Yeah, we're not getting anywhere far arguing like that. A cool thing would be to cobble together a _really_ simple game using Construct, the first one using the game design I've described and the second one your design - and let them poll on which experience was more fullfilling, more fun for them. I'd just love to see raw numbers in that case.

If you're complaining about the trial-and-error aspect of gameplay then it seems to me that you just want game content spoon-fed to you. No offense.

That's exactly what this is not about. You're oversimplifying what I was saying by putting a 'Games should be easy and all users should be able to get through them in the first run' phrase into my mouth.

That's not what I was saying. What I'm saying is that it doesn't make sense to 'surprise' a player by deaths that he can't possibly foresee.

[quote:9fk2tq58]Why should there be a warning that a rock is going to fall on your head? The rumbling room is warning enough. Be careful. Look around. If you get hit on the head, you learned a lesson. Next time you'll be more careful.

Take any person, I don't care how experienced he is in terms of games - I'm _very_ sure that everyone dies at the first try here. I take a step forward, the rock falls down, 1 second later I'm dead. In reality, I'd hear the rumbling, I would hear the rock falling off the cliff and I could jump away. But in games, especially in 2d games, I can't 'be careful', cause I can't freaking look upwards to see if it's safe to take a couple of steps or if the rock's coming down to kill me. It's about putting the player into a dangerous, risky situation - but the wrong solution is to just kill him 5 times until he learned exactly how the rock is going to fall and then let him follow that pattern. That's no fun, that's just being a marionette.

[quote:9fk2tq58]So you're easily frustrated. I can understand that. I don't have the time or energy to master every hard game that comes my way either, and I get frustrated at hard games too. But that doesn't mean that when I have a moment to play a game, I want every step of the game explained to me. It ruins the experience.

Again, not what this is about. It's not about making a game idiot-proof, but if a game takes away a lot of my senses, meaning I can't anticipate what's coming cause I don't see it - nor do I hear it - why the fuck am I getting killed at this stage? I had no chance of knowing beforehand what pattern the designer had in mind. Me running forward, stopping, walking a couple of steps backward and forward again - that's retarted. It's like you'd put a safe into a game and just expect the user to know what the code is, but if he fucks up 50 times, you'll tell him.

It's not about me being frustrated quickly, I'm making a living through being frustrated and solving problems all day - And when I spend time on a game, I want it to entertain me, not making me work my ass off by figuring out what sorta idea the designer had in mind.

[quote:9fk2tq58]The only warning or tutorial any game should need, in my opinion, is "It's dangerous to go alone. Take this." The original Legend of Zelda took me literally months to beat. The joy of that game was in the discovery, in the trial and error. You wander into a place where you shouldn't be yet, you're screwed. No warnings. You learn the rules as you go.

Yet, that was 20 years ago and look at how Zelda has evolved. It's not fucking unfair anymore like Zelda 2 was (worst title in the series), it's fair now. I haven't played the newest title and thought the Cel Shading stuff was pretty boring, but if I compare A Link to the Past with the original Zelda, I mean... there's no comparison in terms of quality.

[quote:9fk2tq58]You want artificially extending gameplay? It isn't getting a nasty surprise that kills you instantly... it's making the player struggle with a wonky, non-standard control scheme for a section of game that has very little to do with the plot. Challenge the Yeti to a snowboarding contest. Fly a bird up the river... and pop balloons. What the hell?

Yeah, but those are just examples of bad design choices. That doesn't mean that expecting the player to know what'll happen in the next scene is a good design choice. It's just less shitty than what you've listed above.

Alright, let's dissect what you're saying.

So, taking Another World as an example again - Chahi definitely tried to give the player a cinematic experience, right? I mean, that's obvious. And what happens for 95% of all users? They'll die like... 20 times within the first 10 minutes. Or even more often. So it's like you're watching a film and every 2 minutes, we're rewinding for 15 seconds and you have to watch all that shit all over again. The problem is that even the checkpoints are way off at times. So, the water scene: You have to shoot that rock thingy to make the water flow out so you can cross a path that you couldn't before cause the waterfall would drag you down and - guess - kill you. Right? Right.

So I do all that, I shoot the frigging rock, I manage to complete the water scene and I keep going. 3 screens later, I'm in the middle of a layout and suddenly an enemy comes from the right, immediately shooting at me and I'm dead. Now, can I just play that scene again? NO! We're put right before the water scene again, so I have to do all of that again just to get to that scene where I just died to get another chance to jump into this trial and error scheme again! Seriously, that's fucked up.

