I just changed my entire game

Hi, my name's Mark and I am making a 
video game called Mind Over Magnet,

a delightful puzzle platformer 
about, you guessed it, magnets.

Now, when I returned to game development after 
my Christmas break, I noticed a very interesting

button in my Steam Developer Dashboard 
thingy that I can't show you due to NDAs.

Basically a button to submit 
the game to the next Next Fest.

This is Valve's online games convention 
where developers can share demos of their

upcoming games and it's proven to be 
a pretty powerful marketing technique.

I mean, look what happened to 
the very addictive poker-based

roguelike game Balatro in 
the last one in February.

So I knew I had to submit Mind Over Magnet 
to the next Next Fest happening in June.

I just needed to create a trailer, make a demo,

and fix one very annoying thing about my 
game that's been bugging me for months now.

Let me explain.

♫ Music ♫

Okay, so my game is pixel art, right?
But the sprites are pretty big.

I mean, compare Magnus to Mario, or Madeline.

This means that a typical room 
might fill up 1280 by 720 pixels.

So if you're playing on a really old 
720p monitor, great. Pixel perfect.

And if you're playing on a nice 
1440p monitor like I do, cool.

We can just double all of the pixels.

And if you're playing on a really nice 4K 
TV, then I can just triple all of the pixels.

But what about 1080p, the most commonly used PC 
screen resolution according to Valve's surveys?

Well, that's a bit tricky because 
that's 1.5 times bigger than 720p.

And, well, there's no such thing as half a pixel.

Now there's a few different 
solutions to get around this problem.

One is to just let Unity figure it out.

Like add an extra pixel here and remove an extra

pixel there to make all of those half 
pixels fit onto a full pixel grid.

But as it sounds, that's not perfect.

And things can look really squiffy and ugly,

especially when it comes to 
angles and curved surfaces.

Another solution is to use anti-aliasing 
to feather all of the edges.

But now everything looks blurry and it's no 
longer that crisp, retro-looking pixel art.

And solution three, or C, forgot what we're doing,

is to keep the pixels as they are but have 
a stonking great border around the screen.

And now everything is small and zoomed out.

Now I tried all of these solutions.

I put in a setting in the menu for pixel perfect 
mode that would add borders around the screen.

I coded up a system so it 
would systematically only

add anti-aliasing if the pixels couldn't 
be multiplied by a nice round number.

And at times I wondered if maybe 
players just wouldn't notice the ugly,

squiffy pixels and it was just a me thing?

But ultimately it felt like I was trying 
to find the least worst compromise.

And I spent a long time going back 
and forth between these things,

trying to find one of these three solutions that 
would kinda, almost, sort of work for the game.

Now of course there is a fourth solution 
but I don't even want to think about that.

And I mean it would work but 
I don't even want to say it.

Okay fine, I'll say it.

I could get rid of all the pixel art and replace

all of the graphics in the game 
with brand new 4K ready artwork.

But that... but that would be mad, wouldn't it?

So I replaced all of the pixel art in 
the game with brand new 4K ready artwork.

I took every sprite, tripled it in size,

and then remade it using Photoshop's vector 
tools like the pen tool and the shapes.

That's every button, every door, every decorative 
object, every magnet, every UI element.

And you know what?

It didn't actually take as 
long as I thought it would.

I mean my pixel art is already pretty chunky 
and made up of simple shapes with basic colours.

I also chose not to use this newfound pixel real 
estate to add in extra details because I wanted

things to still be super readable when crunched 
down to a smaller display like the Steam Deck.

And I also chose not to remake the background art.

Instead I kept that as pixel art and just threw 
on a depth of field filter to blur it out.

But no one needs to know about that, right?

That can be our little secret.

And after all of that work I'm actually 
really happy with how the game looks now.

Magnus has way more charm and character.

The game no longer looks like yet 
another generic pixel art indie game.

And the game just works.

You can put it on a 4K TV 
or a 1080p monitor or the

small screen of a Steam Deck and it just works.

It just works.

And the funny thing is, in the 
end I probably spent more time

trying to find a good compromise than 
I spent just remaking the art assets.

And so I've learnt a really good lesson from this.

Well, two lessons.

One is just plan ahead.

I made the sprites long, long, long before I made

the final camera system and just kind 
of hoped that they would work together.

They didn't.

I mean, there's this button in Unity for 
a pixel perfect camera and I kind of just

assumed I could turn that on and things 
would be perfect, which is dumb and naive.

That is not how it works but I know that now.

But the other lesson is sometimes it feels 
like a good idea to try and find a compromise

or a loophole or a workaround but 
that's kind of the lazy way out.

And sometimes it's just better 
to bite the bullet, knuckle down,

and do the hard work to make an actual real 
solution that's going to work properly.

So in the future I'm gonna save myself some time 
by just skipping ahead to doing the real solution.

Okay, so with that kind of a nightmare out of

the way it was time to move on to 
the next thing - making a trailer.

Now this is something I've put off 
for a long time simply because my

game just isn't done yet so there's not 
a lot of stuff to show in the trailer.

I mean I finished World 1 but if I just show 
that then the whole trailer is gonna be the

same footage of the same turquoise-y sewer 
system over and over again for two minutes.

Not a good depiction of what 
my game is really about.

So I decided to use the ancient 
art of "fake it until you make it".

Quite literally make it in this case.

Because I've got a load of stuff that isn't 
quite done yet like these levels that haven't

been playtested and are currently using the 
blueprint background design I use for development.

I've got some half-finished backgrounds and

some different mechanics that 
aren't implemented fully yet.

