When we last talked, we had launched our Kickstarter beta in July & August. Things went great. Player training was awesome. The ability to create teams, buy plays and ordering your roster went great. We are super stoked with where we left the beta.
Things weren’t all peachy though. We ran into issues with game playback. For those that played the beta you probably know already, but we ran into issues with our matchup replays. Each time a game would be simmed, there would be a slightly different result.
In a nutshell, we created two different versions of the engine: the standard version, which you used to watch game replays, and a simulator which had no art, no overhead of any sort. Each version of the engine would write all AI movements to a text log which would exchange between the two engines. We created a game seed system to guarantee symmetry between the two based on the text logs. Lastly, we created a time scale system (TS) to increase playback speed for use with the simulator engine. The TS was set up with increments of TS0-TS99, with TS99 running at 1000% of TS0. We could simulate games in about 20 seconds with 42kb of overhead at TS99. YAY!
The issue we ran into had to do with Unity’s ability to process collisions with fast moving objects. The faster the object, the more inaccurate it becomes.
Collisions = # of total collisions
End Result = score
As you can see, as the TS increases the more inaccurate the results became. It boils down to how Unity prioritizes CPU processing for the sake of performance. We tried changing the way the colliders were detected. We tried using raycasting for collision detection. Alas, if we crank up to TS99, we get wildly different results than TS0. With 11 players vs 11 players and their fast-flying collisions all happening at the same time, there are A LOT of collisions at any given moment.
The issue is this. The simulator engine would simulate a game at TS99 to its end result. If you then wanted to watch the replay of your game, you could possible get a different result because you’re watching your playback at TS0. Ultimately, the outlook of what we want to do, a “Tecmo” Com vs Com team management w/ a live P2P PVP system, is not possible with the current development tools available to us.
BUT WAIT, THERE’S MORE!
There is light at the end of the tunnel. You are going to call us crazy, but this is the actual plan.
The current Com vs Com Facebook game is going to become its own game – Gridiron Heroes: Coach. GH:C will be a Com vs Com, team management, asynchronous, 1-player experience with touches of multiplayer here and there. For those that are fans of the genre, it will be similar to Top Eleven. This enables us to continue down the path to finish what we’ve started. There will be leagues to advance to, free agents to pickup, players to train and playbooks to be created just like we showed you.
This is the crazy part. We are going to reboot our initial build from 2012. You, the fans, probably remember this. The field view is from the original NES. For those that want to play Tecmo online, this will be your go to. We will offer many, many teams for you to choose from and you can take on your friends…online…..in PVP mode. The results won’t be stored, but hey, you can wager beers or baseball cards and play your friends across the world. Pretty cool, eh? Upon alpha release, there will not be a management piece. This will purely be to play single games.
….It gets better.
We are partnering with TecmoBowl.org with our PVP build. We want to make this as accurate to the original as possible and who better to supply insight than the guys at TecmoBowl.org? Partnering with TecmoBowl.org helps the community have an accurate representation of Tecmo, but with the ability to use it on modern platforms. Our builds are using the Unity3D engine being written in C#, meaning, we will have Webplayer (browser) & PC builds immediately upon completion. If the game continues to gain ground, we can port to other platforms or modern consoles somewhat easily.
I’d like to tell you our endgame with both of these builds, but let’s start with the baby steps and then we can talk big picture stuff later. First up, GET A BUILD IN YOUR HANDS!!!!1
Dave (davefmurray), Casey, James, Dave B (bruddog), Victor & 1 Secret Weapon
P.S. Here is a quick teaser of the PVP build’s progress.
And a super zoomed in screenshot!
P.S.S. We are contemplating a move to Visual Studio and will be applying to Bizspark in hopes of getting approval. Visual Studio 2013 seems like where we need to be.