$10.00
Recommend
5 
 Thumb up
 Hide
20 Posts

7 Wonders» Forums » Strategy

Subject: Ranking the civilzations special abilities: Ephesus B rules! rss

Your Tags: Add tags
Popular Tags: [View All]
Henrik Johansson
Sweden
Järfälla
flag msg tools
mbmbmbmbmb
I have developed a way to compare the civilization special abilities during realistic game conditions, and like to share the results here. The game is very well balanced, don't get me wrong. However, under comparable circumstances the Ephesus B civilization really shines compared to the others. They are listed in order:
The best: Ephesus B. For each wonder step, this civilization receive $4 along with the 2, 3, 4 VP. The wonder steps are cheap, 2 stone, then 2 wood, then (Papyrus) Loom and Glass. The Papyrus is not really a cost at all, since this civilization has a built-in Papyrus, and the way costs are calculated and material regenerated, Papyrus+Glass+Loom is equally cheap as Glass+Loom!
The good ones: Alexandria B, Ephesus A, Rhodes B, Giza B.
Alexandria B gets free materials and goods. Ephesus A has a middle wonder stage producing $9, very useful for buying other VP stuff. Rhodes B has only 2 wonder steps, but producing $3 and $4 along with VP and Military force, much like Ephesus. Giza B has a 4:th +7 VP wonder stage, but can choose to build only 3+5+5 VP if stone is expensive to import.
The worst ones: Babylon A, Babylon B, Giza A. Babylon B has its first wonder stage prized at Loom+Clay, but the built-in Clay makes the actual cost only 1 Loom! However, this is only 3 VP. The middle stage 7:th card extra round is not very useful, the good cards are never there for the 7:th round. Babylon A has a weak 2:nd wonder unless the Green path to victory is decided very early. Giza A has too expensive wonder stages relative to the pay-out in VP.

The evaluations may seem to overvalue money a bit. 4 dollars is less than 2 points at the end of the game. The idea is to use the money for import of resources and use those for buildings, not to cash in dollar to VP.
2 
 Thumb up
 tip
 Hide
  • [+] Dice rolls
ronaldinho @boardspace.net
Taiwan
flag msg tools
mbmb
I found Ephesus (Artemis) B to be the worst amongst the B sides.

I've developed a general strategy around each of the 7 B sides, and my strategy for Ephesus pretty much involves my not using its special abilities. I use them as cheap bail outs when I have nothing better to play on a hand. What I do take advantage of with Artemis B is its starting Papyrus and try to go Green. The cheap bail outs work with a Green strategy because when one goes Green, he won't have a lot of Browns so the cheap wonder plus 4 coins for flexibility work nicely.
 
 Thumb up
 tip
 Hide
  • [+] Dice rolls
Flame Bird
msg tools
What is this comparison you have developed?

I think the A-sides are weaker because fewer abilities provides less variance that a perceptive player can exploit. Also, I don't like how several A-sides require double manufactured goods for the last Wonder stage. With no other cards requiring doubles of the same manufactured good they are cumbersome.

The B-sides however are all interesting in their own right. I've won games with all and lost games with all. Personally I find Rhodes the hardest to play, and like Ephesus, but overall the B-sides are close enough in power that the hand distributions are more important than their inbuilt biases.
1 
 Thumb up
 tip
 Hide
  • [+] Dice rolls
Evil Roy
United Kingdom
Sutton
Surrey
flag msg tools
mbmbmbmbmb
sthrjo wrote:
I have developed a way to compare the civilization special abilities during realistic game conditions, and like to share the results here.


I'd be interested to know what your way of comparing the wonders is. It very easy to be subjective (and my subjective opinion on the strengths of the wonders differs sharply from your results in places).
1 
 Thumb up
 tip
 Hide
  • [+] Dice rolls
ronaldinho @boardspace.net
Taiwan
flag msg tools
mbmb
Speaking of subjectivity, I find the OP's evaluations to overvalue money a bit. 4 dollars is less than 2 points at the end of the game.
1 
 Thumb up
 tip
 Hide
  • [+] Dice rolls
Henrik Johansson
Sweden
Järfälla
flag msg tools
mbmbmbmbmb
Evil Roy wrote:
sthrjo wrote:
I have developed a way to compare the civilization special abilities during realistic game conditions, and like to share the results here.


I'd be interested to know what your way of comparing the wonders is. It very easy to be subjective (and my subjective opinion on the strengths of the wonders differs sharply from your results in places).


