Recommend
4 
 Thumb up
 Hide
29 Posts
1 , 2  Next »   | 

Scythe» Forums » Strategy

Subject: Has anyone attempted to "solve" Scythe? rss

Your Tags: Add tags
Popular Tags: [View All]
Daniel Lowe
United States
Hayward
CA
flag msg tools
mbmbmb
What I mean by this is, has anybody mapped out an optimal points/turn curve for each faction/playermat combination? Obviously this isn't completely possible due to player interaction and would require a lot of playing and testing, but it'd be really interesting to see.
9 
 Thumb up
 tip
 Hide
  • [+] Dice rolls
X. Nostradunwhich
United States
New Hampshire
flag msg tools
Avatar
mbmbmbmbmb
Why would you want to do that to a game. I have never understood the drive to do that to any game.
15 
 Thumb up
 tip
 Hide
  • [+] Dice rolls
Daniel Lowe
United States
Hayward
CA
flag msg tools
mbmbmb
Trying to "solve" a game is just trying to understand a game more. Its an attempt to understand the game at the deepest level it offers.
6 
 Thumb up
 tip
 Hide
  • [+] Dice rolls
Johnson Orr
United States
Salt Lake City
Utah
flag msg tools
Pick 'em Up!
badge
You know my methods, Watson.
Avatar
mbmbmbmbmb
The problem with that level of optimization was already stated in your post, so I won't go over it. I can't even think of a way this game could be conceivably solved. There are too many moving parts, especially in games with more than two players.
 
 Thumb up
 tip
 Hide
  • [+] Dice rolls
flitz flitzer
Germany
Mainz
flag msg tools
mb
Nostradunwhich wrote:
Why would you want to do that to a game. I have never understood the drive to do that to any game.


As a scientist, I can see where the question is coming from. You try to maximize what is possible within a given rule set and environment. It also leads to understanding. This lead to a ruling preventing a combination involving the Russian faction - it's possible to get to the factory insanely fast somehow. It's somewhere here in the forum.

However, as a gamer, I also love to rely on intuition when playing games. I weigh spending time on considering "best" moves (not always entirely clear what that actually means) against the fun or the most "logical"/"consequent" next move. To me, it retains the most fun and also keeps me imersed within the theme of the game - something I enjoy deeply with Scythe in particular.

So I guess for me it comes down to: is it worth your time to crack a game or to play it?
13 
 Thumb up
 tip
 Hide
  • [+] Dice rolls
Eric Hogue
United States
O'fallon
IL
flag msg tools
Avatar
mbmbmbmbmb
philsuess wrote:
Nostradunwhich wrote:
Why would you want to do that to a game. I have never understood the drive to do that to any game.


As a scientist, I can see where the question is coming from. You try to maximize what is possible within a given rule set and environment. It also leads to understanding. This lead to a ruling preventing a combination involving the Russian faction - it's possible to get to the factory insanely fast somehow. It's somewhere here in the forum.


It's not the getting to the factory fast. The forbidden Rusviet combination allows getting 4 stars, including all mechs and full power, in 13 turns, making them favorites to end the game by turn 15.
7 
 Thumb up
 tip
 Hide
  • [+] Dice rolls
flitz flitzer
Germany
Mainz
flag msg tools
mb
Thanks for clarifying.
 
 Thumb up
 tip
 Hide
  • [+] Dice rolls
X. Nostradunwhich
United States
New Hampshire
flag msg tools
Avatar
mbmbmbmbmb
randplaty wrote:
Trying to "solve" a game is just trying to understand a game more. Its an attempt to understand the game at the deepest level it offers.

I get that on one level, but then you get the level where you make a move because you don't focus that much on the math and someone else gets angry because you did not make the "obvious optimal move". Or you get to the point where the Halifax Gambit is found and the game gets labeled "broken".

That was more what I meant, but your point is well made.

1 
 Thumb up
 tip
 Hide
  • [+] Dice rolls
Tor Iver Wilhelmsen
Norway
Oslo
Unspecified
flag msg tools
badge
Avatar
mbmbmbmbmb
Nostradunwhich wrote:
I get that on one level, but then you get the level where you make a move because you don't focus that much on the math and someone else gets angry because you did not make the "obvious optimal move". Or you get to the point where the Halifax Gambit is found and the game gets labeled "broken".

