Hey, sorry I didn't get to it these days. I'll take a look soon, perhaps tomorrow. Sorry for the wait.

Thank you for pointing this out DeXVinogradov, I will fix it in the coming days.

Hello everyone!

Today I found this amazing shader that can generate planet textures!

The only problem is that the shader is applied for the entire size of the layout, meaning you can only have one big planet in the center of the screen. Is there a wizard among you who could look at the effect's fx file and somehow contain the shader to the size of the box it is being applied to?

I'd like to have multiple planets and use them like normal sprites.

So, is it possible by now to actually use this to generate positioned/scaled planets? If not, does anyone know other ways to achieve this?

Hey Sup with that? , thanks! Now I feel silly for not seeing it. Cheers!

Hi guys,

I have a game in which I select objects by clicking on them (more specifically, touching them).

I need an event which deselects all my selected objects when I touch the background, but not when I touch one of the objects.

I though a simple "On tap gesture" would suffice on the background, since the tap wouldn't go through the objects, but I guess I was wrong. Is there a way to make objects "unpassable" for the click/touch ?

Thanks!

Hello everyone,

About a year ago I spent several weeks working on a Construct 2 remake of the wonderful classic Sword of the Samurai by MicroProse.

Parts of the project are in a fairly advanced state now, but other parts still need a lot of work. Since I don't always have the necessary time and for other reasons, I decided to post the entire project and assets on GitHub and open source it for you. This way, if any one of you feels compelled to play around with and/or add to the project, parts of it might someday be finished, yet.

Right now, the game is comprised of 4 projects:

1. The Meta-Game, which includes the (handcrafted and very proud of it) map of Japan, the World Map and the entire family/rivals/story/quest logic and implementation. This one is in the most advanced development state - the logic is nearly finished as it is. The projects needs balancing, in parts visual overhauls and bug fixing. 2. The Battlefield Game, which allows you to control units of independent soldiers of different classes, in a Total War, 2D top-down style (see screenshot). The mechanics here are fairly advanced as well, with the positioning logic being as good as done. It requires samurai sprites for the units (which are already partially included in the Castle/Rescue Game mentioned below - I had someone do them for this project specifically). It also needs considerable fine tuning and balancing, better fight-logic, different starting scenarios etc. This part of the project also has the biggest potential for further improvement, like adding terrain as a factor for speed and strength, possibilities to hide in trees/fields etc. I had originally intended to make this battle engine a stand-alone game in itself - it still has the potential for it. 3. Then there is a Duel Game, which uses Spriter and features a side-view duel. This requires lots of loving all around. 4. And finally, a Castle/Rescue Game which uses a dungeon simulator to create a different castle each time, from which you can save the princess/kill your rival/ etc. The castle generation, some logic and some AI is done, but I'd say this one needs the most work, still.

These are, as of now, separate Construct 2 projects that would need to be brought together eventually but should absolutely be enjoyable by themselves. Here are some old posts I had written about the project.

The code in the Construct 2 projects is commented extensively. Still, due to the complexity of the whole thing, especially the meta-game, you might still need some time to get into it, but if you have any questions please feel free to gather them and ask. I'll try to answer as fast as time allows.

(P.S. You will require several plugins, depending on the project you're looking into. If you have problems locating any of them, please ask and I'll look for them).

(P.P.S. If you use the assets/components of these projects in anything other than this game, I would greatly appreciate it if you mentioned the source and left a comment here)

(P.P.P.S. In case you wonder why sometimes it says "Life of the Samurai" in GitHub, that was my working title)

I haven't, but I think it's not outside the realm of the possible.

However, if you want to look into this beyond game-AI-behaviour, I'd recommend you switch to either the JS version (or another version) of Deep-Q-Learning (here) which you can have running at thousands of ticks per second (instead of 60fps) or focus on some machine learning library, for example by using Tensor Flow.

After 20 minutes of "training" the output is still completely random.. Am I doing it wrong?

Hi Jan. At first glance what you're doing looks correct to me. In my experience, it usually takes around 200.000 ticks (e.g. 3 hours) for the agent to begin behaving intelligently, e.g. to see it converge. This is a downside of Construct 2 - it only allows actions each tick, so you can't accelerate training beyond that.

Also, try setting Action:Learning to turn off the learning process after the training period - this way the agent won't undertake any more random actions and you get a clearer picture of the outputs which are correlated to your inputs.

Hi guys,

so sorry I never got to make a more detailed documentation. I just can't find the time at the moment, I'll try to fit it in sometime.

Until then. here is the capx file and here is the documentation.

If you have specific questions just ask here and I will gladly help.

I'll try to add some tips and hints on the weekend, but Rex is definitely right.

One should definitely invest some time and read up on reinforcement learning - what a neural network is, what inputs and outputs are, what forward and backward propagation are etc, since it's quite the broad topic and at first perhaps not too simple to grasp.

There is a plethora of great sources out there at the moment, as Deep Learning becomes a valuable tool for companies' data analysis. Start with some Wikipedia (like https://en.wikipedia.org/wiki/Reinforcement_learning and https://en.wikipedia.org/wiki/Q-learning) and also google around a bit.

I admit the CAPX is a bit bloated, since there's quite a lot going on, but it's all broken down in only 4 steps (Train, triggerAction, Reward and manageEnvironment):

TRAIN:

The agent gets 9 sensors, and each sensor generates three inputs for the brain - "apple", "poison" and "wall", using the distance to the touched object as value.

(In this step, I also draw some lines from the agent to the touched object using the SensedApple, SensedPoison Tiles, but these are purely optional, I guess I could take them out alltogether).

REWARD:

Then there's the reward built upon interaction with the apples, poison or the walls.

TRIGGER ACTION:

And finally there's the output of the brain (for example "right") transformed into an action ("set angle of motion to current angle +50 degrees").

MANAGE ENVIRONMENT:

This just adds more apples and poison if there's not enough laying around.

If you still find this confusing I'll take a look at it in the weekend and trim it down. Although I usually consider more to be better .