Recommend
5 
 Thumb up
 Hide
15 Posts

Race for the Galaxy» Forums » Sessions

Subject: Race for the Galaxy , myself verses the computer six times in a row rss

Your Tags: Add tags
Popular Tags: [View All]
Charlie Chuckle
United States
California
flag msg tools
Avatar
mbmbmbmbmb
Initial setup 36 VP’s to end the game or 12 settle/developes

Blue(me) = New Sparta +2 military
Red = Alpha Centauri
Green = Earth’s Lost Colony

DevL1’s 4 out of 12 good (New Mil Tact/Exp.Force)
DevL2’s 2 out of 12 good (Space Marines)
DevL3’s 0 out of 4 good
DevL4’s 2 out of 8 good (Drop Ships)
DevL5’s 0 out of 2
DevL6’s 2 out of 12 (Gal Imp/NewGalOrd)

SetL0-1 0 out of 8 good ones
SetL2-3 4 out of 19 (EpsErid/BlaGemMine/AlienRos/GalResort)
SetL4-6 4 out of 12 (ImpArm/GalTrend/NewEarth)

RedL1 = 6
RedL2 = 5
RedL3 = 3
RedL4 = 2
RedL5 = 2
RedL6 = 2
RedL7 = 1

A total of ten developments that I want out of 50 and 8 good settlements out of 39 left.
I’ll consider all 21 conquerable worlds decent.

Round 1
Blue conquers OutLaw World
Red settles a Space Port
Green settles Gem World
Red and Green consume and produce (those cheaters must be working together)

Round 2
Blue, Green and Red decide to explore for more supplies.
Nothing special happens.

Round 3
Blue decides to Develop Drop Ships for a huge bonus to his military.
Green decides to explore again for supplies.
Blue’s Drop Ships get made, and Red drops a huge load for Trade League (big points for trade power cards). Green passes because his explore failed to net him anything good, I assume.
Red’s original choice was to produce filling up everyone with goods, except Blue who is the warrior.

Round 4
Blue and Green moves to settle/conquer, and Red trades.
Blue conquers a Rebel Underground (saving Aquatic Uplift for later), and Green sets up a huge goodie Galactic Trendsetters (double VP’s).
Oh crap, I shoulda made the Aquatic Uplift… What was I thinking.. I could have sold the goods for VP and a card… grr… (mistake #1)

Round 5
Blue Settles, Green Produces(he’s doubling up), and Red Developes.
Blue conquers an Aquatic Uplift Race, Red terraforms a world, Red produces a Space Port, and then Green produces preparing to get double VP’s next turn.

Round 6
Red produces, Green consume(x2), and Blue preps to Produce to gain on a windfall.

Round 7
Red and Green decide to double up on consume(x2) while Blue produces to join in the fun of production.
Red settles a Merchant World, and Blue settles a Distant World for more goodies.
Red gains 4VP, Green gains 6VP, while Blue only gets 1 VP for consuming.
The scores: Blue 14, Red 20, Green 20
Tokens given out 2/4/15. So 15 VP’s till the end of the game. Green is grinding me down, and I have nothing to compete with his hording of VP’s

Round 8
Everyone Produces. No change.

Round 9
Blue Consume Trades, Red and Green Explore.
2 VP’s to Red, 3 VP’s to Green, 1 VP to Blue. 9 VP will end the game.

Round 10
Red and Blue develop, while Green Produces.
Blue places his big Galactic Imperium (not even close to enough points to catch up)
Red builds a New Economy for huge bonuses.

Final Round 11
Red chooses explore, Green consumesx2(to end the game), and Blue settles.
Red makes a Primative World, Green makes Blaster Gem Mines, and Blue conquers an Alien Robot Sentry.
Red gains 2VP, Green hordes 8VP, and Blue gains 1VP. (all the VP chips are exhausted, end of game!)

Red wins with 37. Green in 2nd with 36, and Blue last place with 24. What the heck happened!!

Blue/Red/Green Rematches:
Game 2: 24/35/33 Red wins. 11 short
Game 3: 24/33/23 Red wins. 9 short
Game 4: 25/31/39 Green wins. 14 short
Game 5: 19/30/36 Green wins. 17 short
Game 6: 20/26/28 Green wins. 8 short

To be continued…

I tried beating the computer six times total and closest I got was within 8 points. Ew.. somethings strange. I think they are cheating. Lol, who knows.

The strange thing is that every attempt ended because we ran out of VP chips, not because someone made 12 development/settlements.

