The Electra Engine tool is live for use in building characters for the most advanced conversation system in game history. The core achievement is pairing the complexity of a 3D flow chart system with a simulation view that shows the output in a text message format. This allows writers to approach the conversational choices on both a micro and macro level.
So why is it the most advanced conversation system in a game? It starts with the mechanics from the player’s point of view. When they go to talk to someone or an event triggers a conversation they have multiple choices of what to say like many games, but what is truly groundbreaking is a nested emotional collection of options. Basically the player can select an option that reads ‘hey, do you want to go out sometime’, but then before locking it in they can explode the dialog to reveal what the emotional tone of the option is. Perhaps in this case it would be a combination of friendly and timid. Now when the player explodes this option, it’s not merely a passive observation mode, but actually reveals up to 7 other emotional options each with a combination of tone traits that each person reacts to depending on their personality.
This may all seem to be getting overly complex. After all why would you want to have so many variables going on behind the scenes? Well players live and breathe consistency and rules that they can trust, and more importantly depend on to make strategic decisions and be rewarded for their skill.
What happens next is the players tone is taken into account by the Ai’s personality which combines that factor with it’s ongoing BOND (like or dislike for the player) and arrives at a REACTION that uses a few more aspects of the tone to figure out how positively or negatively she will respond.
Now this may sound like a lot of work for a writer, but actually the system doesn’t just offer organization through the nested depth in the 3D flow chart system. It also has built in functionality to allow for as simple numbers of options as the writer wants per step or Conversation Chunk (1+ player dialog, 1+ ai reply). It communicates the options with less depth as gut instinct warnings so that any combination of writing styles or ebb and flow depending on the importance of the conversation step.
The visual components of the Ai’s body language will help show the positive or negative reactions to the player’s choices are not connected to the Electra tool at the moment, as much of the character art and animation are being reworked, but the entire Conversation system is working and this is an exciting time in which writers can finally get their hands on this revolutionary game creation tool for Forbidden Technique.
One of the last big changes to Electra before Beta Release is the addition of the Premise Graphic within the galaxy view of the conversation system. A core part of Electra is the dynamic step by step calculation of the Ai’s degree of response on a scale of 1-10. But building a dynamic branching conversation tree quickly becomes hard to keep track of in a what it will be like for the player sense after awhile. So the text chat window was made a while ago to simulate how the conversation would go if you choose specific options as a player. Although over time more factors were added to the Ai’s response besides just the tone that the player chooses. Which is accurate to real life, that saying things in the best way from the girl’s perspective does not guarantee a way into her heart.
In fact the discussion of all the factors that play into courting a girl is a long one that can be seen on the team’s youtube channel’s MAGFEST panel: https://youtu.be/-YJI7-VAzrc
So that’s where the Premise comes in. The premise is what is the relationship between the player and the person they are talking to that impacts the calculation of how the person will respond to the player. The 2 exposed parameters in a dynamic sense are Bond and Spark. These parameters specifically change from conversation to conversation and change a small amount during each step within a conversation.
Bond is the friendliness the girl feels towards you. Essentially Positive To Negative with Neutral in the middle, specifically platonic.
Spark is the attraction the girl feels to you which is very importantly separate from Bond.
Now how these properties are changed (improved or harmed) is based on each Girl’s individual Personality. Now the girl’s personality stays the same the entire game which acts as a more solid foundation for understanding what each girl will like or dislike. Personality consists of 3 separate properties. But more on that later!
Forbidden Technique is a Dating game mixed with magic powers combat in VR. The game is build on Bad Plan games’ Electra engine which brings skillful and fun depth to conversation and dialog choice. Progressing in your relationships with girls at school or girls in the spirit world helps you grow your powers to save the world from the collapse of the spirit world into the normal world. Mechanics include dialog choice and tone options within all dialog choices, checking out girls and the risks that come with it, and using your powers to defeat monsters and gain more xp. Follows the story of a normal high schooler who gets caught up in the spirit world and powers, goes to school by day and visits the spirit realm by night.
