Oh, hey, I have a bunch of loud opinions about this! (All personal and subjective, mind you.)
Ultimately, I want a game that will delight me while not getting in my way. These things balance each other out, to some extent (I will cut more slack to a delightful game that is bugging me flow-wise, and vice versa) but I weigh the delight factor a bit higher, and I don’t think I’ve given any nines or tens out purely for technical competence.
How to quantify the delight factor? It’s tricky, but I think it consists of providing a cool and novel experience for the player while connecting with them on a human level. Which, man, that sounds hard. Fortunately, you can fudge it a bit by throwing in something everyone likes, like an adorable puppy or a winsomely quirky love interest, or something ridiculously awesome, like shoulder-mounted shark cannons. (Are they shaped like sharks, or do sharks come out of them, you ask? Both.)
(I’m being a little snarky about the puppy and the love interest, but it does seem that on the whole, comp judges prefer games with appealing characters. Violet without Violet would be a perfectly fine game, but probably not the beloved competition darling that it was, and possibly not the first game I think of when people ask me for IF recommendations. Also it would have to be called something else.)
If you have a delight source in your game, don’t be afraid to amp it up and get a little fanservicey with it. If you start getting transcripts back and find that your testers want to hug their mom, or give things to the cat, or what have you, make sure those actions are implemented and satisfying. (Paying attention to what your testers want to do is a good idea overall.)
Depth of implementation falls partly under delight, and partly under flow. I have sort of a baseline expectation for how much work has been put in describing things and creating systems, and so on. If you exceed this, I am impressed, and you get points. At minimum, anything important enough to be on its own line must have a description. Points docked if the description is along the lines of “It’s a chair, what did you expect?” Any major objects mentioned in the room description should also be present and described. (We can quibble over the definition of “major,” or, alternately, you can implement every single one and impress me!)
I am also impressed by systems and interactions that go above the call of basic functionality, especially if they tie into flow and make my life easier. (Easier transportation back to places, say, or a self-organizing inventory, things like that.)
Rule of thumb: if I am surprised and happy, your score is higher.
So, let’s talk about flow, and how its disruption can kill your comp score. Think like a reviewer here for a minute. It’s November third, you have, ugh, twenty more of these games to play, your kid spent all morning demanding to watch a DVD you’re pretty sure has never existed anywhere, especially in your house, your gut is still messed up from your sister-in-law’s pie, and all you really want to do is hide in the bathtub and play the new Phoenix Wright game. You want to be fair. You feel you owe it to the authors to be fair, and give their game a chance. But, realistically, every roadblock you hit is a new opportunity to ragequit. (The save throw against ragequitting is modified by your current fondness for the game, a number defined by the goodwill it has gained and lost.)
Things that will make people check the hints/go for the walkthrough/stop playing your game:
Game-killing bugs
Insufficiently clued puzzles
No idea what they’re supposed to be doing
The perception that the game is unfathomably complex and they are too dumb for it
A crushing sense of ennui
The feeling that the author clearly didn’t give two farts about their game so why should anyone else question mark
Things that deplete goodwill without necessarily being gamekillers:
Lesser bugs
Unhelpful/default failure messages
Inconvenience (wrestling with inventory limits, the same long sequence of commands required often, etc.)
Unnecessary disambiguation questions (did you want to open the tub of margarine, or Margarine the cat?)
Guess-the-verb/unrecognized synonyms
Lack of implemented abbreviations (X SUPER-ADVANCED MEAT CALCULATION DEVICE)
Things that should be implemented but aren’t (the television is not something you can turn on)
Default messages left in place (some people check for every. single. one.)
Crazy complex rigmarole required to set up the homebrew system the game runs on (although some people feed on this)
What you the author need to do is make sure at least the first part of your game is smooth and compelling. (Preferably the whole game is, but if you have to pick, front-load the best bits and build up that goodwill.) Give the player a clearly defined goal, and demolish anything that gets in the way of them accomplishing it. This does not mean having no puzzles, or trivially easy puzzles, but it does mean that you need to anticipate how the player is going to try to solve puzzles, and fall over backwards providing them with clues and useful feedback. Testing is really, really important for this, because different people’s brains work differently (e.g. everyone but you is a weirdo).
Think of a game of table tennis. Hitting a ball (typing in a command) and having it come back (give a non-default response) is way more fun than hitting it and watching it roll into the corner, or swinging at it and watching it disappear into thin air, or being asked what you want to hit it with. (I may have mixed my metaphor there but whatever my point is probably still stands.)
Now imagine that your game is a physical space in which you have installed all kinds of awesome systems and set pieces and things, and you want the player to see all of it to best advantage, but you yourself can’t be there to give them the tour. You’ll need to make sure that their path is clearly marked and free from obstructions, takes them past all of your showpieces, and doesn’t have them wandering back through the same parts of the facility with nothing new to see. You can’t leave them alone in a giant warehouse and trust that they will notice the elaborate miniature city you have constructed on the ceiling, or find the jetpack that takes them close enough to watch the animatronic miniature citizens, without some signpointing.
I’ve tried to keep this about what I personally want in a game instead of pretending to speak for all judges and reviewers, but I am going to make a blanket statement about most people, and that is that they want to like your game. Really. It can probably feel like every quote-unquote blogger is firing up your game going “Man, I hope this is rubbish, so I can skewer it and feel better about myself,” and I can’t say definitively that no one is, but I honestly believe most people want to spend their leisure time having pleasurable experiences. (I have seen people decline to play any game without listed beta testers, and I have never seen the reverse.)
So, nudge them into thinking your game is good. Test it. Polish it. Think about the experience you want people to have, and honestly assess whether it is delivering that experience. Oh, and watch out for red flags that will give seasoned comp judges the preconception that your game is awful. That is another list I could throw in here, a list of red flags. Man. This post is gonna be long.
Red flags (my personal list, because this is way subjective):
The blurb reads like you wrote it in thirty seconds, because you hate writing blurbs
The blurb makes the experience of playing the game sound awful
It’s set in someone’s dirty apartment
No listed testers
Spelling and grammar errors appear early and/or often
“You can’t see that here” or the like as the response to my first command
Intro text is so long I get sleepy
The player is not allowed to do anything interesting until they have Z’d through pages of exposition
Oh, and if you find yourself writing an ABOUT text explaining why your game is not finished, I strongly urge you to consider not submitting it to the comp until it is finished. If you were watching a movie that suddenly cut to the director saying “Yeah, so there were going to be some explosions and stuff, and the characters were going to grow and fall in love, and I had an idea for a really good bit where everyone is fighting by throwing buses at each other, but then I just ran out of time,” you would probably be fairly well cheesed off. You don’t have to be that guy. Please don’t be that guy.
I will shut up now.