To Sea in a Sieve - how could I improve it?

I’m going to follow Arthur’s lead and ask for suggestions as to how To Sea in a Sieve could be improved for the post-comp release. What would have pushed you to give it a higher mark?

One of the main criticisms of the game from reviewers was that it was too hard. This surprised me because one of my worries when I was writing it was that it was a bit too easy! Which puzzles did you struggle with, and could be better clued?

I’m aware that I made a big mistake with the hint system - there were lots of hints coded for when the player has been seized by the Yateveo tree, but I forgot to add “hint about [thing]” as one of the allowed commands when the player is ensnared, so instead you got the unhelpful message “I cannot, for I entangled entangled in spiny tendrils!” I am very sorry about that, I don’t know how I missed it - but that one has already been fixed in the post-comp release.

I have also made “light” a synonym for “sun”, another one I should have thought of.

It has been a bit of a rough year for me. My girlfriend, who I hoped would be my life partner, broke up with me in January. We are still very good friends and I still love her deeply. Over the summer the interest rates crisis happened and my freelance work, already a bit sporadic, suddenly dropped off a cliff. The cost of living was increasing and I had no money coming in at all. Sometimes writing the game was the only thing that would motivate me to drag myself out of bed, but I my mind wasn’t always on it. Numerous times I found myself staring blankly at the screen with no idea of what I was doing. I’m still a bit all over the place, but work has picked up a bit and I’m better than before. It would be nice to polish up To Sea in a Sieve and move onto the next one!


I haven’t completed it yet but so far it seems challenging and fair so to me it wouldn’t be an Improvement if it became a lot easier. But perhaps if you want to make it easier you could have a difficulty setting chosen by the player?


I very much liked the game and am glad you got a high ranking because you deserved it!

One thing was “You might understand more if you examine [object]” in the hints. I wanted to be able to survive but examining it would cause me to run out the needed time and die, which is annoying.

Good luck with finding income! (Have you got a Patreon? People could donate small amounts there!)


Yes, that’s a technical issue. The hint system is based on the one used in Emily Short’s early Inform 7 game Bronze. Asking for hints should really be an action-out-of-world which would not advance the timer, but some of the hint rules, out of necessity, are instead rules, which for some reason can’t be out-of-world actions. I might be able to make all of the hint rules check / carry out / report, which is something I’ll look into. Good call, thank you!


In Counterfeit Monkey, examining, looking and taking inventory are out of world actions which won’t advance time, and they can still be used in instead rules. I don’t know how easy this system would be to transplant to your game. The basic mechanism is this:

Examining something is acting fast. Looking is acting fast. Taking inventory is acting fast.

The take visual actions out of world rule is listed before the every turn stage rule in the turn sequence rules.

This is the take visual actions out of world rule:
	if acting fast:
		rule succeeds.

There are some complications in the Counterfeit Monkey code that might be relevant to this, I’m not sure. It implements an extra “sanity-check” stage before the “before” stage, and has rules such as Sanity-check doing something other than acting fast during roc-soaring.


Worth a try! Thank you! :slightly_smiling_face:

Update: it works! :smiley:


honestly it was perfect difficulty for me, and I suck at puzzles. while there were bugs (which I think are apparent in my transcript/noted in my review) and things that could be hinted/signalled better, I wouldn’t change most of the puzzles or structure overall. I felt so damn smart with figuring out what to do with the barrel on my own.

I did have trouble with the in-game (non-hint) signalling for how to get rid of the yateveo tree, but I think that was my main “I have no idea what the game wants me to do here” issue.


Hi @pieartsy! Thank you for your transcript, it has been helpful.

Spoilers below:

I think the issue with the tree is that it’s the one and only time you need to ask the Captain something to gain the information you need. What I’ll probably do is have him volunteer the information at some point, perhaps the first time you escape from its clutches.

1 Like

ah yeah that is the crux of the problem, eh? the adjustment would help!


For what it’s worth, I feel like the modern wave of IF is maybe a little “too” afraid to make a challenging puzzle, and I wouldn’t dilute your difficulty at all. I’m probably a minority, of course, but just so you know those voices are out there. I didn’t use the hints in your game (as mentioned in my review/transcripts) EXCEPT when I got thrown off track by similar verb phrasings whose responses turned me away from the real solution. And I believe you already addressed those.

It was my first year judging, and after it was over, I determined that I should probably use a different rubric next time. I had a big traffic jam between 6 and 8 where I clearly saw some games as better than others with the same score (I predicted before the comp started that I was going to find myself wishing for decimal scores.) At any rate, I didn’t go higher than an 8 this year, and Sieve got one of them.
But I will say, which isn’t the game’s fault, that I don’t think I can end up as completely immersed and overwhelmed with satisfaction at the end of a one-room game as that of a larger adventure. Another subjective thing, I guess. But Cap’n Booby went a long way to make it contend!


Double ouch on the life “progress.” Those are, uh, really big. I’m glad TSiaS allowed some latitude to work through both pretty big things.

