Recommend
 
 Thumb up
 Hide
11 Posts

Puerto Rico» Forums » Strategy

Subject: Using the harbor rss

Your Tags: Add tags
Popular Tags: [View All]
The pirate's life for me!
United States
flag msg tools
pineapples are in my head
badge
"There is a Heppy Land Furfur A-waay"
Avatar
mbmbmbmbmb
I've been playing craploads of 4 player PR games against 3 computer opponents.

Last game, I just took a few Indigo/Corn plantations, and generally saved up my money.

I got the large market, to save up money.

Then I bought the harbor, and so every captain phase I was getting double the VPs.

I finished with the two piece building Customs House.


My final score was 53. The second placer had a score of 39 or something.

I'm sure everyone asks this, but how do you cope with this kind of strategy? I'll try it a few more times, but if the computers can't evolve to stop me from winning, I'll stop using the harbor :/

Quite simply, what is the strategy against this?

LOVE THE GAME by the way. I refuse to read strategy articles

Kris24
 
 Thumb up
 tip
 Hide
  • [+] Dice rolls
Chris Linneman
Canada
Vancouver
BC
flag msg tools
Avatar
mbmbmbmbmb
So you refuse to read strategy articles but post strategy questions in the strategy forums? I see.

I'm not sure which program you are using to play PR but I found the commercially available "official" one to be quite weak. I was able to win consistently as an almost pure PR newbie (after only a couple face-to-face games). If you can, try to find some real opponents.

That said, I find the Guild Hall strategy to be a strong contender for beating corn shippers. Basically you try to diversify your production, build the Factory, and leech off the shippers' craftsmen to get income to buy multiple big buildings, including the Guild Hall (which can score up to 10 additional VPs). You end the game on buildings before the shippers can get too many VPs from shipping.
5 
 Thumb up
0.01
 tip
 Hide
  • [+] Dice rolls
Chris Ferejohn
United States
Mountain View
California
flag msg tools
badge
Pitying fools as hard as I can...
Avatar
mbmbmbmbmb
Well, first of all, harbor doesn't double your VPs. I'll assume that was just short hand.

As for countering it:
* Don't let the player get a bunch of corn.
* Don't ever ever ever call produce just before the player.

Out of curiosity, what are the computer players doing? If 2 people are trying this, it makes it a lot harder (since one of them will tend to half fill the boats, fill up the trading post, etc.). This strategy should lead to your being pretty cash poor, and still you managed to get enough for a large market and a harbor. If opponents are playing properly, the game should be almost over by the time you get the harbor...
5 
 Thumb up
0.01
 tip
 Hide
  • [+] Dice rolls
United Kingdom
Southampton
Hampshire
flag msg tools
badge
I'll think of something witty to put here...
Avatar
mbmbmbmbmb
Kris2476 wrote:
Quite simply, what is the strategy against this?

LOVE THE GAME by the way. I refuse to read strategy articles




Facetious answer: Go read some strategy articles.

Short answer, either 1. don't call Captain or Craftsman (instead go build/trade with quarries) or 2. Corn Wharfing.
 
 Thumb up
 tip
 Hide
  • [+] Dice rolls
The pirate's life for me!
United States
flag msg tools
pineapples are in my head
badge
"There is a Heppy Land Furfur A-waay"
Avatar
mbmbmbmbmb
First off, I was reading a few strategy articles before this thread. I guess I was a bit shocked by the "choose captain, and you lose 3 VPs, right now, on round 2".

So I really found some of the articles interesting (Such as general opening, middle, and endgame strategies), but once the player starts breaking the game down into individual actions, and what could have been changed to raise themselves 1 or 2 VPs, then I stop reading.



cferejohn wrote:
Well, first of all, harbor doesn't double your VPs. I'll assume that was just short hand.

As for countering it:
* Don't let the player get a bunch of corn.
* Don't ever ever ever call produce just before the player.

Out of curiosity, what are the computer players doing? If 2 people are trying this, it makes it a lot harder (since one of them will tend to half fill the boats, fill up the trading post, etc.). This strategy should lead to your being pretty cash poor, and still you managed to get enough for a large market and a harbor. If opponents are playing properly, the game should be almost over by the time you get the harbor...



I think it's a really weak computer. It uses CURL software, and it came up first for a "play PR online" Google search, so it fit my credentials.

So I'm not sure if the computers were really giving me a run for my money. I'm reluctant to play with other people online because of that over-analysis of the game.

But thanks for the replies, everyone. I'm sure I'll learn things as I play more.

Teehee! Now I MUST play more, you understand.

Kris24
 
 Thumb up
 tip
 Hide
  • [+] Dice rolls
Chris Gibbs
United Kingdom
Bath
flag msg tools
mbmbmbmbmb
If you want a decent PR challenge against purely computer opponents, set all players to the hard AI at www.tropiceuro.com

(note that a separate much easier AI is also available in the game, which is based on the phial.com AI with some improvements)
 
 Thumb up
 tip
 Hide
  • [+] Dice rolls
Henry Wong
msg tools
The tropic euro AI is generally pretty good, although I feel like it plays some areas wrong:

1) It over values (IMO) shipping vs trading in the early game (turn 2 in particular). For example, it's very common to see P3 w/ two corn and a small market, call mayor/equipper and put his two guys on both corn. This seems silly to me.