If I throw all that stuff together I can make 
some content that isn't 100% going to be in

the final game but is pretty indicative 
of what the game is going to be like.

Side note, I actually think I did 
too good of a job with this because

after I posted the trailer I got 
a number of comments from people

congratulating me on finishing the game 
which I will take, thank you very much.

It's not actually done. It's all fake!

So with this new content I then recorded 
myself playing through those levels.

I didn't have any music on but I did keep the 
sound effects on so those can be in the trailer.

Also, at the risk of talking 
more about screen resolutions,

I recorded everything in 4K but 
the trailer is only 1080p so I

can crop the most interesting bits without 
having to zoom in and lose image quality.

I then threw this footage into Adobe Premiere and 
cut out all of the most interesting little clips -

you know, interesting animations or depictions 
of mechanics or narrative beats or whatever.

Now as for music, I do actually finally have a

composer for the game now so I can 
shut that very large Google form.

Thank you very much for 
submitting your name, everybody.

They're not quite ready to work on the game yet

so for this trailer I just went to my 
favourite audio source, Epidemic Sound.

And the cool thing about Epidemic Sound 
- and this video is not sponsored by

Epidemic Sound but if you go to the 
description and hit my link I will

get paid by Epidemic Sound - is that when 
you download music you don't just get the

full song but also the individual stems 
like the melody, the bass, and the drums.

This meant I could isolate the drums and have it 
so the footage changes on the drum beat like so.

In terms of the narrative arc of the trailer,

I start with just really simple stuff like 
using the magnet beams to lift up blocks.

Then I slow things down to introduce Magnus 
as a character, then show what Magnus can

do as a game mechanic, and then have a 
rapid fire montage of other stuff you'll

see in the later half of the game like other 
magnets, other mechanics, and other worlds.

And then end with a nice call to 
action to wishlist the game on Steam.

Now this is not the best trailer ever made - I put 
it together in like the course of a single day,

I also don't have a huge amount of footage 
to use, and I'm simply not Derek Lieu,

you know, the guy who makes all of those 
awesome trailers for every indie game.

But I think it serves its purpose for now.

And also I posted it on YouTube 
and other socials and it pushed

the game past 30,000 wishlists which is 
pretty good before Next Fest even begins.

And then finally there's the demo.

Now I'm not 100% sure how much 
stuff I want to have in the demo.

Like the trailer, if I just 
have world 1 in the demo,

that's not going to be very indicative 
of what the whole game is like to play.

So maybe it should also have world 
2, or half of world 2, or whatever.

So I need to start making some of that.

But in any case, I think I just need 
to get working on finishing this game.

Because like, going back to Balatro, you 
might have noticed that the game built

up a huge amount of hype during Steam Next Fest, 
and then launched for real a couple weeks later.

Pepper Grinder is kinda similar:

it's launching about one month after 
the end of its Steam Next Fest debut.

And so if I'm going to showcase 
my game at the Next Fest in June,

I should probably look to 
release it in July or August,

and that's not hugely far in the future, 
and so I'd better get on with this game!

Now in the last few weeks I 
have discovered a pretty good,

not to sound too grandiose, production 
pipeline for getting this game done.

It's very much inspired by 
Valve's approach to making Portal,

which I discussed in that video 
on Valve's playtesting philosophy.

It basically works like this.

On Tuesday through Thursday I 
work on new content for the game.

New levels, fancy new backgrounds, or 
new cutscenes. Anything new for the game.

On Friday I work on bug fixing.

I take a few items off of my 
ever-growing Trello to-do list

and try to find a good way to solve that 
bug, and then do some more bug fixing.

And then at the end of the day I take what 
I have, test it, and export a new build.

I then give it out to some playtesters,

usually like a few randos on Twitter, 
but a maximum of three playtesters.

Then on Monday I check my Twitter DMs 
to find some nice playtesting footage.

I watch the whole video through and make as many 
notes as possible for things I want to change,

like bugs, or places that need more polish, or 
levels that are too easy or too hard or too fiddly

or have some exploit that I didn't foresee 
when I was developing that part of the game.

Now I've had some version of this process 
throughout the game's development,

but I feel like I'm really perfecting it now.

Like, in the past, I've given 
builds to too many people in one go.

That just ends up with way 
too much footage to watch.

It's just overwhelming, and if one person 
stumbles upon a bug or an exploit it's

really painful and not very useful to then see 
that same thing crop up in nine more videos.

Also I've given out playtest builds in the 
middle of the week, and then I'm just kind

of stuck because I can't do anything 
until I get the playtest footage back.

So doing it on Friday means people play the game

over the weekend when I really should 
be taking a break and playing Balatro.

And so with this process the game 
is really starting to take shape.

New content starts out a little bit rough and 
experimental but then quickly gets playtested,

and stuff that's been in the game for 
a long time, stuff in the early levels,

has been seen by so many playtesters 
and had so many little tiny tweaks

and bug fixes that it's now starting 
to get seriously polished and good.

And so with this process I feel 
somewhat confident that the game

will be done shortly after its Next Fest showcase.

But I can't do that if I'm standing here 
rabbiting onto you about what I've been doing,

so it's time to get back to 
development for some time.

I've got to run a playtest, 
or watch playtest footage,

or make new content for the game, 
or do bug fixing or play Balatro.

See the aforementioned schedule, compare it 
to today's date, you'll know what I'm doing.

As always GMTK Patrons can play the latest 
build whenever they want but if you're not

supporting the show financially - and I mean 
why not - you can just wishlist the game

now on Steam and you'll be notified when the 
demo is available during the next Next Fest.

Thank you so much for watching 
and I will talk to you soon.