Early this week, we finished modeling seven test tracks – acceleration (a), turning (t), jumping(j), a+t, a+j, t+j, a+t+j, and the turning track for turning. We wrapped up a new demo with all the tracks before the client meeting on Wednesday. The new demo includes the features that one can select a chassis model, select a running track, run the track and collect the data, review the evaluation, customize the car with a few options.
On the design side, there was a big issue that we need to figure out – do we need to pre-generate all the chassis models? In another word, does the final deliverable have to be a standalone playable game? If the answer is yes, then we will have to compute all the chassis models by ourselves. The advantage of this plan is that we can deliver a game that is playable by itself and it is perfect for Autodesk to show their demo. But the disadvantage is that players won’t be able to get a model based on his own performance, since everything is computed by our own data.
Therefore, we brought the demo and questions to the meetings with Autodesk and Bandito Brothers on Wednesday. We went through a quick presentation talking about the progress. Then we received their feedback. The feedback mainly includes:
- Focus on the test tracks instead of the mission tracks, since test tracks are where data is collected and cars are trained. Bandito Brothers found test tracks super attractive and exciting.
- Focus on making the difference of feeling explicitly. An optimized car for turning should perform very differently than the one for jumping, and players should be able to tell the difference easily.
- Pre-computing models for one or two tracks (jumping or testing) is enough, as the purpose of the project for Autodesk is to show off the ability of Dreamcatcher.
After the meeting with Autodesk, we went to meet with Emotive, who is a stakeholder of this project. We talked about our game design and progress, and then discussed how we can use Emotive in this project with them. As the semester is closed to the end and Emotiv doesn’t have a plugin or SDK for Stingray game engine, we will probably use it for playtesting purpose. They agreed with our plan and expressed the willingness to seek for further collaboration with the ETC.
On Thursday, we deeply discussed about the feedback and our future plans. We will narrow down the scope to focus on only two levels – turning and jumping and the combination of the two. We will have 12 customization options for each level so totally 36 models. During the meeting they suggested us to schedule the time with Dreamcatcher team and use the web version of Dreamcatcher to generate massive amount of models in one run. But when we carefully thought about it, we realized that we were sacrificing the obviousness of difference of feeling for more customization options. Therefore, we may carefully consider if it is worthwhile.