This is my objective evaluation method: I have developed a variant to play the civilizations one by one, see Solitaire exercise play, using two passive left and right opponents, otherwise by the original 3-player rules. Then I have developed a software that execute such solo-civilzation games, using an exhaustive search for the optimal solution that produce the most VP, and coins for tiebreak. The software tries each card in each order, and also tries to sell them or build wonders in all combinations.
Then I have selected 18 random draws of 7+7+7 cards, and let each civilization find its optimal play of 6+6+6 cards. This gives a variation caused only by the civilization's built-in resource and its wonder stages.
I then post-process these 18x14 sub-results, and scrutinize some of the games by hand to look for bugs. The outcome is clear: Ephesus B gets the highest result a lot of the time.
 
 Thumb up
 tip
 Hide
  • [+] Dice rolls
ronaldinho @boardspace.net
Taiwan
flag msg tools
mbmb
sthrjo wrote:
Evil Roy wrote:
sthrjo wrote:
I have developed a way to compare the civilization special abilities during realistic game conditions, and like to share the results here.


I'd be interested to know what your way of comparing the wonders is. It very easy to be subjective (and my subjective opinion on the strengths of the wonders differs sharply from your results in places).


This is my objective evaluation method: I have developed a variant to play the civilizations one by one, see Solitaire exercise play, using two passive left and right opponents, otherwise by the original 3-player rules. Then I have developed a software that execute such solo-civilzation games, using an exhaustive search for the optimal solution that produce the most VP, and coins for tiebreak. The software tries each card in each order, and also tries to sell them or build wonders in all combinations.
Then I have selected 18 random draws of 7+7+7 cards, and let each civilization find its optimal play of 6+6+6 cards. This gives a variation caused only by the civilization's built-in resource and its wonder stages.
I then post-process these 18x14 sub-results, and scrutinize some of the games by hand to look for bugs. The outcome is clear: Ephesus B gets the highest result a lot of the time.
I am not convinced.
 
 Thumb up
 tip
 Hide
  • [+] Dice rolls
Evil Roy
United Kingdom
Sutton
Surrey
flag msg tools
mbmbmbmbmb
sthrjo wrote:
This is my objective evaluation method: I have developed a variant to play the civilizations one by one, see Solitaire exercise play, using two passive left and right opponents, otherwise by the original 3-player rules.


Your method sounds interesting and is certainly not subjective. However, I would be wary of drawing conclusions from this as your solitaire play is significantly different from a multi-player games of 7 Wonders.
1 
 Thumb up
 tip
 Hide
  • [+] Dice rolls
Henrik Johansson
Sweden
Järfälla
flag msg tools
mbmbmbmbmb
Let me share two results out of the 18x14 ones, the same cards dealt to Ephesus B and Giza A:
Age I:
1: Ore Vein, 2: Glassworks.
3: Ephesus builds East Trading Post, Giza sells it for $3.
4: Theater, 5: Stockade, 6: Barracks, (7: discard Altar).
Both have 2 military strength, and beats left passive opponent, 1 VP.
Age II:
1: Brick Yard, 2: Quarry.
3: Ephesus builds Temple B1, Giza builds pyramid A1. They put Dispensary underneath.
4: Ephesus builds Temple B2 using Press, Giza builds Press.
5: Statue (chained), 6: Walls, (7: discard Laboratory).
Both get 6 VP for beating both left (3 mil.) and right (2 mil.).
Age III:
1: Ephesus builds Temple B3 using University, Giza sells it for $3.
2: Ephesus builds Builders Guild, Giza sells it for $3.
3: Gardens (chained), 4: Pantheon, 5: Palace, 6: Fortifications (chained), (7: Discard Study).
Both get 10 VP for beating left and right.
Result: Ephesus B: 56 VP and $1, Giza A: 46 VP, $2.
The money changes for Ephesus B: $3 from start, after Stockade $2, after Brick Yard $1, Quarry $0, B1 $4, B2 $5, Walls $4, B3 $6, after Pantheon $4, after Palace $1.
Giza A: $3, after sale $6, Stockade $4, Brick Yard $3, Quarry $2, after 2 sales $8, after Pantheon $6, Palace $2.
 
 Thumb up
 tip
 Hide
  • [+] Dice rolls
Mark Levine
United States
New York
New York
flag msg tools
mbmbmbmbmb
The problem is that the solitaire exercise is too different from the real game to make the results comparable. Your example has Giza discarding 3 times for money [including twice in the 3rd age!]; while perhaps valid for solitaire, that would be horrid in an actual game. Because in the solitaire game the passive players never give you money for resources, and with the passing hands [vs the static hand] more likely to get a chance to build a yellow card for money or money/points, the money from Ephesus is going to be relatively more powerful.

