@JohnnyBlack BGE does not like, do not use.
BGE is great, I like it.
To everyone talking about the BGE not being up to par when it comes to performance, I decided to do a simply benchmark to see what the speed impact is when you switch the ‘V-synch’ mode from ‘on’ to ‘adaptive’ (in current builds). the setup is six highly subsurfed solids (that are rotating), and two sunlamps with VSM shadows. There really isn’t much in the way of logic going here because V-synch mainly concerns graphics. Formatting and snipping has been done so as to fit the information from both tests into a single image.
Now I thought that we might get a little more speed out of the rendering portion of the engine, but the total improvement indicated to me that a lot of people here are going to like this.
I’ll let the image do the talking.
From 30 FPS with initial performance spikes to 50 FPS with no spikes or slowdowns. Very impressive to say the least, so impressive in fact that perhaps the V-synch mode should be set to adaptive by default. It may not be enough to convince some to continue their projects in the BGE (citing raw graphics features on top of performance, but it should lessen the chance of serious slowdowns).
The biggest question I have to ask is why there are so few good BGE games but plenty of good renders and animations. Naturally Blender attracts a certain type of casual user, since it has no barriers to entry, and so there will always be a disproportionate amount of newbie work around. That said, the games area seems to contain almost exclusively newbie work. My guess is that the BGE is not as well-known in dev circles as the renderer is in CGI circles; or worse, that it has a reputation as a “noob engine” due to the aforementioned newbie projects.
And of course the only way to get past this is to stop complaining and make some good games.
It’s also worth noting that there are lots of crappy Unity game projects out there as well, for the same reason. But Unity has its own PR/marketing department that can spackle over those. So in the end we need two things: good games, and advocates.
Because the amount of work to create a good game is orders of magnitude above what it takes to make a good render or a short animation. Most of the games that people would classify as “good games” have 3-6 full time programmers and 10-30 full time artists. You either get artists that can make assets and wire them together with logic bricks or developers that might not produce the best graphics.
Minecraft has to be one of the most popular and successful indie games of all time and its graphics are worse than most of the BGE game challenge projects. The main reason they are terrible is not because they have to be, but because they managed to convince people that the low-rez cubes were a stylistic choice. It was total genius because it must have saved countless development dollars on art assets.
I personally really likes so many game engines that available, whether it is open source or propietary. BGE is one of those. But i completely agree with those quotes. And in addition, allow me quote from Ton’s twitter yesterday:
“We do care about BGE, but generic tools have priority now. The future might bring back a lot of BGE work too.”
So, let’s get back to work on our games, shall we? 
@vknw123 - JohnnyBlack didn’t say that at all. His first post said that he liked the engine.
People seem to think Minecraft has poor graphics when it is indeed a matter of preference. Some people love the graphics. Some don’t. It doesn’t have bad graphics absolutely.
- WALL OF TEXT AHEAD -
I agree with PlantPerson and Goran. A good game may take longer than a good render or animation, but that’s not to say that a good render or animation is a one or two day thing. They can take months or years, just like games. It’s about dedication, and being willing to learn the things you don’t know to pursue your project, or make a project within your means, and polish it, regardless of the tools.
People have been doing the best with what they had for years. People have been making full 3D engines in Megazeux, and making entire life simulation games in a console window. The tools don’t really matter that much; it’s about how you use those tools. How you get what you want across as effectively and efficiently as possible with what you have.
On the topic of the BGE particularly, if you want to make any kind of game, then you should probably learn how to script, as sdfgeoff put it. You can make a game with logic bricks, but you can’t really polish it without using Python. Fortunately, it’s not hard, really; Python is on the easier end of the programming languages out there.
If you don’t know how to do something in the BGE, just hop on the (fairly active) forums, and ask how to do it. Ask why something’s not working correctly. Ask how something should be / is done “in the real world”, not for a logic brick setup, or a pre-made script. Convert that concept to the BGE as best as possible. If you just jump ship when you face a problem because the BGE’s not pre-made to do something, you’re going to face the same issues in whatever engine you choose. You’ll run into a problem in which something’s not working correctly, when it’s really that you’re not doing something right, and you won’t want to stick around to trouble-shoot it.
