Privateers

[bob_x]

A good strategy for the human player is to build lots of privateers. This strategy relies on the computer player stupidity. The AI responds to the privateers by building massive caravels, which just get killed by the privateers, giving the human player Great General points. The AI doesn't even use Galleons which it views as a transport ship and not the most powerful offensive ship in the game before Chemistry is researched.

When harassed by large numbers of privateers, the AI should do what i would do, which is ignore them while rushing to get Chemistry so I can fight them off with frigates.

 

[notque]

First, that needs to be generalized for all mods. So ignore them while rushing to get a tech that will increase your boat strength.

That could be bad idea depending on how far away you are. Should you always rush to match boats, or are there other issues that may be of more importance? Maybe you need more money or research to make it to the tech for boats.

 

[Roland Johansen]

The AI should also know how to defeat stronger units with weaker units.

A privateer will on average (with equal promotions) kill 2,2 caravels. 2 caravels attacking a privateer have a chance (27%), 3 caravels will likely win against a privateer (66%) while 4 will almost certainly win (89%). This makes it inefficient to kill privateers with caravels, but maybe better than the alternative: losing all seafood and losing lots of trade.

A privateer will kill on average 1,2 galleons. 2 galleons have a good chance to win (73%), 3 galleons will almost certainly win (96%). This means that grouping galleons in groups of 2 or 3 is a decent way to counter single privateers. The first galleon will mostly lose, but the second one will likely kill the privateer meaning that such an encounter will often result in one privateer lost versus one galleon lost. The investment in galleon is of course high.

 

[jdog5000]

There's a partial fix to address this in there already ... the AI has a built in ability to consider sending a group of weak units to kill a stronger one, but the problem is that it doesn't group Caravels into stacks of more than 2 (and often just 1). I tried having these smaller groups recognize that they have friends who will join in the battle, but should probably revisit this part of the code.

 

[Roland Johansen]

There's a partial fix to address this in there already ... the AI has a built in ability to consider sending a group of weak units to kill a stronger one, but the problem is that it doesn't group Caravels into stacks of more than 2 (and often just 1). I tried having these smaller groups recognize that they have friends who will join in the battle, but should probably revisit this part of the code.

Sounds good.

I noticed that you were only talking about caravels and not about galleons. Does this also work for galleons?

 

[bob_x]

Even when the computer player doesn't attack the privateers with the caravels, they leave them floating on the water where the privateers can kill them, and then the computer players build MORE caravels. The computer players should be building research, not caravels, at this point.

Yeah, it's possible to take out a single privateer with two galleons, but human players who use a massive privateer strategy will stack them together for safety, and as I wrote above, the computer players aren't even smart enough to build galleons instead of caravels.

 

[Roland Johansen]

Yeah, it's possible to take out a single privateer with two galleons, but human players who use a massive privateer strategy will stack them together for safety, and as I wrote above, the computer players aren't even smart enough to build galleons instead of caravels.

So, shouldn't the AI then learn to use galleons? Depending on their discovered technologies, it can be the better option. Teching towards chemistry (+ astronomy) might take too long. It depends a bit on the situation. But when the AI is losing large numbers of sea food resources and trade while it has a number of galleons (maybe built for other purposes), then it could instead group those galleons together to try and stop the privateers.

For the human player, it is also not attractive to pillage and pirate an AI player when he/she loses a number of those privateers to stacks of galleons. Only in an all-out war situation would this be beneficial to the human. In a peaceful situation, the pirated and pillaged gold doesn't compensate the losses in hammers invested in privateers.

Note that if a human player is in a position to build a massive stack of privateers, move them towards the AI and the AI still does not have a counter against those units, then the AI is generally in a bad position. The human player must have a significant technology lead. The human player can use this lead in various ways and chose the privateer route. It's pretty hard for the AI to deal with a human with a significant technology lead.

 

[bob_x]

I guess there are several issues here:

(1) I get the impression that the AI thinks that the Galleon is a transport ship and doesn't realize that it's the most powerful offensive ship before Frigates. I think this should be fixed independent of whether or not the Privateer issue is addressed.

(2) Apparently, when the AI realizes that it's being attacked with ships, it responds by building ship. Now normally, this is pretty good thinking. If I see Privateers mucking around my waters, I immediately build Frigates in response.

However, I am also aware that Privateers are powerful ships compared to Galleons and especially to Caravels. And the Privateers can only harrass naval squares and not take over the cities. Thus, the best response to massive Privateers is to do nothing until you can research Astronomy and Chemistry.

(3) I think that on Pangea maps, the AI doesn't place emphasis on getting Astronomy, which leaves it at the mercy of the human player who can tech a big lead in naval ability, taking advantage of the AI's boneheaded response.