What are the actual advantages of building your own engine?

[Squire Grooktook]

A bit curious about this. It's often said that building your own engine is the way to go if you want to squeeze out maximum horse power. However such statements are generally made on the subject of large budget games, featuring high tech 3d graphics. For a comparatively simple 2d game with sprite based graphics, what exactly is there to be gained from choosing to build ones own engine as opposed to picking a pre-made one?

[Cagar]

Squire, if you want to make a shmup, just get Gamemaker Studio. It has everything and more that you will ever need for making a sprite-based shmup, it's simple, and it's fast as hell. Thanks to my recent projects, I'm starting to be pretty confident in my gamemaker.
Wanna team up actually? I've got a shmup half-done already and I'm willing to share it with you. The development froze because I lazied and lost interest in the genre during that season.

[S20-TBL]

Think of it like cooking your own meals: at least you know what went into your dinner, and at most--assuming you have the cooking skills and reliable equipment to match--you can make exactly what you want for what you need to do.

The problem is gathering enough know-how to actually pull it off. Do you have the time to learn something from scratch? That's usually the biggest barrier getting in the way. Those early arcade developers knew electrical / electronics engineering and used it to their advantage machine coding ROM and processor chipsets to do things, for example, but not all of us have that luxury.

That's why having pre-built engines nowadays is a Godsend. Of course, the problem with using premade engines is that you're essentially limited by the engine architecture itself unless you can figure out workarounds, but my personal belief regarding them is: if it ain't broke...you know the rest.

Game Maker's a decent engine once you get past its limitations. Granted, it's a tad inefficient when it comes to memory handling, computing algorithms and especially graphics processing, but once you get past that you can actually create above average material if you put in enough time and effort.

[Emad]

Making a custom engine can maximize the performance of your game, if you know how to do it.
Is it worth it?
If you want to truly learn how games work and you are willing to spend many months on it then yes, it's a very good experience.
If you just want to make a game don't even think about it. Pick any existing engine and go with it. Game Maker is a tad buggy and unoptimized but overall is fine if you don't have to render 2k+ entities on screen.

[Cagar]

Making a custom engine can maximize the performance of your game, if you know how to do it.
Is it worth it?
If you want to truly learn how games work and you are willing to spend many months on it then yes, it's a very good experience.
If you just want to make a game don't even think about it. Pick any existing engine and go with it. Game Maker is a tad buggy and unoptimized but overall is fine if you don't have to render 2k+ entities on screen.

There's no way you can make a shmup too heavy to run with gamemaker, unless that is specifically your goal. Otherwise this is just misinformation.
(If we don't count single-core processors and windows XP)

[Emad]

There's no way you can make a shmup too heavy to run with gamemaker, unless that is specifically your goal. Otherwise this is just misinformation.
(If we don't count single-core processors and windows XP)

Yes, because OS and number of cores have a HUGE impact on single threaded games.
I don't understand why you are saying that I'm making disinformation, I stated that if need to have 2k+ entities on screen it's better to avoid game maker. It's well known that gm doesn't handle well too many objects. If your shump doesn't need so many entities (it shouldn't) then it's ok.
I'm speaking from direct experience, my game at the moment doesn't go over 1000 entities and runs ok. More than that and the fps begins to drop on my mid high end 2.5 years old laptop.

[Cagar]

Well I knew that the way I described old computers would be nitpicked with the exact way you did, but still didn't bother to edit it. Goddamnit I admit that it was a mistake.

I'm just reading heavy implications that gamemaker being badly optimized somehow prevents making a shmup with it in any way, which is absolutely not true.
Anyways, I've been defending gamemaker dangerously much here lately, so I'll stfu for this thread instead of looking like a marketer.

Still, it's a tad ironic to read this from a guy who got a 1-stage long touhou fangame out after.. a few years of development?
That's called building your own engine I guess. Have fun with the good optimization at least when you have it.

EDIT: Please don't tell me that you're using GM though

EDIT2: Just so that this post isn't so offensive, i'll also answer to this:

For a comparatively simple 2d game with sprite based graphics, what exactly is there to be gained from choosing to build ones own engine as opposed to picking a pre-made one?

You get the praise from l33t c0d3rs in an unknown forum because your code is shining, optimized and 100% self-made.

[Emad]

Well I knew that the way I described old computers would be nitpicked with the exact way you did, but still didn't bother to edit it. Goddamnit I admit that it was a mistake.

I'm just reading heavy implications that gamemaker being badly optimized somehow prevents making a shmup with it in any way, which is absolutely not true.
Anyways, I've been defending gamemaker dangerously much here lately, so I'll stfu for this thread instead of looking like a marketer.

It's funny because the whole thread screams "use gamemaker". I'm also saying that!

