I finished Braid this morning. It was a lot of fun… I got through all but one of the puzzles, but had basically no clue on that one so I cheated and used a walkthrough.  Once I heard the solution, it turned out to be something I totally should have figured out.  Ah well. You can get through all the worlds easily enough without solving all the puzzles, so if you’re stuck I’d suggest trying other puzzles and coming back to the ones you’re stuck on.

I have two non-spoiling comments on the ending:

  1. The timed part at the end was really dumb.  Games do this all the time (“Let’s take our careful, thoughtful gameplay and add time pressure so it’s more exciting!“) and it always results in a completely different kind of game that the players have not been trained to play. Maybe that was on purpose as some sort of artistic statement, but that doesn’t make it any less dumb. I ended up finding a video on YouTube that showed me I was about 0.5 seconds behind the required pace and had to back up and do half the level over to get through one section.  The margin of error was way too small and the whole level was extremely frustrating.
  2. I don’t get it.  I don’t know what the story was about, in the end. I had fun anyway, though, so it doesn’t matter.

Anyway, it’s well worth the $15. I suggest you pick up a copy and give it a try.


6 Responses to “Braid”

  1. Kevin replied on :

    Surprisingly (given my usual take on timed segments) I didn’t mind the last bit at all. I think it’s because I could rewind, so as soon as I realized I’d made a mistake that would cost me too much time, I could immediately rewind and fix that mistake. So if I missed a jump or didn’t jump off a ladder early enough, I could say ‘whoops’ and fix it right then. It was like having an editor for gameplay.

    As for the story: Raymond finished the game and the optional content. Tynes sent me a message when he saw me playing it: ‘There is more. It starts in world 2-2 and continues outside your front door.’

    Whether that reveals enough of the story to make it make sense to you, I don’t know. However, I can say that the obvious first step is to google any of the text in the epilogue that is obviously a quote from something.

  2. David Hunt said on :

    The Penny Arcade comic ( sums it up pretty nicely. I was impressed with the ending, but the gameplay felt out of place in that level.

  3. Joe said on :

    The problem is that the margin between success and failure on that last level was so small that you pretty much had to do the whole thing perfectly to survive. When I went back and did it the time that actually worked it just happened to be the first where I didn’t have any extra 1/4 second delays. That might be fine for a pure platformer, but I certainly wasn’t prepared for that kind of hectic pace.

  4. Kevin wrote on :

    I guess my problem with timed sections like the end of Halo 1 is not that it’s timed; it’s that the time between me realizing I’ve screwed up and me getting to try that particular task again (a missed jump, a bad turn, whatever) is generally measured in minutes. If I make a mistake, I spend the rest of the next attempt anticipating that moment where I screwed up, getting tense and stressed out, and knowing that even if I don’t make that mistake again, I’m likely to make another and have to start over anyway.

    I didn’t have that problem in Braid’s final level because the time between me screwing up and me attempting that task again was measured in seconds; as well, any future mistakes wouldn’t require me to do that particular challenging bit again. It was like a continuous rolling save point.

    It helped that I’d already watched Raymond do part of it, so I knew going in that I’d have to do it perfectly, but even so I screwed up a dozen or more times, including one long screwup where I discovered I’d done the wrong thing half the level earlier.

    If every platformer could give me that sense of constant recording of my successes and erasing of my mistakes, I’d play a lot more platformers.

  5. ejengstrom commented on :


    Have a puzzle for you:

    Customer has a ULC front end running under Tomcat/J2EE connecting via hibernate to Oracle RDBMS.

    They have an application framework that is overdesigned/underdeveloped. One of the framework methods was to return a complex table view to the ULC client. A lot of rows. With even moderate rows of data the results were taking 38 seconds to render in the client.

    Upon profiling, the individual client in ULC was building the columns and preparing for the results and taking a lot of time doing so. Then it checked Hibernate for cached information, threw out the pre-rendered columns and started population from cache.

    It then queried the database for changes when done, before presenting to the client and replaced all the effort with the results of the query. 38 seconds.

    A single two letter change was made in the framework that reduced the overall transaction to about 5ms.

    Once the change was made the table was pre-rendered and loaded quickly with the results. What was the two letter code change to the framework and where was it implemented in the methods?


  6. Brady commented on :

    The timing tolerance on the last level had to be that tight or it wouldn’t have played back correctly. Really, you’re being forced to conform to a scripted sequence. Having that last rewind work right is sort of the key point in the whole game.

Leave a Reply