1 00:00:00,840 --> 00:00:06,160 JONATHAN BLOW: It was very clearly the case that more ideas came out of the development process, 2 00:00:06,160 --> 00:00:11,880 and ended up in the final game, than I put into it as a designer. 3 00:00:11,880 --> 00:00:16,289 The process of designing the gameplay for this game was more like discovering things 4 00:00:16,289 --> 00:00:22,600 that already exist than it was like creating something new and arbitrary. 5 00:00:22,600 --> 00:00:25,060 And another way to say that is that there was an extent to which 6 00:00:25,070 --> 00:00:27,970 this game designed itself 7 00:00:31,760 --> 00:00:37,600 This is Game Maker's Toolkit, I'm Mark Brown. 8 00:00:37,600 --> 00:00:43,730 That was Jonathan Blow talking about the rewindable platformer Braid at the Game Developer's Conference 9 00:00:43,730 --> 00:00:45,170 in 2011. 10 00:00:45,170 --> 00:00:49,440 What Blow's describing here is a philosophy of game design that he used when making both 11 00:00:49,440 --> 00:00:56,390 Braid and The Witness where rules and puzzles were discovered through programming and play-testing, 12 00:00:56,390 --> 00:01:01,079 rather than designed through the implementation of some preconceived idea. 13 00:01:01,079 --> 00:01:06,569 So with the Mario-like platformer Braid, he started with a mechanic - the ability to turn 14 00:01:06,569 --> 00:01:09,700 back time by a practically unlimited amount. 15 00:01:09,700 --> 00:01:15,719 In the process of coding that, new ideas emerged. If he was rewinding the position of everything 16 00:01:15,719 --> 00:01:20,539 in the world, he could choose to not do that for certain objects, and thus make them immune 17 00:01:20,539 --> 00:01:25,749 to your ability to manipulate time. A rule was born. 18 00:01:25,749 --> 00:01:30,810 After implementing these new rules, Blow could play the game and look for consequences that 19 00:01:30,810 --> 00:01:36,270 he perhaps did not foresee. Like how if a moving platform was immune to time travel, 20 00:01:36,270 --> 00:01:41,099 the hero could rewind to a point where the platform is no beneath his feet, and would fall 21 00:01:41,099 --> 00:01:46,319 down as soon as he stops manipulating time. 22 00:01:46,319 --> 00:01:47,489 That's kinda cool. 23 00:01:47,489 --> 00:01:52,989 So each puzzle became an illustration of one of those phenomena, so that by solving it, 24 00:01:52,989 --> 00:01:58,319 the player would stumble upon that interesting fact about Braid's unique universe - the same 25 00:01:58,319 --> 00:02:02,099 fact that Blow himself discovered while programming the game. 26 00:02:02,099 --> 00:02:08,020 A similar process was used in The Witness, where Blow made rules and puzzles by exploring 27 00:02:08,020 --> 00:02:14,219 the mechanic of drawing lines on a grid. Play testing this showed Blow that he was often 28 00:02:14,219 --> 00:02:18,180 partitioning grid cells - perhaps that could become a rule? 29 00:02:18,180 --> 00:02:22,709 Which leads to situations like this. This puzzle is pretty easy to solve: you just loop 30 00:02:22,709 --> 00:02:27,409 around here and you're away. The next puzzle looks identical but you'll notice that the exit 31 00:02:27,409 --> 00:02:32,670 has moved. Now, using that same solution will cut off your access to the exit. So you have to 32 00:02:32,670 --> 00:02:35,640 solve it like this. 33 00:02:35,640 --> 00:02:41,480 Here, the mechanic of drawing a line inspired a rule about partitioning cells which had 34 00:02:41,480 --> 00:02:47,140 the consequence of cutting off your exit, which led to a puzzle illustrating this fact. 35 00:02:47,140 --> 00:02:52,780 Describing the invention of this puzzle type at IndieCade in 2011, Blow said... 36 00:02:52,780 --> 00:02:55,860 JONATHAN BLOW: That came from asking these little known questions. 37 00:02:55,860 --> 00:03:00,740 It didn't come from a top-down imposition 'I want to make a puzzle type that... blah' 38 00:03:00,740 --> 00:03:06,540 Rather, it came from this very simple process of exploration very early in development. 