at least some sensors, and still makes it worthwhile putting multiple sensor modules into a design. There should also be some kind of disadvantage to using it, such as making the ships of the player that has it "visible" for +1 space further away than normal, and/or making a 15-space radius around the planet where it is built permanently visible to ALL players.
Reply #8 Top
A much bigger issue is the AI's lack of interest in senors! The AI will not research sensors, not even the inexpensive first one which gives all ships a +1 boost to their sensor range! This places the AI at a large disadvantage, as a skilled human player will quickly research this branch and build "Eyes of the Universe".


I have only built Eyes of the universe once in 30+ games, yet on suicidal lvl won every game. The Eyes of the universe is not unbalanced, does not need to be removed or altered, indeed its perfectly balanced. All it does is make the game easier to play not win. Its really just a useful playing tool for lazy players. My 2-6 early survey ships and the rest of my fleet, keep me just as well informed on whats going on in the galaxy.

I consider it a waste of resources.
Reply #9 Top
IMO, yes. Those ships with 15 sensor range cannot pack as many weapons or armour, so they need to be defended, and there will be far fewer of them. The advantage of having every ship have a huge sensor range is tremendous, espcially early on in the game. The strategic advantage it gives is immense - being able to keep an eye on the AIs means that I can maintain my defences with a much smaller fleet.


Putting weapons and armour on them is daft. You'll have to pay maintenance and they won't be able to pack enough of a punch for it to be worthwhile. To defend them, just withdraw them (no weapons or armour means plenty of space for engines to make them faster than any attackers) or send a fleet to intercept.

You can cover an entire gigantic map with roughly 60 of them, which is not much considering for the vast majority of the game you have no need to see most of the map.

The Eyes of the universe is not unbalanced, does not need to be removed or altered, indeed its perfectly balanced. All it does is make the game easier to play not win. Its really just a useful playing tool for lazy players.


Spot on (from a lazy player)



Reply #10 Top
Well megaloob I guess we have different play styles. All I can tell you for sure is that getting it early on when you're playing a gigantic map is NOT a waste of resources!

Putting weapons and armour on them is daft.


OK, random, so you have dedicated sensor ships to build. EoTU is cheap; it costs about as much to build as, say, 3 good, fast sensor ships, and it instantly "gives" you as many sensor ships as you have vessels. And you can build scouts and survey ships that are faster and longer-range because they don't need sensor modules.
Reply #11 Top
Brad has said the AI continues to ignore sensors, for performance reasons.


IIRC, the original AI ship-design timing of once per game year was also imposed due to performance reasons. Since the "AI uses CPU-intensive algorithms" to design ships more often was introduced, I have noticed improved gameplay with no performance hit whatsoever. My computer's decent, a 2.4 gb P4 with 1 gb RAM, but it's also five years old.

I haven't tried it yet, but next time I play a gigantic galaxy, I think I'll sell/trade the sensor techs to everyone as soon as I get them, then see what happens. I'll also be interested to look at the enemy ship info screens to see if they actually put sensors on their ships.

If, as I suspect, there is no performance drop, I would like to see another option added to allow the AI to research and use sensors.
Reply #12 Top
Yes, to MarshallONeil
I always trade/give away the sensor techs, they are non-military and the AI does not
use them very well (check out the minor races ships, they seem to use sensors better
than the major powers !) The AI manages the fleet intel task by using scout ships.
The AI then REMEMBERS the entire map (much better than a human player)
This balances the player advantage of active sensor nets somewhat...
Reply #13 Top
OK, random, so you have dedicated sensor ships to build. EoTU is cheap; it costs about as much to build as, say, 3 good, fast sensor ships, and it instantly "gives" you as many sensor ships as you have vessels.


Completely agreed. So what? That doesn't make it particularly unbalancing, just convenient. I said 60 of them to cover a gigantic map, but since the really useful place to have them is near your borders, all you actually *need* is about a dozen, and even that's more than enough. So it saves you maybe a couple of grand over the course of a game on a gigantic map. Big deal. It does save you loads of annoying micromanagement though, whether it's due to constantly moving ships out of danger (the sensor boat approach) or having to mess about putting a specially designed ship with less weapons and extra sensors into every fleet.

If you want to talk unbalancing, look at the broken mind control center, psionic weapons, the spin control center, even that one that makes all your diplomacy offers worth more (if you happen to use it)...they all offer far greater rewards. EotU is one of the few wonders that has it right IMO.
Reply #14 Top
think I'll sell/trade the sensor techs to everyone as soon as I get them

Don't expect much for them as the AI does not place much value on them.

Even if the AI does not place them on its ships, if it gets the first sensor tech, its ships will get +1 to their sensor range (even ships without sensors get a base sensor range that increases with the hull size).

A few fast ships with sensors help me explore more anomalies than the AI. Those $2500 Billion anomalies help keep me from going broke in the early game!