The basic strategy, at least from what I noticed from the computer's style of play. The computer would build up about two good's producing planets, then start it's deadly cycle of produce, consumex2, produce, consumex2.... The worst thing... and I'm kinda suspicious about it.. Is that the computer seemed to be double teaming me, they would produce/consume in the same turn. I rarely saw both computer players picking the same job. They must be programmed to predict each others move. I'm calling them cheaters, and we'll see how I do verses human players in the future.
3 
 Thumb up
 tip
 Hide
  • [+] Dice rolls
Al Sefcik
United States
Bolingbrook
Illinois
flag msg tools
Co-op??? NOOOOO!!!
Avatar
mbmbmbmbmb
I've beaten the AI about twice out of a hundred tries. Either I'm the worst RFTG player on the planet, or the program cheats.
I'm going with the latter...
1 
 Thumb up
 tip
 Hide
  • [+] Dice rolls
Kester J
United Kingdom
flag msg tools
badge
Avatar
mbmbmbmbmb
The AI doesn't cheat as such. I suspect there is a bit of "groupthink", where the AIs know how each other play (the same as you would with a regular opponent) and can take advantage of it. The AI is certainly a lot easier to beat in 2-player than multiplayer, although I'm not sure whether it's relatively easier once you account for the lower expected win rate.

As a general piece of advice, the AI tends to be very cautious in multiplayer, going more for high-quality tableaus or produce/consume engines than fast rushes. This means that it's a good idea for you to be equally cautious. If you aren't, you'll often be the only one calling develop or settle on a given turn while others draft off it and refill their hands via trade or produce. The result is that you end up in a poor position for the next turn without having got ahead in tableau size.
4 
 Thumb up
0.25
 tip
 Hide
  • [+] Dice rolls
Filipe "Faol" Henriques
Portugal
Lisbon
flag msg tools
mbmbmb

There is something strange about choosing the actions. I've seen the computer choosing consume trade, with no products in its worlds!!! if there is no settle in that round, then its a complete waste...unless the second pc player gives a little help!
1 
 Thumb up
 tip
 Hide
  • [+] Dice rolls
Serge Levert
Canada
Vancouver
British Columbia
flag msg tools
badge
Avatar
mbmbmbmbmb
Kester wrote:
As a general piece of advice, the AI tends to be very cautious in multiplayer, going more for high-quality tableaus or produce/consume engines than fast rushes.

Interesting, i wonder if that's something i should apply to my large multiplayer games. I'm such a rush fiend, sometimes when i hit ~10 cards and have ~8 points, i'm like "uh oh". :P
3 
 Thumb up
0.25
 tip
 Hide
  • [+] Dice rolls
Mark Delano
United States
Stamford
Connecticut
flag msg tools
mbmbmbmbmb
entranced wrote:
Interesting, i wonder if that's something i should apply to my large multiplayer games. I'm such a rush fiend, sometimes when i hit ~10 cards and have ~8 points, i'm like "uh oh".


In a multiplayer game the difference between one person rushing and two people rushing can be huge, much bigger than between two and three. In that sense it depends quite a lot on the choices made by your opponents, and another reason why I prefer to play a series of games to just one or two. I feel like I'm just learning their play style in those first couple of games.
2 
 Thumb up
 tip
 Hide
  • [+] Dice rolls
Eric Jome
United States
Franklin
Wisconsin
flag msg tools
designer
badge
Avatar
mb
I love this thread!

People... guess what? The computer isn't cheating or using a gimmick to win. It's just very, very good at the game. I've played hundreds and hundreds of times against it. I've gone from being a pretty good player to being well... people don't like to play me anymore in real life because they say I'm too good.

Watch how it plays. See if you can learn something from it. The first big lesson you'll notice is - when someone plays Settle or Develop and you have a card you can play, you play it. It doesn't matter if it is good or bad or if you were saving for something else. Play it.

Another surprising lesson? Never play Settle or Develop.

I love you Keldon Jones. You've made me 1000x better as a player.
3 
 Thumb up
 tip
 Hide
  • [+] Dice rolls
Jean-Denis Gill
Canada
flag msg tools
mbmb
I just start to test the AI. In 1vs1 I'm able to beat it 7 times in 10 games.

I gonna try multi but I'm not use to play this way.
2 
 Thumb up
 tip
 Hide
  • [+] Dice rolls
Eric Jome
United States
Franklin
Wisconsin
flag msg tools
designer
badge
Avatar
mb
teuehikan wrote:
I just start to test the AI. In 1vs1 I'm able to beat it 7 times in 10 games.

I gonna try multi but I'm not use to play this way.


I also find it very easy to beat it in one on one play. You have a great deal more control over how the game turns out. Of all actions selected, you select 50% of them and you can do combos. The game is much, much harder in multiplayer.

