I think the dev's goal has been towards balancing the difficulty not so much towards making difficult puzzles that are rewarding to beat, but towards making reasonably challenging puzzles that become rewarding to revisit and optimize as you get better at the game. In a way it sort of mirrors when you've just started learning programming, when every task takes large amounts of thinking "how am I even going to solve this?", implementing a solution that is sure to not be very optimal or even well written, vs. being an experience programming, when "how do I solve this?" is only the first step towards "how do I make my solution _better_?".

I hope we keep getting more Zachtronics games for a while; there really aren't too many games out there that are comparable to them.

> I think the dev's goal has been towards balancing the difficulty not so much towards making difficult puzzles that are rewarding to beat, but towards making reasonably challenging puzzles that become rewarding to revisit and optimize as you get better at the game

That's been my impression as well. TIS-100 is still has a high replay value, because optimize optimize optimize!

Yes, TIS-100 is super fun to optimize!

Shameless plug: I think I made the Signal Averager about as fast as it can possibly be https://www.youtube.com/watch?v=XBfEcxvJ6aY

Factorio is another one for optimization. The main difference is that Factorio is a single puzzle that you optimize continuously throughout the game.

Not a single game I know drives home the concept of "technical debt" as much as Factorio does. Every time I want to automate manufacturing of a new thing, I end up wondering whether I should do a quick&dirty hack on existing pipelines, refactor half of my base, or just build a new one elsewhere...

(And you can't just always keep choosing "build a new one elsewhere", because this increases a literal attack surface and invites literal bugs.)

Never though about it that way, it really seems that Factorio is a very good representation of actual software development without actually containing coding (except for the logic you can do through the cables if I remember).

I've found all the games to have high replay value for the same reason. It doesn't really matter how hard the game is because there are plenty of challenges beyond just passing the level.

What does "stupidly difficult" even mean? Does it just mean you did not manage to solve all the puzzles (yet)?

Honestly, I find that Spacechem is one of the very best game ever. I am really happy that Zachtrorics continues trying making better similar games, as it generally does make great games (I immediately purchase all their games - I only missed Opus Magnum as I was busy becoming a father when it launched), but so far, I was not able to feel the joy I felt when my Spacechem reactor finally worked, watching all the chucks fitting together (I know, I really need to try Opus Magnum which is more similar to Spacechem than other games in the serie - but my impression is that the UI looks much more busy, and simplicity is one of Spacechem strong point, which makes it enjoyable on a tablet for example). I'll admit that Spacechem had the advantage to be the first of this kind I tried, so it could be one of the reasons I loved it so much.

Also I am a bit sad that making games easier to beat is even a goal at all - it's great to try to make them easier to play / to start for non-programmers and beginners. But I find that having a few levels that are really challenging makes a game strictly better. I know that many people think "I paid for the whole game, I need to be able to finish it all reasonably quickly", but personally, I think it could just mean that the game was simply not challenging enough.

For what it is worth, I did beat Spacechem (whatever it means, I have the "beat the game" achievement on Steam), but I am stuck on TIS-100 and Shenzhen I/O (I spent much less time on this one though). Exapunks does feel easier than both for now (the programming language is slightly "higher level", so maybe it feels more familiar).

"stupidly difficult" to me means Nethack. You should try it if you're into rpgs :)

I like Nethack :) Btw, that reminds me of Dwarf Fortress and the whole "losing is fun" idea, of which I am a great fan as well, and do not find anything stupid about it ;)

I'm quoting from the linked Gamasutra post, which shows the attrition rate per level.

Sorry, my bad, I did not read this article. Zach himself is calling the game "stupidly difficult". I still strongly disagree with the term though :) Plus seeing the graph, I find that it does not seem to be the case - I mean most levels in the main campaign seem to be over 80% success rate (hypothesis: "defense" levels are not available in the Android version, which could be a reason why they seem to have a lower success rate)

80% completion rate with 20 levels works out as a ~1% chance of ever finishing the game (which i'd say counts as fairly hard). Looking at the graph, the pass rate seems to be higher, but there are also a lot more than 20 (non-optional) levels, so I would guesstimate the finish rate to be less than 1% (90% per level for 40 levels is also around 1% to finish the game, for example).

https://steamcommunity.com/stats/SpaceChem/achievements according to steam achievement "beat the game", it's actually 4%. I do agree that it makes it a difficult game to finish. Then again, my point is that I think it's fine not to be able to finish a game, especially a puzzle game which have no story. You just have to spend more time on the puzzle you are stuck on, so in a sense, you get more game time for your money :)

SpaceChem does have a story though. I gave up on the game after it got too difficult, and at that time I was disappointed that I wouldn't find out what happens in the rest of the story.

Opus Magnum had some puzzles that were rather tedious, I think maybe more than average. There's always a point in these harder puzzles where you know what you need to do and then slogging through the execution is daunting on its own.

Interestingly, I liked the Spacechem the most (probably because the program is represented in the most visual form of all their games, and it has excellent music and it's very calming to watch a resulting solution), although I didn't finish the last optional puzzle and any of the expansions.

I would love to go back and finish those, but what holds me back is that I would have to replay the whole game. All the tricks I learned about how to play it I have already forgotten.

+1 for SpaceChem being stupid-hard. In spite of that I thoroughly enjoyed it.

SpaceChem wasn’t their first, they had a few prior with long names, the only one of which I remember is Codex of Alchemical Engineering. There was a steam mecha one, and an IC design one.

KOHCTPYKTOP, Engineer Of The People was the "IC design one". It's an abstraction of NMOS semiconductor design, making it a difficult game to approach. Fun, though!

http://www.zachtronics.com/kohctpyktop-engineer-of-the-peopl...

Is that an enjoyment or difficulty ranking?