I said this awhile ago, but if you want the BGE to get better, use it to make games. This will make the engine more popular, and give it a visible “pedigree”. They don’t have to be excellent, block-buster games. Just polished, completed games. Not “BlenderArtists Finished Games Forum” finished games - “Distributing Games on your own site as executables in zip files” finished games. “Participating in non-BGE game jam” finished games - real games. Games that make people say, “Have you considered working in Unity?”, to which you can blow their minds when you reply, “Why?”, not because Unity’s bad, but because you’re happy with the tools you’ve been given. You should be happy not because the tools are perfect, but because they’re perfect for you.
P.S. Thanks to those who praised (?) me, but I’m not a great BGE user. I’m just fairly experienced with the engine, and up until recently, had a lot of time on my hands. If you’re newer to the engine, or just think I’m at some sort of “pinnacle” level of capability, just keep working with it and trying interesting projects and tests.
I’ve made “noob” posts on the forums. I haven’t known what I was talking about on my posts before. I’ve made little test blend files, trying out little techniques and things. I’ve made tons of game projects that never panned out, in engines all around the Net, haha. I’ve abandoned (never started?) projects in Game Maker, BGE, jMonkeyEngine, C++ / SDL, Python / PyGame, ActionScript / FlashPunk, etc. If I moved to Unity or another engine, my inability to finish games would continue; it wouldn’t change. That’s a problem I have to work through. The BGE is not really what limits me.
I am just a user like everyone else, except now, I pretty much know how to do what I want to do in the BGE, though I’m still learning. If you want to get good with anything, keep practicing. Making games is better than just practicing something, though, because you get results. You can have a finished product, or a working toy, or something to just mess around with in your spare time.
BGE is more than capable of making good games, chasing realism and the highest fidelity of graphical experiences is rather silly and anyone trying would have to be willing to spend a lot of time and effort into it, games aren’t a life simulator with a twist. Tight mechanics, a good story and a carefully laid out pace, any two combination of these makes for a good game no matter what it looks like. And the BGE lets you do that, it’s a game engine with a GUI. Trying to make a game from scratch with open source material would take you ages to even get to the GUI stage with a reasonable performance.
Like Goran said, BGE is just a tool among many that do, and it offers so much for being open source, it leaves so much to customization that you can do anything you want with it. You direct access through the API and the shader programs, input and output, a profiler and an easy to understand debugger, even the modeling/animation suite is right there in the software, come on now.
It’s up to you to figure out what game you want to make and just make it happen with programming, and the cool thing about BGE is there is always a way to make it happen, heck you can even get down to the C++ if you’re not comfortable with Python and make it work, logic bricks are useful and well optimized but they’re nowhere near enough, and if you neither like nor know how to code and program then honestly you shouldn’t be trying to make games, period.
The very first time I finished my model and got it to move around, interact with the world was a really unforgettable moment, I was hooked ever since. I can’t live with myself unless I finish my game exactly how I want it, no matter what, there are so many concepts, papers and just ideas out there for you to explore, a game after all is made with a loop of geometry data, an update each frame and a render/draw on the screen and BGE lets you do anything with each step so you can’t exactly blame the tool.
Sorry for the long post, I don’t usually do this as I’m kind of a lurking member of this community.
Jmonkey and the torque 3d engines are both open source. At least Jmonkey can publish without problem to android.
finally, someone who thinks like me.
Yeahhhhhhhhhhhhhh!!! 2014, i make my blender game
and do not make posts if you haven’t practiced,
but that’s what makes bge great, you can make any kind of game because you have access to the code.
In BGE, skill + persistence + knowledge + motivation + good plan&layout = kick ass game
You talk about graphics and then this, proudly blender!
Not absolutely perfect but that was blender 2.49 and we are in version 2.68, so anyone that says he can’t create really good graphics may have not searched well enough
Same with any other engine. Do you think because it’s not blender you suddenly can’t make the game you want to make. It’s best to choose the engine that is geared most to what you want to make that will have the features that you’ll need so you can focus on making a kick ass game and not on trying to extend the editor.
Just like me minus the pc spoilage part. I intent to make my own game in 2014 with blender
As sexy as that image is I think there is a little more to the gaming experience than poly count.
But blender can do what a normal game engine can do. Proof- Yo frankie
Kamiyama: I think Ace Dragon was trying to point out a) behind the scenes dev work continues and b) trying to counter one of JohnnyBlacks points about performance.
Smooth engine performance means more resources for other things.
@nathan: So why should I choose blender over another engine with more features that can do everything that the bge can?
Exporting time makes up for coding time if you are experienced.
It’s all about choice.