The player has obviously already beat that passage of the game - Why would make him repeat an annoying passage in the game all over again if he dies in a completely unrelated scene further on in the game? That's what I loved about Braid: Blow didn't just stack up Puzzles we ALREADY COMPLETED over and over again, varying them slightly, just so you're playing for 20 hours and not for 6 hours, instead he gave you 6 hours and those 6 hours and fucking good 6 hours.

The only reason why we take that shit is because it's a game and we've been conditioned in the 80s that 2d games are hard and we have to beat them by trial and error (and that was only true because most games were based on coin-ops, so you needed to die every 2 minutes to throw in another coin). Now, I consider my time to be valuable and if I buy a game - and games usually aren't really cheap - I want to be entertained, right? Of course I also want to be challenged, but there's a difference between annoying the player by killing him on purpose and letting him repeat annoying passages again and again, just so the players dies more and more often, which in turn means that he'll have to invest even more time to beat the game and challenging him in a fun way that the player could even realistically manage in his first run.

That laser scene in Megaman 2 is one of the most fucked up design mistakes I've ever seen in any game. You literally jump into the scene and you're dead a second later if you can't predict the future. Another World has something very similar, with the falling rocks. So you're entering the layout, you take a step forward and a rock falls onto your head. THAT'S FUCKING STUPID.

Give me anticipation. Tell me that the rocks are falling. In reality, I'd see that the freaking environment looks like it'd fall apart too, so I'm very careful to begin with - but if you don't give me shit and just kill me, I'm gonna hate you. And you don't want your customers to hate your game.

I certainly agree with you on the fact that games have become too easy in the recent years and I also hate the frigging in-game tutorial levels that can take up to an hour to complete. That's nuts, right. But in the end you're just building something for someone else to have fun with, to inspire him, to tell him a story - Don't let the player zip through the game in an hour, but killing him every 2 minutes will make him put away the controller sooner than you could imagine, so you're not on target.

What? Say I wanna create a layout with a blue sky in the back - and as I walk across the layout, the sky gradually becomes darker and greener. That sorta stuff would work nicely with a HSV Slider that could be triggered.

And it'd still be a bitch to adjust the cones dynamically for each enemy type. A cone detector behavior where you could directly modify the cone and where the obstacle stuff is in there (my guess is that you shoot a vector and if the vector hits a solid, it stops detection from that point on) without me having to do workarounds would be neat. You could use them for most genres anyway, whether it's top-down or a sidescroller, it just makes sense.

Some good points here.

This is the first one I'd like to address:

Casual Fun vs Stupidly Hard

I like to die. I like to play a game and die die die die and die some more. I don't like to spend $100 on a new game, and finish it in 2 days without a challenge. 'Ninja Gaiden'... I died more playing that game than I did playing 'Another World'. I think that a lot of people just get too frustrated when playing games that are too hard... so there needs to be a difficulty adjustment ranging from "cake walk" to "fucking impossible bullshit wtf"... then make it a dynamic self-adjusting difficulty so if you die 5 times trying a section of the game, it automatically drops the difficulty down a slight amount or at least prompts the player with something like "you sure do die a lot, would you like to make the game a little easier?".

There's a big difference between dying because the player screwed up, made a mistake or by just expecting the player to know something that he can't possibly know beforehand.

Example: In Megaman 2, I don't remember the enemies name, but there are countless stages where the developer just expected you to fucking know that - as soon as you'd enter the next screen - a laser would be fired at you and if you don't jump in the first second after entering the screen, you'll die.

That's fucking frustrating. There's NO reason why a designer should do that except trying to artificially making the game harder / more frustrating / longer, because you cannot know that shit.

Because we talked about Another World, Another World has a few really shitty examples in it too:

For example, the passage where the player needs to swim around - how in three damn fucks am I supposed to know that the left path will lead me to a place where I can gasp for air again, whereas I am dead if I take the right path? This example isn't quite as unfair, cause that can happen in real life too - you take the wrong path and you lose.

But in a game, even if you want your game to be a real challenge - don't do that. It'll just frustrate the player and the first reaction in everyones mind would be: "How the fuck am I supposed to know that beforehand?"

Of course a lot of old school games from the 80s basically always worked that way. You screw up in a level to find the right strategy and then you try it again until you got it - but you can't complete it in the first run, cause it's fucking unfair. In todays time, if you're unfair and I have to repeat a certain part 30 times, you can bet your ass I'm gonna turn off the console or the PC and do something better with my time.

Challenging a player != expecting him to know things he can't possibly know.