39 00:03:06,540 --> 00:03:12,340 While Blow may have largely abdicated the duty of designing puzzles to, I dunno, the 40 00:03:12,349 --> 00:03:15,549 universe, he still has some important roles to play. 41 00:03:15,549 --> 00:03:20,840 First, is making sure the ramifications of each change are explored to the fullest. In 42 00:03:20,840 --> 00:03:25,440 The Witness, Blow asked how every part of the game could be twisted, and that includes 43 00:03:25,440 --> 00:03:28,840 the grid, the cells, the line, the environment, and the panel. 44 00:03:28,840 --> 00:03:33,480 And in Braid, you'll notice that the consequences of each rule change are explored by every 45 00:03:33,489 --> 00:03:34,730 object in the game. 46 00:03:34,730 --> 00:03:39,469 In the world where objects can be immune to rewind, for example, there are puzzles where enemies, 47 00:03:39,469 --> 00:03:44,489 keys, doors, clouds, platforms, and even the player character have this property. 48 00:03:44,489 --> 00:03:50,219 Blow's second job is to present the resulting puzzles in a way that will give the player the best 49 00:03:50,219 --> 00:03:54,579 possible set-up to discover the interesting fact at the heart of the conundrum. 50 00:03:54,579 --> 00:03:59,670 For example, he frequently uses misdirection to lull you into making a seemingly obvious 51 00:03:59,670 --> 00:04:05,170 move - only to show you that this is not correct. In the Braid puzzle "Hunt", you're told to 52 00:04:05,170 --> 00:04:09,319 kill all the monsters but they're set up in a way that if you kill them in the most obvious 53 00:04:09,319 --> 00:04:11,519 sequence, you're unable to solve the puzzle. 54 00:04:11,520 --> 00:04:17,100 Misdirection like this stops the player from brute-forcing the puzzle and failing to grasp the interesting 55 00:04:17,109 --> 00:04:22,130 fact. And showing the player why something doesn't work is often part of that fundamental 56 00:04:22,130 --> 00:04:25,810 truth that Blow is illustrating in each puzzle. 57 00:04:25,810 --> 00:04:31,500 The designer also uses sequences, pairings and reprisals. If you come across a simple 58 00:04:31,500 --> 00:04:36,180 puzzle - like this one about trying to unlock two doors with one key - you'll likely come 59 00:04:36,180 --> 00:04:38,610 across a more substantial version in the same area. 60 00:04:38,610 --> 00:04:42,810 And by using familiar layouts in different worlds, with different rules, you can see 61 00:04:42,810 --> 00:04:48,380 how the consequences have changed. This level is essentially repeated in Worlds 2 and 4, 62 00:04:48,380 --> 00:04:52,310 but the way time works in each means the solution is unique. 63 00:04:52,310 --> 00:04:56,520 Jonathan Blow also subverts the rules you're used to. In the level Irreversible, you have 64 00:04:56,520 --> 00:05:01,340 to realise that you must not use your rewind powers. And throws in traps, to catch out 65 00:05:01,340 --> 00:05:06,080 those who aren't thinking hard enough. In this level, the wacky way that time works means 66 00:05:06,080 --> 00:05:09,250 only one of these gates can be opened... 67 00:05:09,250 --> 00:05:14,979 Blow's final job is to be ruthlessly curatorial, and edit out mechanics, rules, and puzzles 68 00:05:14,979 --> 00:05:20,710 that lack a sense of surprise, or overlap with each other, or fail to say anything interesting. 69 00:05:20,710 --> 00:05:25,190 Both Braid and The Witness were spin-offs of games that were shelved because their main 70 00:05:25,190 --> 00:05:30,580 mechanics didn't present a rich enough space to explore. And Blow killed off rules, like 71 00:05:30,580 --> 00:05:35,699 Braid's weird turn-based world, because their consequences weren't surprising, or the rules 72 00:05:35,699 --> 00:05:37,979 felt contrived. 73 00:05:37,980 --> 00:05:42,360 But where Jonathan Blow will differ from other designers is that he deliberately left stuff 74 00:05:42,370 --> 00:05:46,669 in, even if it wasn't fun - simply because it was interesting or would make the game 75 00:05:46,669 --> 00:05:48,990 feel incomplete to remove it. 76 00:05:48,990 --> 00:05:53,780 Like this super weird puzzle where a key can bumble along on its own. It is, after all, 77 00:05:53,780 --> 00:05:57,870 a surprising and interesting consequence of this game's universe. 