So it is more about finding in what manners a game is broken? Like if it has one particular optimal strategy making choices immaterial? One hopes such things have been caught in playtesting.

However, since a given playtesting group probably is at an even level when it comes to experience with the game, things like "sitting on the wrong/right side of the new player" problem in Puerto Rico would not be caught there, only once it is out in the wild. Then the problem becomes the players who stick to playing the broken game, and get annoyed at players who haven't learned the "boss pattern" style of play...
1 
 Thumb up
 tip
 Hide
  • [+] Dice rolls
Désirée Greverud
Sweden
Stockholm
flag msg tools
Avatar
mbmbmbmbmb
Nostradunwhich wrote:

I get that on one level, but then you get the level where you make a move because you don't focus that much on the math and someone else gets angry because you did not make the "obvious optimal move".
I have never understood this. If someone gets mad because my move doesn't fit into their plans, good. Why should I play the way my opponent expects me to? If their strategy is based on my "optimal" play (a thing I don't believe truly exists) then my job is to disrupt their strategy. I laugh at people who get mad when other people play "chaotically" and it ruins what they are trying to do. Too f'in' bad. Find a new strategy then.
2 
 Thumb up
 tip
 Hide
  • [+] Dice rolls
Steve G
United States
West Grove
Pennsylvania
flag msg tools
mbmbmbmbmb
randplaty wrote:
What I mean by this is, has anybody mapped out an optimal points/turn curve for each faction/playermat combination? Obviously this isn't completely possible due to player interaction and would require a lot of playing and testing, but it'd be really interesting to see.


I've thought about this a lot and think its a fair question. A large part of Scythe is about being as efficient as possible with your actions. Do you have to wait until you're in the moment of your turn to decide what's your most efficient play? You probably start thinking about it on your opponenet's turn. But it's often best to plan a couple turns ahead and string your actions together for better efficiency. So where should you draw the line? After every game of Scythe I find myself thinking about the game, and thinking about what I want to do differently to improve the next one. So I've got no problem with thinking about my optimal opening long before the game begins.

But as so many others have pointed out it depends on your faction/player mat combo, and your opponents (though not so much in you first several turns!). It also depends a lot on your personal preferences on what actions you value most.

I love encounter cards. My first couple turns are usually planned around getting my character to that first encounter. After all its only two moves away. But I hate moving without getting to do the bottom row action (or combat!). Obviously you can't do the bottom row action every turn, but with trade or produce I feel like I'm getting something even without the bottom row. Simply moving feels like I didn't accomplish much so my first couple turns are usually also planned around getting the resources I need to do the bottom row actions each time I take the two moves to get my character to the encounter as quickly as possible.

But someone else who doesn't value encounter cards as much as I do might see all of that as suboptimal. A similar case can be made for objectives, or the building placement bonus or which bottom row action to prioritize completing first for a star. I think any "solution" to Scythe would have to be a personal one to fit what you value most.
1 
 Thumb up
 tip
 Hide
  • [+] Dice rolls
Steve
United States
California
flag msg tools
mbmbmbmbmb
randplaty wrote:
What I mean by this is, has anybody mapped out an optimal points/turn curve for each faction/playermat combination? Obviously this isn't completely possible due to player interaction and would require a lot of playing and testing, but it'd be really interesting to see.

Player interaction? In Scythe?
2 
 Thumb up
 tip
 Hide
  • [+] Dice rolls
MyBGG Name
msg tools
mbmbmbmbmb
Nostradunwhich wrote:
Or you get to the point where the Halifax Gambit is found and the game gets labeled "broken".


...what's the Halifax Gambit? I'm coming up short in my searches.
 
 Thumb up
 tip
 Hide
  • [+] Dice rolls
Désirée Greverud
Sweden
Stockholm
flag msg tools
Avatar
mbmbmbmbmb
mybggname wrote:
Nostradunwhich wrote:
Or you get to the point where the Halifax Gambit is found and the game gets labeled "broken".


