Saturday, 21 July 2012

ir/rational Redux - Postmortem

My word. It's done better than I expected.  In under a fortnight...

  • ir/rational Redux is just about to hit 100,000 plays on Newgrounds
  • 300,000 plays web-wide
  • 4.11/5 stars user rating on Newgrounds
  • 6th highest scoring game this month
  • 40 pages of user reviews and comments

Play Session Stats
But let's look at some play stats rather than a bunch of self-congratulatory ones. Just under 50% of everyone is quitting out in 1-5 minutes. This sounds high, but doesn't particularly surprise me - I know the second I saw a massive page of text in a Flash game I'd consider quitting out; and if you make it past that to the second or third puzzle you're going to realise quickly enough whether or not the game's for you.

Once past the 5 minute barrier, though, there's a dip in the drop out rate and if you stay longer than 5 minutes chances are you'll stay longer than 10 too. I've got you hooked.

The average playtime for those that play past the first couple of puzzles is 15m-26m. This is great, because it means most people who play on, finish the game, or at least spend 20 minutes trying and then rage quit on level 7 or level 9 (difficulty spikes, bad design).

- 28% of everyone who lasts more than 5 minutes spends more than half an hour playing the game.
- 3% of the same take about an hour.

This are invaluable numbers when it comes to thinking about what's next for the concept. The biggest and most obvious hurdle is this - when you can't just adjust a value and give the bad guys twice as much HP, how do you manage the difficulty curve for a game which takes some people ten minutes, and some people 5x that?

Flash Marketing
I've learnt very quickly about flash game marketing during the last fortnight. Reddit is your friend, lover and guardian angel. My day 1 post on there - no doubt aided by my shameless credit name-dropping - immediately put me about 10,000 clicks ahead of everything else on the site. This in turn got me a daily prize and into the Hot New Games list. The clicks and good reviews I scored from there got me onto the front page. I'm now off the front page and the hot new games - and clicks have dropped off quickly - but the game should pick up again any day when it makes it onto the Best Games This Month list.

Netting some good website write-ups helped, especially when they came in the first week. Due to the way the portals feature games, you're better off with 50,000 plays in week 1 than with the same (or even more) spread out over a number of weeks.
"[I]t contains nifty logic puzzles, darkly amusing writing from Penumbra scribe Tom Jubert and has been reduxed so hard it looks and sounds brand new."Rock, Paper Shotgun 
"No doubt the central premise of ir/rational Redux could have come off as incredibly dry, but overall this is a supremely engaging work. A unique brand of dark philosophical humor is present throughout, and the puzzles manage the right balance of posing a challenge to advanced logicians, while remaining welcoming to the novice."JayIsGames
"The conclusions you have to reach are sometimes obvious, but the challenge of the game is figuring out how to reach them. While the game does take a brief, stupid side turn into the politics of video game censorship, most of the puzzles are amusingly high-minded."
"I can guarantee you've never played a puzzle or escape the room type flash game like this before so you are in for a treat."
A special shout out also needs to be made to Jay of JayIsGames for buying me the full version of the Clickteam game maker package so I could release the game unbranded. Massive compliment - that man knows his flash.

What I Learnt About Designing Games
(Or a list of all the crap I did wrong)

Difficulty & Designing Logic Puzzles
I approached the puzzles from the wrong direction. I started with a story, and tried to come up with logic puzzles that fed into it. As a result they feel crowbarred in - while they sit nicely with the story, the difficulty curve is all over the place, and worse: what makes those puzzles difficult is the way I've programmed the solutions, rather than the logic itself.

In the next game I'll begin with the logic, building complexity slowly, and set the game in a scenario more flexible in the opportunities it provides for puzzles.

Art is Important
I know, obvious, but I've been genuinely surprised by how polished people are finding the game - because behind the scenes, of course, it's all over the place. What's key here is this: this is essentially the same game I released three years ago, only with pretty pictures. The difference in response, though, has been massive. Good art is important.

How to do Flash Right
I've learnt a bit about the nitty gritty of Flash development as well. Don't get me wrong - I know nothing about Flash itself, but I know a bit more about alpha transparency and file compression and in-game ads and analytics. 
  • I know how to make a Flash game that's less than 18MB now (use as little alpha transparency as possible, use a non-lossy image format, don't store your text as image files, compress everything!). I also know you can make a successful Flash game at more than 5x the usual file size.
  • Implement Mochi for ads and analytics
  • Don't release buggy games
  • Don't use drag and drop software unless you really can't program (in which case do use it)
  • Do come up with a great idea and implement it to at least a mediocre standard.
  • KISS: Keep It Simple, Stupid.

And What I Learnt About ir/rational as a Concept
The single most positive part of this whole experience is the sheer volume of player discussion and feedback. There is something fundamentally strong about this gameplay concept - it makes people feel smart, and it makes people want to talk about the game. I'm certain it's performing far better than it otherwise would simply because people are talking about it so much. On Newgrounds there are games with 3x as many plays as ir/rational, but with less than half as many comments. This game is a viral marketing gift. End of.

The other thing that players are really appreciating is the originality of the thing. I remember being in my early teens and discovering PCGamer magazine, and a new value in games: 'innovation'. Nowadays it feels a bit old hat to talk about it, but if I had a penny for every review along the lines of "I've never played anything like this, 5 stars" I'd have... well, more than the $100 dollars the ads have earned so far. There's a lot of clones and fluff on Flash portals - doing something smart and different picks you out from the crowd, even if you don't implement it that well.

What's Next?
I'm currently working on a development of the ir/rational DLC concept. Only very different. I'll know if it's going anywhere in a month or two.

You can play ir/rational at these fine places:


  1. Cool, let us know how it works out. In about two months (if I'm not crammed too much with my thesis) I'll have my concept for the logic-game I told you about earlier worked out much better too. I've made some progress at least, knowing that making money off of this kind of thing isn't an option. But places such as newgrounds are still good places for people to find your games, so one must adhere to their standards and whatnot anyway... Any problems of any kind with that?

  2. Do you think that the spike in difficulty is related to the fact that some puzzles rely on using the contra-positive of the given conditional?

  3. @ Ava
    Problems with adhering to newgrounds standards? Not really. There's a nasty mac bug in the game to do with scrolling that I can't fix and is rather unprofessional to leave in, but it's not put off any sponsors. Any other requirements have just been about branding etc, which is easy enough to slot in.

    And my solution, at any rate, is going to be to try to make money off this kind of thing regardless. You just never know.

    I'm sure that wouldn't help, but I think the real problem was just a lack of considerartion on my part for upping the difgiculty clearly and consistantly. I don't teach enough in the game, and sometimes people flounder as a result.