-
Hey, guest user. Hope you're enjoying NeoGAF! Have you considered registering for an account? Come join us and add your take to the daily discourse.
Opus Magnum - Zachtronics latest. Coming to Steam 10/19
- Thread starter BrokenFiction
- Start date
It's reasonably clear that it means to continue that piece of dialogue though :V
Meanwhile next in the series of "throw more money at it to make it faster": Voltaic Coil
Meanwhile next in the series of "throw more money at it to make it faster": Voltaic Coil
I speel my dreenk
Member
Nice, the Airship Fuel and Hangover Cure solutions particularly are satisfying to watch. I look forward to making the most elaborate, convoluted garbage later on just to watch it run.
That happens to me like 2 out of every 3 games >_>
My current album of solutions https://imgur.com/a/vaatV
They're mostly fast solutions and I try to keep cost and area overhead to a minimum too but I don't try tooo hard. Basically if it looks pretty and efficient I consider it done for now. Will optimize later.
Also will definitely take more leaderboard challengers. Quote for a link to profile.
I think my first-pass approach is mostly due to completion time having a definite algorithmic minimum (reagents can only generate so fast) whereas area and cost would take far longer to approach an ideal state.
Definitely going to go back and do them that way after optimizing completion times.
Definitely going to go back and do them that way after optimizing completion times.
First solution i'm proud of! This one's the Hair Product.
https://i.imgur.com/a2tycAy.gif
Cheap, fast and small!
https://i.imgur.com/a2tycAy.gif
Cheap, fast and small!
The only learning-code game I have played was Human Ressource Machine
and I got stuck into a certain part where my brain couldn't continue, like part 3 or something
Will I be able to enjoy other Zacktronic games like TIS-100 or SHENZHEN ?
I found it quite similar to those games. TIS is fairly simple to get into and as long as you dont beat yourself up trying to get the targets for each level you can kind of proceed at our own pace
Okay, I just beat the main campaign. "As one of the first players to complete the main campaign of Opus Magnum" I got a URL sending me to an actual store page to buy an actual physical trophy which only ships within the USA.
Nice gesture, thanks, but uh. Yeah.
Overall, I'd have to rate this below both Spacechem and Infinifactory. It's still a good game and the levels/mechanics are obviously fresh while I've played those other two to death, but both a rather relatively small amount of content and some rather one-size-fits-all methods of doing certain things (probably again due to the lack of content) makes me sit here feeling rather unfulfilled having finished the campaign.
I'm rather sure from prior experience with community-sourced content that the Journals are just going to be pages and pages of gimmick levels, but I'd love to be proven wrong.
Story was neat, albeit also rather short and somewhat predictable and not really going anywhere, with a script that didn't really jump out. Characters feel like nothing more than a vehicle to move you from level to level, which is surprising after how much personality Shenzhen's cast had. It honestly feels like I've finished Arc 1 and two more arcs are missing, or something.
Don't get me wrong, I'm going to be getting my fill and then some but overall, it's hard to recommend at full price over Infinifactory instead for someone looking to jump into Zachtronics.
(Yeah, I'm aware it's Early Access)
Nice gesture, thanks, but uh. Yeah.
Overall, I'd have to rate this below both Spacechem and Infinifactory. It's still a good game and the levels/mechanics are obviously fresh while I've played those other two to death, but both a rather relatively small amount of content and some rather one-size-fits-all methods of doing certain things (probably again due to the lack of content) makes me sit here feeling rather unfulfilled having finished the campaign.
I'm rather sure from prior experience with community-sourced content that the Journals are just going to be pages and pages of gimmick levels, but I'd love to be proven wrong.
Story was neat, albeit also rather short and somewhat predictable and not really going anywhere, with a script that didn't really jump out. Characters feel like nothing more than a vehicle to move you from level to level, which is surprising after how much personality Shenzhen's cast had. It honestly feels like I've finished Arc 1 and two more arcs are missing, or something.
Don't get me wrong, I'm going to be getting my fill and then some but overall, it's hard to recommend at full price over Infinifactory instead for someone looking to jump into Zachtronics.
(Yeah, I'm aware it's Early Access)
UcchedavÄda
Member
How long did it take you to beat the main campaign?
I kinda feel the same way TBH, even though I haven't beaten this or Shenzhen. Managing to figure out a puzzle in this doesn't feel as satisfying as it does in Shenzhen or Spacechem, but I've also had a lot more fun going back to puzzles to optimize them.
Even though the checkout procedure sucks, Zach will send patches to international players. There was an option for International, and even though you have to enter something USA-like in the address form, you can put your actual address in the note box.
cpp_is_king
Member
Just picked this up. In a way I'm glad this isn't another another insane programming one. For the Shenzen I/O I ended up writing an actual compiler backend for LLVM so I could just run everything through its optimizer.
