As you may have noticed in the meanwhile Christer has started to integrate my extended dice stats code to the site. This post is for those interrested in the current status and near future of the dice stats.

Currently this has two benefits, firstly the dice stats are available for more (hopefully all) games again and secondly some more data is no visible on the match report page.

Visualization is still work in progess and is done solely by Christer. So we will hopefully see more stats appear over the next days/weeks.

So far I am aware of a bug that causes Blitz! turns not to show up in the game timeline. Also I have been approached with a few additional feature requests. Most of those are already included in the current code but not visible yet. A few though would require more code.

I decided to suspend my work on my other hobby projects in favor of this one. The prime objective is of course to get the bug wiped out and then see what aditional features would indeed make sense and could be implemented.

Unfortunately I will have to do quite some other work first. When I first decided to work on the dice stats tool it was more a fun project to see if it was possible and I had to dig into the client code to see how the replayer works and then plug my one code into that logic.

The result was a working but really messy peace of code with a clumsy architecure and no tests. The turn over detection code added in the extended stats at least received a rather extensive test suite but still the code itself is kind ugly.

I spent several hours if the last days to find the reason for the BLITZ! bug and I think I found the problem. But fixing it would require some restructuring of the code and this could very well break existing functionality.

In order to avoid that I decided that I will first put effort into creating regressional integration tests to ensure that future changes will only change what I actually want to change. Additionally I plan to spend time on refactoring the code into smaller chunks and if possible decoupling from the client code so that future updates of the client can be adopted faster and with less risk.

Some of you might say "he should have done that in the first place" but that is how it goes with hobby projects. Honestly I did not expect people to enjoy the feature that much, so I just tried to get it out in the first place.

I know it might be a disappointment for many of you that the bugfix will not come anytime soon, but I hope you can understand that this step is necessary before I can move on.