...what's the Halifax Gambit? I'm coming up short in my searches.
it's "Halifax Hammer" and it's a "broken" strategy in A Few Acres of Snow
 
 Thumb up
 tip
 Hide
  • [+] Dice rolls
Philip Morton
msg tools
Avatar
Some random thoughts because I find the idea interesting:

Solving it doesn't seem plausible, but it sounds like you're asking about something weaker than solving it. If it was solved, you could tell after handing out the mats who was going to win (if the factory cards and encounters were known information, anyway?)

If the question is "With this combination, is it better to focus on Enlist or Upgrade?", it seems more likely...

Taking opposing player actions into account would, I think, be really hard, but most games of Scythe with my group at least don't actually involve that much interaction until the last turn--it's mostly warily eyeing each other and then not actually attacking. You could maybe simulate a game with no opponents, find out what's optimal in that, and then see how vulnerable the resulting strategy looks to disruption. Maybe assume that it can attack for a win if it meets a threshold of power/cards, though this would miss strategies like "Saxony who's really good at lowballing people." You might have to do something to stop it from assuming it gets to hold the factory for free though....and you'd have to make some assumptions about Enlist trigger frequency...

Second issue I see with such a plan, with no pressure to end the game, if you ask it to optimize for points it's going to think it's optimal to draw out the game accumulating resources infinitely. You'd have to do something like have it optimize points/turn, or, more likely I think, set a target number of turns--maybe different ones assuming a long game or short game, as I'm pretty sure the optimal strategies for various mat combinations change depending on whether anyone else is rushing.

Third issue, the cards--encounters and factory cards--aren't consistent, so you're either monte-carlo-ing it or assuming some average resources/options? Maybe there's other standard ways to address things like cards here, I don't know.

Not sure what else I'm not thinking of, but it seems like maybe a simulation like that would then be able to tell you how many points you could expect to get with a specific pair of mats going Upgrades+Mechs or split Enlist/Build/Mech or whatever, assuming that you don't get crowded out of whatever hexes the simulated strategy thinks it needs.
2 
 Thumb up
 tip
 Hide
  • [+] Dice rolls
Paul Ferguson
Australia
Brisbane
Queensland
flag msg tools
badge
Avatar
mbmbmbmbmb
I play to have fun, not to analyse. People that do miss the point of playing games. It should be about the interactions not the best way to play said game.
 
 Thumb up
 tip
 Hide
  • [+] Dice rolls
Danny Perello
Canada
Salmo
British Columbia
flag msg tools
badge
Want DnDeeples? I'm on Esty as MagnificentMeeples.
Avatar
mbmbmbmbmb
itmo wrote:
I play to have fun, not to analyse. People that do miss the point of playing games. It should be about the interactions not the best way to play said game.

Your language is a little harsh. Shouldn't people play games in whatever way brings them enjoyment? It isn't like there is only one way here. Some play for the interaction, some play for the puzzle, some to fight boredom and on it goes. There is no one way to play...
5 
 Thumb up
 tip
 Hide
  • [+] Dice rolls
Philip Morton
msg tools
Avatar
itmo wrote:
I play to have fun, not to analyse.

Analyzing is fun!
11 
 Thumb up
 tip
 Hide
  • [+] Dice rolls
BG.EXE
United States
flag msg tools
I favor Euros, but I like pretty much all games.
badge
Make the kinds of posts you want to read.
Avatar
mbmbmbmbmb
I try to "solve" all games the best I can. I believe very few board games are TRULY solvable in scientific terms (as in, few have a strong proofed solution) but playing optimally is a joy in itself. That doesn't mean I'm not fun to play with or I take 4 hour turns. But saying people who want to dig deep into strategic optimization are playing wrong is ludicrous. I mean, go play The Colonists without trying to optimize things and see how fun it is. Probably not very.
2 
 Thumb up
 tip
 Hide
  • [+] Dice rolls
Keith Santamaria
United States
Texas
flag msg tools
For me personally I just love optimizing things, it's almost as fun as the game itself. There is a thrill as you get closer and closer to the perfect line and you feel how much better you are getting.

I mean sure once the game is solved then you "beat" the game and now its over, but there 1000s of more games to go and try to do the same.