I want to keep my sanity this time.
I want to keep my sanity this time.
So here's some hardcore theorycrafting.
Completion time is essentially the time it takes for creating one product, multiplied by the number of products (generally 6), and then with an overhead of how long it takes for the last reagent in a product to get from generation to when it's dropped off in the product area.
In other words, if
Tp is optimizable simply based on how fast Reagents can generate output, which is hard-capped at two cycles per reagent. (one cycle for an arm to grab, another cycle for that arm to move away)
Tl is optimizable based on these factors:
- moving a reagent to any function (bond, unbond, transmutation, product goal) is one cycle
- changing arms delays by one cycle sinceyou cannot release with one and attach with another in the same cycle, it throws an error you cannot grab an element on the same turn it moves into a tile or is generated. edit: Corrected this statement.
Therefore to optimize Tl, the absolute last reagent in a product should always have each cycle after it's generated contributing something to the assembly of the product with no wasted moves. This is most simply done by having one arm carry that last reagent from start to finish, but other solutions exist.
To demonstrate, here's Hangover Cure
In Hangover Cure, since each product requies the reagent to output three times, Tp = 2*3 = 6. The last reagant in each product takes 3 cycles (calcification, bonding, goal), or Tl = 3
Cycle completion time hence is (6 * 6) + 3, or 39.
It should theoretically be possible with a more elaborate setup to have the last reagant be the middle Elemental Water, reducing Tl to 2 (bonding, goal) and an optimized time of 38.
And here's Airship Fuel
Since the Product requires 4 elements and we have 3 generators, the calculation gets a little more complex. The solution relies on grouping 3 products together (12 elements needed) and since 3 reagents take 8 cycles to generate that, we end up with Tp = 8 and N = 2
Tl is 3 cycles, similar to the solution for Hangover Cure. HOWEVER, due to a collision in sending finished Products to the goal, the last one has to be delayed for 1 cycle, hence the final cycle count is (8 * 2) + 3 + 1, or 20.
It should theoretically be possible to create a far more complicated setup where the last three reagents generated complete one product instead of two, eliminating that collision, but I don't think I really even want to begin to fathom how that'd work.
edit: I'm noticing a bug in the GIF generation in Airship Fuel, heh.
edit2:
edit3: Found this 19 cycle Airship Fuel solution on Reddit
insert another pokerface here
Completion time is essentially the time it takes for creating one product, multiplied by the number of products (generally 6), and then with an overhead of how long it takes for the last reagent in a product to get from generation to when it's dropped off in the product area.
In other words, if
Code:
time to make one product = Tp
# products = N
cycles for last product to complete = Tl
completion time = (Tp * N) + TlTp is optimizable simply based on how fast Reagents can generate output, which is hard-capped at two cycles per reagent. (one cycle for an arm to grab, another cycle for that arm to move away)
Tl is optimizable based on these factors:
- moving a reagent to any function (bond, unbond, transmutation, product goal) is one cycle
- changing arms delays by one cycle since
Therefore to optimize Tl, the absolute last reagent in a product should always have each cycle after it's generated contributing something to the assembly of the product with no wasted moves. This is most simply done by having one arm carry that last reagent from start to finish, but other solutions exist.
To demonstrate, here's Hangover Cure
In Hangover Cure, since each product requies the reagent to output three times, Tp = 2*3 = 6. The last reagant in each product takes 3 cycles (calcification, bonding, goal), or Tl = 3
Cycle completion time hence is (6 * 6) + 3, or 39.
It should theoretically be possible with a more elaborate setup to have the last reagant be the middle Elemental Water, reducing Tl to 2 (bonding, goal) and an optimized time of 38.
And here's Airship Fuel
Since the Product requires 4 elements and we have 3 generators, the calculation gets a little more complex. The solution relies on grouping 3 products together (12 elements needed) and since 3 reagents take 8 cycles to generate that, we end up with Tp = 8 and N = 2
Tl is 3 cycles, similar to the solution for Hangover Cure. HOWEVER, due to a collision in sending finished Products to the goal, the last one has to be delayed for 1 cycle, hence the final cycle count is (8 * 2) + 3 + 1, or 20.
It should theoretically be possible to create a far more complicated setup where the last three reagents generated complete one product instead of two, eliminating that collision, but I don't think I really even want to begin to fathom how that'd work.
edit: I'm noticing a bug in the GIF generation in Airship Fuel, heh.
edit2:
edit3: Found this 19 cycle Airship Fuel solution on Reddit
insert another pokerface here
truestatic
Member
You inspired me. My 38 hangover cure.
It was actually extremely simple. I don't want to know what optimization would look like for actually hard puzzles though.
Messofanego
Banned
That's pretty cool of them.
undu
Member
I didn't think I was doing anything special
I'm having a hard time with later problems, actually