I typically play with Rebel V Imperium, takeovers on, with goals, 3 players. This gives, I feel, the least opportunity for me to beat the computer... that is, the hardest way to play. With more players, there is more chance for bonus actions I might use to get ahead or for the AI to trip over each other and delay the game end in my favor.
1 
 Thumb up
 tip
 Hide
  • [+] Dice rolls
Charlie Chuckle
United States
California
flag msg tools
Avatar
mbmbmbmbmb
There is an option to undo a turn if you suspect the computer is cheating. Since the computer never changes it's choices, even after you change yours... It's a fact that it's not cheating based on your decisions.

They do however... Seem to be conspiring together as a team. Since, they do know each other very well. Therefore, it is a 2 vs 1 game.
1 
 Thumb up
 tip
 Hide
  • [+] Dice rolls
Charlie Chuckle
United States
California
flag msg tools
Avatar
mbmbmbmbmb
Ok, I've played about 5 more times in 2 player mode.

The computer is very predictable. If you have no producing worlds, the computer will assume you will settle. Basically, it might produce even if it can't produce anything at all. The computer assumes you will settle, giving the computer a chance to settle and produce a good in one turn.

Also, the computer will produce while full on goods. Meaning, if you don't consume, then the computer wastes a turn. From what I understand, the computer only does this if you have goods to consume, and your good producing worlds are less than their own.

So, logically; correct choice is to counter the computer's moves by playing explore, or producing with him. It forces the computer to waste a turn producing nothing.

You can always undo the turn if you suspect the computer of cheating. However, it always chooses the same option whenever you redo the turn. Meaning the computer is betting with the best odds of winning. Somebody did a lot of math to predict the percentage of winning choosing this vs that. However, being able to predict the computer will always choose the same thing (it always goes with the best odds) is it's weakness.
1 
 Thumb up
 tip
 Hide
  • [+] Dice rolls
Charlie Chuckle
United States
California
flag msg tools
Avatar
mbmbmbmbmb
Oh, I forgot. When you have one to ZERO cards in hand, the computer will settle/develop 100% of the time.

I had one problem that I couldn't figure out. Is there any reason why I cant develop Expedition Force. That is pure cheating if I have 6 cards, choose develop, and the computer wont let me develop this card... I don't get it.

I can post a picture of it later, if you like, the card is dimmed but other develop cards are not.
1 
 Thumb up
 tip
 Hide
  • [+] Dice rolls
Brendon Russell
New Zealand
Auckland
flag msg tools
badge
Avatar
mbmbmbmbmb
Charliechuckleberry wrote:
I had one problem that I couldn't figure out. Is there any reason why I cant develop Expedition Force. That is pure cheating if I have 6 cards, choose develop, and the computer wont let me develop this card... I don't get it.

I can post a picture of it later, if you like, the card is dimmed but other develop cards are not.


Sounds like you'd already developed Expedition Force previously. You can't have more than 1 copy of the same development.
4 
 Thumb up
 tip
 Hide
  • [+] Dice rolls
Guy Srinivasan
United States
Kirkland
Washington
flag msg tools
Avatar
mbmbmbmbmb
Charliechuckleberry wrote:
However, being able to predict the computer will always choose the same thing (it always goes with the best odds) is it's weakness.

Yes, one thing that might improve the AI's play is adding a "compute the best mixed strategy" layer. Past a certain level of human play, this should do better versus humans. (Below that level, it should do worse.)

Charliechuckleberry wrote:
I had one problem that I couldn't figure out. Is there any reason why I cant develop Expedition Force. That is pure cheating if I have 6 cards, choose develop, and the computer wont let me develop this card... I don't get it.

I can post a picture of it later, if you like, the card is dimmed but other develop cards are not.

Prediction: you have already developed one copy of Expedition Force, and the second is illegal.

Edit: What Brendon said. Edit: Brenden's name is not spelled that way.
3 
 Thumb up
 tip
 Hide
  • [+] Dice rolls
Charlie Chuckle
United States
California
flag msg tools
Avatar
mbmbmbmbmb
scwont wrote:
Charliechuckleberry wrote:
I had one problem that I couldn't figure out. Is there any reason why I cant develop Expedition Force. That is pure cheating if I have 6 cards, choose develop, and the computer wont let me develop this card... I don't get it.

I can post a picture of it later, if you like, the card is dimmed but other develop cards are not.


Sounds like you'd already developed Expedition Force previously. You can't have more than 1 copy of the same development.


Damn, I just noticed that! I already had that development out. That's why I couldn't play it.
1 
 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.