This is going to be a long one–even though you fixed so much during TSiaS testing and the original draft would’ve done well, I think. And I enjoyed working with you as you looked to make TSiaS even better. But I know first-hand all too well how it can be to not quite have the time to put in all the features or testing you want to!

There’s always more one can do…but it has diminishing returns to scale! Eventually you will find things more than good enough and move on, so we can enjoy your next work.

So I thought I’d add something that’d add the most bang for the buck (or pound or euro) and also spill some general thoughts, and I legitimately think other parser authors might be able to use the general concepts.

Hinting is tricky–and I still make mistakes with it, but I can fob off a lot of work on the game, I’ve found. And it would’ve been veering way too much to fix the hinting instead of the actual game.

I’ve found that just having hinting relatively stable was a relief for me–one less potential variable that could go terribly wrong, and I’d be okay with one or two not working perfectly, and it’s something I could check when I wasn’t feeling original.

First, though, I wanted to combine two posts above me. Again, this may be an interesting feature, or it may be frippery. But it might not take long to implement if you decided to!

Let’s put these two things together: give the player the option to decide whether or not examining takes a turn.

I don’t think this would be too bad. In your “every turn” rules you could have:

every turn (this is the new boat sinking rule):
    if examined-something is true:
        now examined-something is false;
        continue the action;
    now examined-something is false;
    (check for the boat sinking)

You’d just need to set examined-something to false after bailing as well, or you would get an extra turn.

This next bit is nontrivial and more about process. And not in the “you must have up-to-date development processes” mold but more the, hey, this is probably simpler than you think, and you didn’t know how to do this, and I bet it’d be neat and fun. Because programmer-testing gets tedious (well, for me–certainly I should’ve done more for BBKK!)

And I think this is a general thing parser game writers can and should all do, or something like it, if we include hints.

I think I might’ve discussed this as a pie-in-the-sky thing you probably had no time for while you were still building the story and fixing implementation. But post-comp, I think it’s worth a try, because if it works well, it (or just the general process) might be something to put in place going forward for Mars in a Minibus. (I hope you don’t wait 20 years to write that!)

This would be testing commands to check for hints. They caused reviewers some frustration, but once you know what to do, it’s a source of relief. (Well, it is for me, knowing that at least the hints work reasonably!)

I’d make a separate module. (Perhaps this is an extension I could sent to Friends of I7. But it is not hard to write on your own and tweak as necessary.)

volume testing commands - not for release

include hint checking every turn by JJ Guest

Then in the module…

every turn when hint-check is true: try generalhinting;

every turn when hint-check-each-object is true:
    repeat with X running through touchable things:
        say "Hints for [X]:[line break]";
        try object-hinting X;

[note: we could tweak this further by defining something as worth-hinting, so we have less text bloat. E.g once it's served its use, it could be flagged as not worth-hinting.]


You could even take players’ transcripts and run things. For instance, cutting and pasting everything after > with a GREP, then removing the >, and then placing HINTCHECK ON and HINTCHECK OFF as commands at the start and end. You could even have

test walkthrough with "(walkthrough that works despite randomness)".
test hintthru1 with "hintcheck on/test walkthrough/hintcheck off".
test hintthru2 with "hintall on/test walkthrough/hintall off".
test hintthru3 with "hintall on/hintcheck on/test walkthrough/hintall off/hintcheck off".

This may seem like a lot but it may take less time to implement than you think, and then you’ll have something you can look at and regression-test if you need. (Well, ideally.) You may have to fix random events if your walkthrough goes through that.


I am one of the people who found the game too hard! I really enjoyed the writing and interactions, but I leaned heavily on the hints for almost the whole game, so I didn’t get to have those moments of feeling smart. It started when I couldn’t figure out what to do for the very first action, so that set the expectation early that I would need help. Besides that first action, I also wouldn’t have expected the captain’s clothing and pockets were important, since descriptions of people are usually for flavor, plus there were so many other interesting things to look at. So I think for me, the game seemed hard because I had so much trouble at the beginning.

There’s nothing wrong with a difficult game, though! I think I was surprised at the difficulty because of the otherwise lighthearted vibe, but it’s fine if the game is meant for people who are more patient than me.


Sorry @SomeOne2 but I think I missed your meaning in that you felt examining objects should not count against the timer. I’m not sure about this myself, since I interpret examining as “examining closely” and being synonymous with searching. I do think that LOOKing perhaps ought not to carry a time penalty however. @aschultz’s idea of making the time penalty on examine optional is definitely worth considering though.

Thanks for your kind words! I’m pleased with how the game placed. The last few weeks before the deadline were a struggle and I’m still waiting for the fog to clear. I did make a few silly errors but I shouldn’t be too hard on myself.

Milliways is an incredible achievement by the way! I admire your ambition and I very much enjoyed the result. :smiley:


I kinda liked that everything took time: made the time pressure to bail feel more tense, at least at the beginning before you threw some things out.