This is also why Rhodes B does well, and Alexandria B should do well because it guarantees you access to resources you might not get with the static draw and should save money. The static hands also make science / the 7th card worse, so it makes sense Babylon does badly, but I'm surprised Giza B does well given it's rather expensive, especially if Giza A is doing badly. I'd also expect Zeus A? (3/free building/7) to do well (about the same as Ephesus A) because of the cheap wonder costs, and the free building is nice in this format.
6 
 Thumb up
 tip
 Hide
  • [+] Dice rolls
ronaldinho @boardspace.net
Taiwan
flag msg tools
mbmb
Also, side As are shit. I'd only compare side Bs with side Bs.

I am pretty sure there are other issues with this excercise as well. I doubt real games would play out like your simulations.
1 
 Thumb up
 tip
 Hide
  • [+] Dice rolls
Henrik Johansson
Sweden
Järfälla
flag msg tools
mbmbmbmbmb
KoolKow wrote:
The problem is that the solitaire exercise is too different from the real game to make the results comparable. Your example has Giza discarding 3 times for money [including twice in the 3rd age!]; while perhaps valid for solitaire, that would be horrid in an actual game. Because in the solitaire game the passive players never give you money for resources, and with the passing hands [vs the static hand] more likely to get a chance to build a yellow card for money or money/points, the money from Ephesus is going to be relatively more powerful.
[snip]

Thanks for the response. I have thumbed it.
My goal is of course to make this as objective and 3-player-look-alike as possible. I strongly concider adding export income to the simulation, that would keep more money around, perhaps ending Ephesus B's dominance. But I can not see how Giza's discarding is horrid? Its the optimal tactic: "Shall I build this University for 1 green VP, or sell it and perhaps afford a Palace or the Pantheon later on?". If anything, it is unrealistically good insight into the yet unseen hands.
And I dont think my simulation is too different, different yes, agreed. See A (Nearly) Resourceless Victory, it looks almost as if my simulation made the report. I actually has started to recreate it.
 
 Thumb up
 tip
 Hide
  • [+] Dice rolls
Mark Levine
United States
New York
New York
flag msg tools
mbmbmbmbmb
It's certainly valid for the simulation/solitare version. But it's horrid for a real game because discarding for money is a 1 point play, when in age 3 you want everything to be 5+ points. Sometimes it's unavoidable [don't get a useful yellow card in the first few turns, get passed a bunch of useless red/yellow/green cards], but if it happens twice you're not going to win. And then there's the fact you're discarding your first two cards of the age for money; if you did that in a real game, you're probably not going to get a chance to build the university/pantheon after that, as the other players built the best stuff already.

The big differences between the simulation and the resourceless victory are that the discards for money are at the end of the second age, the player is able to play a cards for points+money, and science is actually viable in the resourceless victory since all the science cards are in the game [simulation only has expected value of 4 science cards + the odds of the science guild].
1 
 Thumb up
 tip
 Hide
  • [+] Dice rolls
ronaldinho @boardspace.net
Taiwan
flag msg tools
mbmb
Okay I just read your method of simulation as linked.

Any conclusions drawn from such a method is pretty irrelevant. It is not realistic. At all.
1 
 Thumb up
 tip
 Hide
  • [+] Dice rolls
Henrik Johansson
Sweden
Järfälla
flag msg tools
mbmbmbmbmb
The static hand with few green cards is resembling the fact that the other players goes green, and not much green cards comes your way. That way the static hands resembles the rotating hands with three players: You just can't build whatever you like, but perhaps 7 cards to choose from is too little. You actually get to see 7+6+5 cards in real play, and then 4+3+2 already seen cards. That would perhaps break Ephesus B's dominance to have a larger static hand! I will try more, but the computing effort will increase.

Discarding in the beginning of the age is just an optimization to save iteration cycles, not the recommended way in real play. The software just gathers the $3 discards at the beginning, with static hands the ordering of the builds/discards is always the one reaching the optimal result, and discarding late never contributes towards the optimum, so the software just discards in the beginning. I could change the simulation into discarding as late as possible, keeping the optimal outcome intact, it would certainly look better in writing.

I have done some initial tests keeping the money in the game as a fair amount of export income: The result does not change anything, Ephesus B just gets more money too, converting it into more VP.