2) It's end game calculations aren't good. For example, the very last player often produces rather than manning an unoccupied large building (w/ no other unoccupied large buildings on the board), or doesn't build despite having more money than anyone else. In general I win my fair share of 1 or 2pt games, which you wouldn't expect - an AI should crush a human at these sorts of limited space problems.

3) It waaaaay over produces in general. For example, it often produces and is then immediately shut out of the boats during the following shipper.
 
 Thumb up
 tip
 Hide
  • [+] Dice rolls
Chris Gibbs
United Kingdom
Bath
flag msg tools
mbmbmbmbmb
Hi Henry, you are referring purely to the CPU Hard players right? From the general comments, it appears as though you are referring to the CPU Normal AI (or a mixture of the two).

I spent quite a bit of time hardcoding the optimal play for endgame conditions for the hard AI, especially with regard to manning unoccupied large buildings, so I guess you may be referring to the CPU Normal, which is considerably less advanced.

Regarding point 3), again that seems to be a criticism of the normal AI, as the hard AI is very good at a producing/shipping strategy, especially for positions 3 and 4 in a 4 player game. (Positions 1 and 2 will usually go for a building strategy)
 
 Thumb up
 tip
 Hide
  • [+] Dice rolls
Henry Wong
msg tools
I'm definitely talking about the hard AI. I've actually never played the normal AI. I don't know if you can break out stats per human player but my name on TropicEuro is henrywong2003.

So for example, the last player on the last round frequently produces. Since number of goods is the second tie break, this isn't usually the best choice - for example if the computer can afford the same tier of building as whoever is in the lead then it's better to pick builder, which nets you $1 relative to the leader. It also doesn't seem to typically max out its large building bonuses (eg: picking "settle" when it has the residence).

re: shipping - Here's something you can fairly easily reproduce (comes up about once a game for me). If produce has $1 on it and no other roles have any money and the AI is the last player to go in the round then even in the mid/end game the AI will often pick produce (at least in 3p, which is what I usually play). It's hard to make blanket statements like "this is always wrong" but... um... it's usually wrong. =)

Anyway, not trying to be overly critical - the AI in general is pretty good and it's a lot of fun to play against. Keep up the good work! =)
 
 Thumb up
 tip
 Hide
  • [+] Dice rolls
Chris Gibbs
United Kingdom
Bath
flag msg tools
mbmbmbmbmb
henrywong2003 wrote:
for example if the computer can afford the same tier of building as whoever is in the lead then it's better to pick builder, which nets you $1 relative to the leader.

Yeah at the moment the hardcoded endgame behaviour doesn't do as extensive analysis as this, but I can look to adding this in future. Most of the behaviour at moment for the final round is to deal with whether to pick equipper or not and which building to choose, but things like this and also projected shipping relative to other players are possible future improvements.

henrywong2003 wrote:
It also doesn't seem to typically max out its large building bonuses (eg: picking "settle" when it has the residence).

I've hardcoded them to choose planter during the last round if the residence is occupied, but during the game the actual behaviour is down to the "DNA" of the player.

Basically, the AI uses genetic algorithms, with a set of players that have evolved strategies over thousands of games in each position, so the actual strategies used by each player across different games may be quite a bit different, depending on which underlying DNA is being used (chosen at random for each position).

The full source code for the hard AI is available by the way, in the form of an Excel spreadsheet - please drop me an email or Geekmail if you would like me to send over the source.

henrywong2003 wrote:

re: shipping - Here's something you can fairly easily reproduce (comes up about once a game for me). If produce has $1 on it and no other roles have any money and the AI is the last player to go in the round then even in the mid/end game the AI will often pick produce (at least in 3p, which is what I usually play). It's hard to make blanket statements like "this is always wrong" but... um... it's usually wrong. =)

Yeah there will be situations like that, but quite often the DNA behaviour will choose to pick a role with the most cash on it, even if it isn't necessarily the best overall choice. Also note that the players that start with banana will often heavily favour a produce/shipping strategy, so may well pick producer a lot more often than you might expect. Judging by the stats, that strategy seems to work well for them

henrywong2003 wrote:

Anyway, not trying to be overly critical - the AI in general is pretty good and it's a lot of fun to play against. Keep up the good work! =)


Yeah no problems, the main aim was to create a separate AI that was significantly harder than the normal AI and also competitive against decent experienced players, and looking at the stats it seems as though this has been achieved, but there is still plenty of room for improvement

Chris
 
 Thumb up
 tip
 Hide
  • [+] Dice rolls
Sven Uilhoorn
Netherlands
flag msg tools
paulclarke339 wrote:
Kris2476 wrote:
Quite simply, what is the strategy against this?

LOVE THE GAME by the way. I refuse to read strategy articles




Facetious answer: Go read some strategy articles.


Maybe because he wants to enjoy the game and explore different facets of it before reading everything about it without a fun introduction. Maybe for the same reason I have: to not become to dominant/experienced when playing with new players. Out of the six times I played, each had one or two newbies out of 3-5 players.

It annoyed me too because I didn't read the italic text in the game manual and thought every delivery to the ships got double the points. Since I'm the one with the game in my friends group and we all assumed that was how the harbour should've been interpreted, that resulted in the last 5 winners all having used the harbour because they were unstoppable point scorers.
 
 Thumb up
 tip
 Hide
  • [+] Dice rolls
Front Page | Welcome | Contact | Privacy Policy | Terms of Service | Advertise | Support BGG | Feeds RSS
Geekdo, BoardGameGeek, the Geekdo logo, and the BoardGameGeek logo are trademarks of BoardGameGeek, LLC.