78 00:05:57,870 --> 00:06:02,750 Because for Jonathan Blow, a puzzle is never just a puzzle. It's a communication of an 79 00:06:02,750 --> 00:06:06,970 idea from the designer to the player. And solving the puzzle is the player's way of 80 00:06:06,970 --> 00:06:09,419 saying "I understand". 81 00:06:09,419 --> 00:06:14,419 And I think "I understand" is a significantly different concept to "I finally figured it 82 00:06:14,419 --> 00:06:19,449 out", which is how many puzzle games operate with their arbitrary steps and intricate sequences 83 00:06:19,449 --> 00:06:21,479 and red herrings and obtuse mechanisms. 84 00:06:21,479 --> 00:06:27,460 But the puzzles in Blow's games feel more fair. And that's why this design philosophy 85 00:06:27,460 --> 00:06:31,960 isn't just about letting the design help direct you to the next rule or the next puzzle - it's also 86 00:06:31,960 --> 00:06:35,480 about helping you make better, and more honest puzzles. 87 00:06:35,490 --> 00:06:39,830 Braid and The Witness introduce all the elements upfront and teach their mechanics quickly 88 00:06:39,830 --> 00:06:44,610 with introductory puzzles - from there the harder puzzles are only about understanding 89 00:06:44,610 --> 00:06:49,400 the consequences of those known mechanics in different set-ups, combinations, and layouts. 90 00:06:49,400 --> 00:06:54,759 And the puzzles can be blisteringly simple. Most are about exploring just one idea and 91 00:06:54,759 --> 00:06:59,720 the stages are small enough so you can consider all the moving parts at once. And there are 92 00:06:59,720 --> 00:07:05,409 no, or very few, red herrings, and also few arbitrary steps to finish. Once you've found 93 00:07:05,409 --> 00:07:08,449 the solution, it's relatively effortless to execute it. 94 00:07:08,449 --> 00:07:13,039 So solving a puzzle in this game isn't like solving a Rubik's cube or trying to guess 95 00:07:13,039 --> 00:07:18,250 at the answer to a riddle. It's simply seeing something that was there all along. The answer 96 00:07:18,250 --> 00:07:23,250 was right in front of your eyes, if only you knew the right way to look at the world. 97 00:07:23,250 --> 00:07:25,639 Kinda like those hidden puzzles in The Witness. 98 00:07:25,639 --> 00:07:30,300 So that "a-ha!" moment you get when solving a puzzle isn't about finally putting together 99 00:07:30,300 --> 00:07:34,900 all the pieces or finally understanding what the hell the designer was asking you to do, 100 00:07:34,900 --> 00:07:38,139 but it feels like you just saw the world a bit more clearly. 101 00:07:38,139 --> 00:07:43,169 As Jonathan Blow told Gamasutra, "the more that a puzzle is about something real and 102 00:07:43,169 --> 00:07:47,740 something specific, and the less it's about some arbitrary challenge, the more meaningful 103 00:07:47,740 --> 00:07:51,420 that epiphany is". 104 00:07:53,240 --> 00:07:54,289 Thanks for watching! 105 00:07:54,289 --> 00:07:59,490 One of my goals with GMT is to pass on the philosophies of the best game designers around 106 00:07:59,490 --> 00:08:02,580 so you can use their ideas in your own games. 107 00:08:02,580 --> 00:08:06,400 If you're interested, I've put loads of links in the description where Jonathan Blow talks 108 00:08:06,400 --> 00:08:08,270 more about the process. 109 00:08:08,270 --> 00:08:12,490 And it's not just for puzzles games - Blow reckons that this process of letting the design 110 00:08:12,490 --> 00:08:16,509 dictate the rules and mechanics could be used in other genres, too. 111 00:08:16,509 --> 00:08:20,009 As always if you liked the show you can leave a comment, give me a thumbs up, subscribe 112 00:08:20,009 --> 00:08:25,000 on YouTube, or even support the show financially on Patreon like these endlessly awesome gold 113 00:08:25,000 --> 00:08:25,699 tier supporters...