Discarding for money very late is a 1 VP play, but discarding early is a way to prepare for building VP using money for import of material and goods, when your opponents does not provide the money for you.

The optimal play that my evaluation method find is not a guaranteed win in a real game. But it is the best you can do under the circumstances! So discarding twice early in age 3 gives you the most VP you can get, not a guaranteed win.

In theory one could do a much better investigation, but in reality the computing process sets a limit to what can be done. I'm quite pleased myself with the achievement so far, and thankful for suggestions for improvement.
 
 Thumb up
 tip
 Hide
  • [+] Dice rolls
ronaldinho @boardspace.net
Taiwan
flag msg tools
mbmb
Quote:
The optimal play that my evaluation method find is not a guaranteed win in a real game. But it is the best you can do under the circumstances! So discarding twice early in age 3 gives you the most VP you can get, not a guaranteed win.
I'll take your word for it. Which is why the simulated situation is unrelated to reality because, unlike the simulation, that is not the way to net the most VP in an actual game.

Quote:
I'm quite pleased myself with the achievement so far
As long as you've convinced yourself.

Quote:
in reality the computing process sets a limit to what can be done.
Pretty much.
1 
 Thumb up
 tip
 Hide
  • [+] Dice rolls
Dave F.
Netherlands
Schelluinen
Zuid-Holland
flag msg tools
mbmbmbmbmb
The majority of strategy from the game comes from interaction with other players and your simulation is severely lacking in that aspect.

In a live game, you may receive $ for your resources and spend some of it to gain resources with a positive VP expectation. Your opponents may also make decisions which are sub-optimal for themselves, but hamper you too. If neither of your opponents builds any resources you can use, you're pretty much done. I don't see bots making trade alliances easily too. It may happen in a game that you are constantly buying X from a player while that player is constantly buying Y from you, providing a sort of symbiosis. You might go an undeniably hard all green strategy and your opponents could notice this and start chucking greens,

I think the finesse of the game is in spotting these little things and acting and reacting accordingly. An algorithm which only maximizes your own VP count is too limited, you need an algorithm which also limits your opponents' VP counts,

For expert players, B boards will most of the times be better because it gives them more choice. An A side will only give you something average, while a B side may give you something wonderful or something awful.

Besides, I highly value winning in any game. I'd rather get 1st and 3rd than two 2nds, even though they average the same. I'd rather get 1st and 4th than two 2nds!
1 
 Thumb up
 tip
 Hide
  • [+] Dice rolls
Henrik Johansson
Sweden
Järfälla
flag msg tools
mbmbmbmbmb
I'm happy to see some indirect support for my claim in threads Some Complex Strategies For 7 Wonders and Analysis of the various wonders from our games this year where proper games have been analyzed: My top and bottom ranking have quite a good correlation with these threads (except for the simplified headline claim that "Ephesus B rules"). From the followups above in this thread the impression might emerge that there is little correlation between my presented results and real gameplay. I say it's more than that: I say it's quit a good correlation!
Another good correlation: http://www.boardgamegeek.com/article/10978409#10978409
 
 Thumb up
 tip
 Hide
  • [+] Dice rolls
A Goldman
msg tools
mbmb
I like Eph B because it puts coins in the game early. Eph B is particularly powerful if player can build wonder steps I and II in age I, for no more than 2 coins per step.

But I think that some of the other wonders can be powerful. Rather than view wonders as stronger or weaker, I view them as delivering a greater or lesser variety of scores. I think Babyl is extremely variable, and Halicarnassus would be too. Note that Halic is more powerful with more players.

In a 3 player game, as one of the higher ranked strategy posts noted, Olympus B and Alexandria B may be the two most powerful.
 
 Thumb up
 tip
 Hide
  • [+] Dice rolls
Henrik Johansson
Sweden
Järfälla
flag msg tools
mbmbmbmbmb
KoolKow wrote:
The problem is that the solitaire exercise is too different from the real game to make the results comparable. Your example has Giza discarding 3 times for money [including twice in the 3rd age!]; while perhaps valid for solitaire, that would be horrid in an actual game.

Given the same cards, my comparison is done using the most optimal usage of those cards by that civilization. There is no 5 VP card available for the Giza to buy! The best move, proven by exhaustive search for several hours on an 8 CPU Linux machine, is to cache in the coins. That situation is very real in any face-to-face game. Theoretically, any multi-point buy is better, but it is not universally available!
 
 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.