I mentioned the trouble I had in my review: that I was surprised I could just beat up the captain (unless he had a weapon, and even then it just was trickier), or struggle free of the inescapable carnivorous tree, but once I got over that hurdle I was mostly OK. I do think that experience sorta put me in the same boat as Sarah where I was predisposed to lean on the hints more than I might have otherwise though…

(Edit: ugh. gah. pun not intended. how did I do that accidentally?)

So… hmm. For me it might have helped if the intro and some of the early descriptions had emphasized how slippery or tough or resourceful you are rather than making you sound like a scrawny failed thief… dunno.


I am glad to know there were meant to be hints for this part!


I had trouble with the “struggle” command because I was trying all sorts of actions that I would consider equivalent to struggling, but the game said I couldn’t do any of them. So I was completely stymied as to how I was supposed to escape if I couldn’t move at all. Sicne “struggle” is a nonstandard verb, it would never have occurred to me to try it even with the strong hint from the captain.

So sorry to hear about the year you’ve had. Despite my difficulties with it, I quite enjoyed the game and I now plan to play through your back catalogue!


I get this, and I think I do need to make this clearer. I get the sense that the people who had less of a problem with this were those who had previously played To Hell in a Hamper, which included some of my testers. The Captain’s descendent Hubert is obviously a pushover, whereas the Captain comes across fiercer.

Regarding the pockets:

At the start of To Hell in a Hamper there is literally nothing else you can do but look in Hubert’s pockets, since he and his overcoat are the only things left in the balloon. Obviously not everyone has played Hamper, but I hoped that in Sieve making the player a pickpocket would be enough of a clue that you were supposed to do the same.


I’m sorry to hear about the bad year you had in your personal life. It’s even more impressive that you finished the game despite all of that. I hope there’s some light on the horizon.

I did find the puzzles too hard for me. (That’s just one data point, of course, I suspect every puzzle will strike some players as too easy and others as too hard.) In my review I gave some specifics about my frustration points and what kind of clues might have helped me.

After a bit more ruminating I did have two higher level thoughts about how difficulty is perceived by the player (OK, perceieved by me, but I suspect it generalizes), that I think maybe relate to why this game felt notably frustrating among the several games where I spent a decent amount of time in the walkthrough.

(1) Sequencing / flow of game: it feels worse to get stuck at the very beginning of the game. I really like the way @lpsmith put the relationship between getting stuck and the player’s trust in the author:

So I think that makes the very opening parts of the game especially important, because that imaginary bank account where I credit and debit trust with the author is likely at a low initial balance.

When the game starts with something I can accomplish (maybe solve a puzzle, but also just explore an area, talk to an NPC, or learn backstory from clues in the environment) that fills that imaginary bank account a bit, and if a tricky puzzle shows up later I’m going to try harder at without giving up.

Applying that here, TSiaS sticks out a me a bit as maybe the only game in the comp where I felt like I didn’t accomplish anything on my own without the walkthrough / hints. And that is a bad feeling.

So I think the feeling of difficulty could be changed just by adjusting the opening parts of the game to give the player something even easier to do, even if all of the puzzles were kept the same after that. (Iirc, you had to open by putting water on the gun, which doesn’t seem that hard, but you’re in a boat full of possible things to mess with, and as I recall the text that suggests focusing on disarming the captain comes AFTER you try to mess with the captain’s pockets, which I was, errr, pretty far from thinking of doing.)

(2) Size of game / ability to work on multiple puzzles in parallel: it feels worse to get stuck on task A when task A is the only task you can possibly work on. My brain processes things in the background, so often in a slightly bigger game I’ll find myself trying to, say, decode the knots in a macrame wall hanging in the salon when I suddenly realize, “aha! I should see what happens if I use the shoe polish on the projector in the observatory!" So that (fictitious) game made both puzzles feel easier by giving me the chance to work on them both at the same time.

In TSiaS, it felt like there was one correct path through the puzzles because they often gave an item I would need for the next one. That’s a perfectly fine structure of course but I think it does make it more frustrating when a player hits a bottleneck and can’t do anything else.


I meant use @JoshGrams’s code (sort of) and if you get a hint for something where the hint is “Ye might understand more if ye looked closely at [the something]”, then if the immediate next turn is EXAMINE [said something], then it’s a free turn. Which you could do with in-quote brackets that set free-examining to the object.

I understand what you mean. But sometimes I’ve examined an object (or otherwise fiddled with it), and am stuck so I go to the hints, and I’m told to examine it and then the next hints can’t even help me. So I wasted a turn. In the end I went to saving my file, asking fur a hint, examining, asking for a newer hint, and restoring.

But, you know, players can also use that as a cheat method. I don’t think it’s necessary, just maybe a little help.

The actual game itself I think was just hard enough, and satisfying to get through. I don’t like too easy games. Just maybe a bit on the hints would help.

Thanks! :heart: My next game is going to be smaller and less ambitious - like yours, maybe. Except horror …

I already failed the “ambitious” part :rofl:

I can’t wait for your next game. 2026, maybe?


Why is it that so many parser games got within the 70s range this year? The close calls were way too many!

1 Like

This is a hack, but you can also do:

First carry out asking for hints: abide by the instead rules.