That is assuming games are simple. They are not and have seemingly infinite branches and different outcomes, so it will never be solved.

But damn is trying for it super fun.
2 
 Thumb up
 tip
 Hide
  • [+] Dice rolls
Scott Allen
United States
Freeport
IL
flag msg tools
designer
Avatar
mbmbmbmbmb
Chrondeath wrote:
If the question is "With this combination, is it better to focus on Enlist or Upgrade?", it seems more likely...


I was just thinking along these lines today:

For each base game faction and player mat combination (25 possibilities), what is the "best" / preferred approach?: Upgrade, Enlist, Build, Mechs, based on available starting resources, costs of actions on the boards, etc.

I don't think this "breaks" the game (because of all the other wonderful variation built into the game: Factory card, encounter cards, objectives, structure bonus, etc.). But, I think it would help beginners, especially those playing against the ruthless and efficient Automa, have at least a fighting chance.

(I played a solo game last night against one Automa. I lost something like 76 to 64. I played 2 more turns, got my 4th mech and my 4th recruit and the score would have been 84 to 76 (or so). So, I was on the right path, I just wasn't efficient enough to get that 6th star before the Automa did. So, I'm just trying to figure out: OK, if I am Nordic + Industrial it's best to do _______ and this ________ early on to maximize my efficiency.)

Hope this all makes sense.

Thanks.
1 
 Thumb up
 tip
 Hide
  • [+] Dice rolls
Scott from Vancouver
Canada
flag msg tools
^^^

This is a great response for people who ask if they should buy Scythe in part to play solo. For people who are into solving, playing vs Automa is awesome. You can take as much or as little time as you want to try and find a 'perfect' path.
1 
 Thumb up
 tip
 Hide
  • [+] Dice rolls
Brendan Slade
msg tools
mbmb
I think the problem with solving the game is that there is no point playing anymore which kind of defeats the purpose.

Like I was reading a choose your own adventure book and I mapped out all the options to find the optimal choice for each decision and it while it was very effective it kind of ruined the book at the same time.

Obviously you always want to be searching for better strategies but if you just sit down and step by step work out what the best strategy is then the game will likely be over as a fun past time.

That being said the player interaction makes it hard to plan out the perfect game. You can plot out perhaps the most ideal opening moves for each player/board combo though.
 
 Thumb up
 tip
 Hide
  • [+] Dice rolls
Kayla Bivens
United States
flag msg tools
mbmbmb
Honestly, I think that attempting to "solve" Scythe is an interesting idea, if only for trying to determine how complex it can get (i.e. is it more/less complex than Chess, Go, etc.).

Given that Chess is only "solved" for up to 7-piece (including both kings) board-states, and Chess is a game where pieces are removed rather than added, I don't see Scythe being solvable in the foreseeable future, either.

Granted, in Chess, this complexity only makes the early game board-state combinations significantly large, whereas in Scythe, the early game has a relatively small amount of combinations that later explode possibly even beyond Chess' board state combination counts.

This excuses that starting faction & player mats are randomized, let alone objective, factory, encounter, and combat cards, structure bonus, airship passive, airship aggressive, and resolution tiles, and then whatever all else is coming with the third and final expansion.

So, just the starting states of Scythe are significantly larger than even the first 2 turns of a chess game!
1 
 Thumb up
 tip
 Hide
  • [+] Dice rolls
Brandon Zappala
United States
Tampa
Florida
flag msg tools
mb
So funny enough, the "scythe statistics calculator" offers a "TEMPLATE" for winning the games. Though in no way does it indicate "do this to win!"

The ending power of a winner is either 9 or 16.
The ending popularity of a winner is 7-8 or 13.

The player who places the 6th star wins 75% of the time.

And the player with 6 stars likely has stars for (in order of win rate):
1) objective 1
2) recruits
3) mechs
4) a single combat (but not 2 combat stars)
5) workers
6) reaching 16 power

One can attempt to follow this template to get a win, but this is the statistics of "results" and don't necessarily indicate a formula for success! Especially because sessions of Scythe are so situational and interactive. This is what makes Scythe such an awesome game!
3 
 Thumb up
 tip
 Hide
  • [+] Dice rolls
1 , 2  Next »   | 
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.