Personally, I'm not a fan of different difficulty modes - In fact, I hate that, cause you need to balance your game 3 times and in most cases only 1 mode really works well, cause it's the one mode that the dev spent the most time with balancing it. A game should start slowly and let the players learn the controls and the conditions of the game. Then, gradually increase in difficulty - but not by being unfair, but by making the challenges harder and expecting more from the player. Yet, don't expect him being able to predict what's gonna happen in the next level.

33) http://www.davidhellman.net/blog/the-ar ... id-part-4/

I really liked this article by David Hellman about how they created the art for Braid.

It seems like the idea was a little different than Braid. If I interpret it correctly, they were able to play the runtime directly in the editor, which would be pretty cool. That's one feature I loved about the CryEngine - you can just jump into your layouts from the editor and don't have to let another window pop up, etc.

Maybe you'll guys find some inspiration reading that

32) A 'Recorder Object' would be cool. So that'd just be a simple global non layout object like mouseKeyboard and we could set a key to it - then, at runtime, we could hit the key and the recorder object would start doing it's thing. We could record pre-defined movements for each sprite (character) to help with the cinematics stuff and could mix it with pre-defined animations.

Valve has something similar which they've used to do some of the TF2 cinematics:

http://media.moddb.com/images/groups/1/ ... mmaker.jpg

As far as I know, you basically record one character after the other, can set the camera and stuff and get a couple of export options.

This idea isn't well thought-out yet, but it'd be cool having a simple to use method to create in-game cinematics that aren't player controlled.

31) Fuck the idea with my 'level sheets' - that's a load of crap. We can just organize everything in folders. I created a quick mockup:

<img src="http://thomasmahler.com/files/construct/images/projectMockup.jpg">

30) I have a new idea for a behavior: A 'detector cone' object.

So, basically, if you've ever played Metal Gear Solid, you know what this is about:

<img src="http://thomasmahler.com/files/construct/images/radar.jpg">

We basically would assign 'cones' to a sprite and define it as their 'detector' behavior. So, we could trigger events like:

If player is being detected by enemy1's cone - enemy 1 is alerted.

The cones should be parametric, so we could define the cone angle and the radius directly in Construct - so we could dynamically create enemies that don't see very well and others that see really well, etc.

Also, the cone should be able to be stopped at obstacles. So, if I stand behind a wall and the enemy is on the other side of the wall, if his cone detects me, the event shouldn't be triggered if there's a wall between us.

This could be used for enemies in general in a multitude of genres or even for those desktop tower defense like games, where every turret would act according to such a cone detector.

This would pretty much allow us to simply emulate the visual AI of enemies - so they can actually see or not see a player and their detection 'skill' is based on the creators input.

Now, how could we solve this for sound? Like, if the player has to sneak through an area, how do we create a gameplay event that'd help us define the area in which enemies can hear the player if he starts being noisy by running, jumping, doing something noisy in general?

Now, our ears don't work like our eyes in the sense that they're not bound to a certain angle. We can hear what's behind us, but we can't see what's behind us.

So the solution should be simple - a sphere! In the same way we attach the cone to emulate an enemies visual field of view, we could create a sphere to basically emulate his ears. So, if the player makes noise inside of an enemies 'audio sphere', we could trigger an event that'd alert the enemy.

So, for a sneaky game or generally just to make our enemies smarter and sorta emulate a simple AI that could be very convincing already, we could add events and base triggers on them. So, we could make a private variable that would define how much noise the player makes - and this'd just be an additive variable.

So, the base noise would be 0. If the player just walks around, he doesn't make any noise at all. If the player starts running, we set the private variable to '+20'. If the player jumps while he's running and lands, we add another '+20' the moment he hits the ground. If the player activates some switch, we could add '+10', etc.

Because of that setup, we could again create enemies that hear well and enemies that can't hear shit. That could make for some very interesting gameplay schemes, where the enemy has to outsmart those relatively 'simple' systems, but it sorta feels real, cause you don't see the cones and spheres.

So if we give our enemies cones for sight and spheres for sound our enemies could actually SEE and HEAR us, which could be a HUGE plus for a lot of games.

Metal Gear Solid on the original PSX was awesome because the enemies seemed a lot smarter because of that - and imagine a zelda-like adventure where you have to sneak inside of the castle and actually be smart about your movement and about when you start running, activate switches, etc.

Also, it'd help for these kind of games:

The problem with 'Z' for jump is that on a lot of european keyboards the Z and Y keys are swapped - so you need to hold your hand in the middle of the keyboard, which is kinda awkward.

That'd be really cool to have!