The Electra engine is a level editor but for dynamic conversations (advanced kind of branching dialog tree) that the game reads as a mod file that will be open to any user to create their own female characters to interact with and try to date in the game both at school and in the spirit world.
Good news! Forbidden Technique has made another massive leap forward. The clothing, personality, and body type randomization system to generate a school full of unique people is in it’s first working stage.
Yes, I don’t currently have a belly mesh so there’s a hole in the middle of them, but otherwise I can generate a massive variety of shirt colors and pattern combinations. At this point the randomization is pretty random. I choose a bunch of colors to tint and shade different patterns and it picks several with no real color pallet.
So yeah they walk to class sit down in their seats, then after the class ends get back up and walk to their next class till the day is over. These generated student files are saved each day and so persistent relationships are trackable with each one as well.
The body randomization is not immediately the most visible because textures don’t change and neither do normal maps, only the physical mesh shape and size. Unfortunately it will likely stay at this stage of variety for a short time while the rest of the mechanics are all combined into a truly playable demo.
A big part of me feels like I’m still goofing around when it comes to Forbidden Technique. I’ve got so many half baked features, all of them with testable promise for myself, but not ready for demoing. I need to get other people in the action for feedback as soon as possible, but I’m caught trying to run tech demos of everything at once. Parts of which are good for long term planning and re-design of interlocking parts.
Other parts of it makes progress stagnant and daunting. It’s kind of like short term memorization because so many of the features are so complex. When I was setting up a system for selecting magic powers I actually had to completely scrap it all and do it again after sketching some flow charts because true organization was needed. But even with a handy flow chart to look at, it takes about an hour or 2 when I haven’t looked at the powers code for more than a day before I’m making real progress.
This turns out to make jumping back into the half built features a little more painful and slow than it needs to be. I just need to commit to a specific system and follow it through till it bumps against other missing features with nothing else to be done. So far I’ve got a power selection system, a complex VR movement system for the player, a power channeling system, an Ai generator that also generates class schedules for each student and where they need to head depending on a game time system. There is also the beginnings of the voyeurism system where the red debug boxes triggers experience gain which is boosted by many many factors, and some of those factors then contribute to a detection system which is not very good at the moment. Not very good meaning I always get caught staring at the boobs.
I think the priority I need to go with is the XP gaining system AKA Boob Watching. It’s really the core mechanic of the game that is completely different from all other games. Not only is getting it right important, but it’s the feature people need to play to know what the game is about and whether they like it. The powers are cool, the story is cool, the open world is cool, the vr player movement is cool, but those are just coincidentally awesome features. Those parts give meaning to the stealth peeping, but the farming mechanic is the definition of the game. All other features should be inspired by the core voyeurism feature and that means finishing it first complete with consequences for female disapproval and awareness mixed with social standings etc etc as it gets more refined.
Tons of games have path tracking, no one cares about the work going into path tracking. I must focus on SXP (sexy experience points).
A couple weeks into brainstorming and I’m excited enough about the project to start running tests. The core experience gaining mechanic is voyeurism. Essentially gain powers by looking at the hot girl in class without getting caught. Higher risk, higher reward for getting in close, but still evading suspicion.
The tricky part is that this is a stealth system in plain sight. Traditionally in games you have any guard or NPC catching you in line of sight triggering an alarm within some time, but this mechanic is more subtle and thus more complex to get to feel right.
Line of sight is important and the angle difference from where the target girl is looking to looking at you is one of the most important factors. But it’s a risk factor, not a cause by itself of being ‘Caught’. Being caught includes the girl facing your direction somewhat and observing you looking south of the face for a segment of time based on your distance from her and the accuracy (obviousness) of your gaze.
Adding on there’s the social implications/influence at play. Does the girl know you well? If not she’ll be more suspicious. Did you engage her in a conversation first? If not, why are you staring at her!! It’s a fun mechanic to play around with which is this sort of awkward meta stealth.