Even if the AI knows how to get sensors it still doesn't know how to translate this into superior tactics. The AI learning what to do with sensors would only help if it knew to attack transports, sneak past fleets to invade lightly defended planets, hit military starbases first and launch Blitzkreig attacks. Without this the AI simply has to smash the player with greatly overwhelming force over the course of a few dozen turns.

My first target is the AI's starbases mining resources (especially military resources). Once I have a fleet strong enough to take out the mining starbase (most are undefended) and enough constructors to build a starbase with all mining modules plus weapons and defense, next to each of the AI’s mining starbases, I start the war by taking them out and rebuilding them in a single turn. Then I take out the AI's fleets and then the ships defending its planets. Finally, I send in a wave of fast troop transports to capture his best planets. Once the remaining planets are revolting to me I make peace and wait for them to flip, unless I'm impatient in which case I keep sending troop transports until they surrender. After my first wave of troop transports captures several AI planets, the AI rush builds troop transport to take them back. Having fast attack ships (with 15 sensor range thanks to Eyes of the Universe) is useful to intercepting enemy troop transports that appear in orbit around those remaining planets.
Reply #15 Top
Well megaloob


Insults r always a sign of someone losing a arguement  
Reply #16 Top
I just wish they would increase thge sensor range for planets to 8-12 instead of 4. I don't think it should be 15 however. Its not like planets don't have satellites telescopes ect. that would at least give them a 8 sensor range. Not as much as ships since planets rotate and ships have a 3 dimensional view of the galaxy.
Reply #17 Top
Not as much as ships since planets rotate and ships have a 3 dimensional view of the galaxy.

The planet rotating would not matter if it had several observation points. In inaddition to ground based sensors, they would have space based sensors.
Reply #18 Top
The AI then REMEMBERS the entire map (much better than a human player)
This balances the player advantage of active sensor nets somewhat...



That's an interesting point that I had never considered. However, much more important that simply knowing where things were at one point in time, EotU gives you a much better chance to keep an eye on where fleets are moving. That's why I find it so worthwhile, and no, I'd rather not have to manage X number of scout ("sensor") ships. There's plenty of micromanagement opportunity, most of it much more interesting than that.
Reply #19 Top
The AI then REMEMBERS the entire map (much better than a human player)
This balances the player advantage of active sensor nets somewhat...


Still, he does not know the position of my un-escorted transports, or the positionning of my sneaky fleets.

EoTU once allowed me to locale 3 ennemy fleets that were coming from a point I did not had any intelligence. I manage to scramble a defence in time.

It's not that the AI would not use if efficiently, it's that WE use it too efficiently. We are intelligent, damnit, they are not!

Reply #20 Top
With the whole thing about AI priority with sensors and performace - it would be really great for an option to have them prioritise them more and use them. As an option, people with old people wouldn't have to worry - but for people with good rigs it could be a viable option. I mean, I'm running an overclocked Conroe E6600 with 2gigs of memory, I want all options added in please
Reply #21 Top
The other way of fixing this would be to make people play more like the AI. Make sensors be even bigger and cut the limit to something like 8.

What kills me about this game is all the mechanics that the AI is not good at. It just avoids them. EITHER FIX the AI to play the game that you designed, or change the game so that the AI isn't at a huge disadvantage.
Reply #22 Top
I read somewhere in the Developer's Journal where Brad says something to the effect of (and I paraphrase here) "It's not clear if there's any such thing as a fog of war at the higher AI levels".

Since I mostly play at the higher AI levels, I appreciate Eyes of the Universe to get me close to the functionality that the AI enjoys.
Reply #23 Top
I read somewhere in the Developer's Journal where Brad says something to the effect of (and I paraphrase here) "It's not clear if there's any such thing as a fog of war at the higher AI levels".

I have read that the AI never cheats; at higher levels it gets bonuses (to their economy, MP, SP, research, influence, etc.).
Reply #24 Top
I have read that the AI never cheats; at higher levels it gets bonuses (to their economy, MP, SP, research, influence, etc.).

Did you read anywhere that the AI doesn't get a bonus to sensor range as well?

It took awhile to find the quote but here it is:

For instance, at 200 bonus, the sensors are probably so powerful that they have no real fog of war.


WWW Link

You'll need to scroll to the top of the thread. The quote is buried in a rather long post entitled Fairness in Computer AI by Draginol in the GalCiv Dev Journals forum.
Reply #25 Top
When I played my first Metaverse game I selected Painful thinking the AI was getting a 50-75% bonus because on Crippling I thought it got a 125% bonus. It turns out on Crippling it runs at 125% (only a 25% bonus) and on Painful at 105% (only a 5% bonus, no wonder it was so easy!). So likely the AI player on Incredible (Masochistic, Obscene, and Suicidal) ships’ sensors run at 200% (100% bonus). Since the AI does not research sensors (and thus does not get the +1 sensor bonus from the first sensor tech) and thus does not place them on its ships, so depending on the hull size the AI players' ships will have sensor range 2-4 (with the bonus 4-8). So the AI players’ ships will still have shorter sensor range than my ships with "Eyes of the Universe". Planets get a sensor range of 4; I wonder if the AI planets get a sensor bonus.