Halo 4 Technical Publications

Games Other Halo
1

One thing I haven’t been able to locate in regards to Halo 4, is any technical publications or presentations written by 343 developers/artists. Bungie didn’t produce massive quantities of these type of publications, but they were always available: Bungie Publications

The only in-depth articles I’ve read were about how “343 hired people who hated halo” (Gamasutra Article), and there was a great art production article describing the light mapping and compression technique used in Halo 4 (Unable to locate the original).

I also found that the Havok Physics, AI, and Animation Engines were all licensed for use in Halo 4 (Havok Games: Halo 4).

Back to my original question, where are the 343 publications concerning Halo 4? As a multiplayer game engineer, I find that articles or presentations that highlight a specific game mechanic in technical way can not only be educational, but also give me a sense of respect towards what was put into the game. There is a lot of talk about how Halo 4 has made a “turn for the worse” and “Halo is dead,” and I think that if there is any merit behind the mechanics added/removed in Halo 4, they should be documented and published (at least at a mid-high level). Just my opinion of course.

It’s documented that Bungie spent millions testing multiplayer (with real players from all corners of the world via open beta) and collected data over 4 installments of the series. From playing Halo 4, it seems that the entire multiplayer core was re-written to something that performed much worse than it’s predecessors. Not to mention, there are noticeable issues with host selection and join-in-progress seems mostly broken to me. Of course, none of it is explained, at any level. The visual aspects of the Halo engine; however, have increased tremendously, and being that the hardware is pretty much the same, that’s quite an impressive accomplishment.

Microsoft has never been that good at being “open” in terms of their software, and while in some sense, I can understand being closed. However, being open doesn’t hurt the success of a title or series (look at what Steam has done over the last few years).

If someone has more sources to technical articles concerning Halo 4, please let me know.

2

I loved seeing bungie’s AI demonstration and the articles about their philosophy on game AI. I always thought the first Halo had the best AI. The video showing how it worked was amazing. Idk what happened to the AI in the later games though. Maybe I am just blinded by nostalgia though.

3

I’m sure there are many people interested in reading technical papers on Halo. It’s a fascinating game and as a System Architect I’ve also looked for in-depth material on the game without much success. Some information regarding game design can be gleaned from Microsoft/343’s job recruiting site, but a 343 authored tech publication would fill in a lot of questions. And it’d be super interesting!

4

343i has a history of being closed and vague about… well, everything. I too would enjoy technical publications concerning Halo 4, but that’s probably too much to expect of 343i. The only behind-the-scenes we’ve experienced is the 30 second montage of Halo 4 development bugs released last week.

5

Corrinne Yu has given some Siggraph talks. I am not sure they are publicly available, however.

6

I was able to find this post concerning host selection. Very constructive post with some great suggestions:
Halo Waypoint Forums: Host Selection Improvements

That forum posts also links to two Microsoft research papers:
Matchmaking for Online Games and Other Latency-Sensitive P2P Systems
Measurement and Estimation of Network QoS Among Peer Xbox 360 Game Players

I also found this while digging through the research archives. Not sure it applies to Halo 4, but it might: True Skill

Note that the first paper appears to have been written in 2009, so it may not even be relevant. However, I would agree with the OP in the forum link that it seems that host selection is sadly based on nothing more than a geographical position estimated by each players IP. What’s interesting is that the research suggests that probing takes too long and players have no patience. However, it seems like Halo 4’s matchmaking is the slowest of the last three installments (Halo 3 and Reach). This, may just be my experience due to my “poor” geographic location :slight_smile:

Here’s a quote that really irritates me (from the first research paper, section 5.5):

> So far, our analyses have assumed that matchmaking relies
> solely on prediction for latency estimation, i.e., that there is no time
> to perform network probing to determine latency. We are primarily
> motivated by that scenario for game matchmaking where users
> are very averse to waiting, or for games where it is prohibitive to
> probe all potential traffic paths because they have an all-to-all rather
> than client/server communication pattern. Nonetheless, there may
> be other application scenarios where some network probing can be
> used to supplement latency prediction.

Ok, so probing isn’t helpful and error prone (based on the entire paper except this one paragraph). Oh, but you can use it to supplement the suggested algorithm? Maybe I am misunderstanding.

It also seems that they’re proposing this implementation as some kind of “catch all” solution. That may explain the lack of matchmaking filters in Halo 4, and perhaps some of the problems with join in progress (that’s just an oblivious guess).

Either way, just my meaningless 2 cents. I’ll probably revisit some of these papers when I have more time to review them in depth. If I find anything else interesting, I’ll post here.