Still, it's a tad ironic to read this from a guy who got a 1-stage long touhou fangame out after.. a few years of development?
That's called building your own engine I guess. Have fun with the good optimization at least when you have it.

EDIT: Please don't tell me that you're using GM though

Well, I see that you have edited the post.
I started this game by making my own engine C++/OpenGL a few years ago with long pauses here and there. Then I got fed up and ported everything to gamemaker earlier this year. It sped up significantly the development.
I don't know if you read thoroughly the posts or not, though. I said that I'm using it.
But I don't want to start arguing about something we both agree so let's drop it.

[rfeese]

Advantages? I think it mostly boils down to more control over how the game engine works, and possibly some legal ownership and portability issues. What that actually means to you depends on your goals from the project.

Building your own engine allows you to control the low-level details of how your game works. If you have the technical knowledge and time to work out all these details, it can be a very rewarding process. If you control all the source code for the game engine, it may give you more freedom as to how you licence and distribute the game, but this also depends on other factors, such as any other code libraries you may be using. Building your own engine may also allow you to target platforms that might not be supported by an existing engine. On the other hand, a commercial engine may have support for more platforms than you can easily support unless you have experience building for different platforms.

I've gone the way of building my own engine, because I come from a developer background and I want to learn how to build a game engine. Along the way, I have discovered that I also need to build tools to be able to build the game assets to run in my game engine. This all takes time...Much more time than I ever imagined, even though I had read stories from other developers. Although I haven't used a tool like GameMaker, I suspect one could save enormous amounts of time using such a tool if you are more of an artist or game designer than coder.

[trap15]

Writing engines is fun.

[CocoPops]

@trap15 not for the faint of heart!

[Squire Grooktook]

Thank you all for the replies.

Maybe rephrasing the question just a bit might better. I know that creating your own engine gives you more under the hood undestanding or control of your game, I think a better question though is "for 2d shmup gameplay, is there anything of importance you might need a custom engine for?". It seems to me that it shouldn't be too hard to optimize a sprite based shmup on most engines, which is really all you'd need at the end of the day? (unless you wanted co op or online leaderboards or something I guess, and of course portability and licencing and other such issues)

Squire, if you want to make a shmup, just get Gamemaker Studio. It has everything and more that you will ever need for making a sprite-based shmup, it's simple, and it's fast as hell. Thanks to my recent projects, I'm starting to be pretty confident in my gamemaker.
Wanna team up actually? I've got a shmup half-done already and I'm willing to share it with you. The development froze because I lazied and lost interest in the genre during that season.

Yeah, I was actually thinking I'd buy Game Maker and start fooling around with GML and GMOSSE over the winter. I've been taking C++ and Game Design classes lately, the latter of which involved Unity. I actually had quite an enjoyable time learning the engine and building a demo for an arcade action platformer. I was thinking I'd continue to learn it and work on small projects every week in order to build some minor experience, both with scripting and game design.

I'd actually love to team up, though sadly having not touched GM (yet), I don't know anything about the language and would need some time to catch up on it. I would love to later though if I start to get comfortable with it.

On a side note, I am actually working on a computer science major in school. So I'll probably eventually have the know how to build something as complex as an engine in the long run. However, by that time I fear I may already be employed and have no time to consistently work on something as dauntingly complex and time consuming (as rfeese pointed out) as a game engine, even a 2d one. So I could end up going either way in the long run.

[BPzeBanshee]

If you're getting savvy with C++ then GML itself shouldn't be too difficult to get the hang of.

I actually found ActionScript 3 (for Flash) not all that bad for my Warbird school project....once we got a teacher that stayed consistent for more than 2 lessons instead of moving the goalposts, anyway. In fact the GM conversion from the original Flash Warbird took about 10 minutes tops because the way I coded the enemy spawn code and associated custom scripts in a switch-case-break allowed me to literally CTRL-C/CTRL-V considerable portions of the code straight into GM8.

As to the subject question, aside from avoiding preconceived ideas from programmer elitists it does seem like the sole benefit is knowing what you've made back to front, thus being able to (in theory) debug more easily. Personally I see no point in reinventing the wheel right now but it is something I'm considering getting around to doing one day.

[BryanM]

Performance should be really, really low on the list of reasons to make your own engine. There's only one primary reason: control. Control over how your assets can be built. How good is Game Maker at allowing you to make random levels? Procedurally generated space ships?

It's just easier to slop together something that makes DirectX actually usable over a couple weeks than to dick around with something you can't ever be sure will meet your needs for forever and ever.

[nasty_wolverine]

if you have time and patience, build an engine.
if you dont, use something of the shelf.

[n0rtygames]

What are the actual advantages of building your own engine?

Expanding your knowledge and having total control/responsibility over your codebase.

That's it.