2017 Update

The past six months have been some of the busiest ones at HearthSim, which is incidentally why updates on this blog have been sparse. I thought it was about time I gave an update on what we’ve been up to!

Since last September, we have released HSReplay.net Articles, where we have been directly posting updates on HSReplay.net as well as Hearthstone Meta / Replay analysis articles.

These articles, as well as other articles currently in the making, have all been guest posts by Baney, Dereck Toker and HearthVis. We have been blessed with very enthusiastic researchers of all backgrounds, wishing to do research based on HearthSim data. As a result of that, we have opened a mailing list for research projects. We want to extend our full support projects researching AI and HearthStone simulation. If you are in this situation, reach out to me on Discord and I will invite you!

HSTracker, a macOS port of Hearthstone Deck Tracker, has joined the HearthSim Developer Community and added support for HSReplay.net. It’s been growing steadily since, and @ifeherva contributed a very impressive C++ port of HearthMirror to use in the tracker, which we will soon switch over to in HDT as well.

Sunwell, our card renderer, has gotten a massive update. It’s been ported from JavaScript into TypeScript, has received a plethora of performance improvements, fixes and new features, and a standalone NodeJS card renderer is now also availale. Both Sunwell and HearthstoneJSON are now available in the NPM HearthSim org.

A new simulator, SabberStone, has joined our ranks and together with @Citiral and @BertP, @darkfriend77 has done incredible work on our server and simulation infrastructure.

We published decrunch, a Python library for interfacing with the crunch.h DXT compression library. This allows us to support crunched textures in UnityPack.

Speaking of UnityPack, we are currently in the process of rewriting it in C++. Although the design is solid, parsing is in need of a second iteration - and the need to interface with UnityPack in other languages has repeatedly come up. The next generation of UnityPack will therefore be a C-compatible library, with bindings in Python, C# and Swift.

We’ve been made an official Discord partner after opening a new Discord server for the greater HearthSim user community. This website (HearthSim.info), as well as the GitHub HearthSim organization, are now referred to as the HearthSim Developer Community, or just HearthSim Developers for short. This contrasts with the new community we have been cultivating, the HearthSim User Community or just HearthSim Community for short.

Our Discord servers are:

Of course, the HearthSim Developer Discord is still being mirrored on both IRC and Gitter, thanks to @42wim’s great work on Matterbridge.

The lovely HearthBot, our Discord bot for card search, has also been made available as a public bot. Send a PM to HearthBot#4400 on Discord with !help and/or !invite for instructions on how to invite it to your own server!

We also went to BlizzCon 2016, which was an incredible experience. We were invited to the Blizzard campus, met up with a lot of Team 5 people and had an amazing experience.

But of course, the main thing we’ve been working on is HearthSim Premium. Expenses have been really high since the initial release of the site; we are processing well over a million replays every day. The infrastructure behind this requires a business model and we think our Premium offering will attract players of all sorts and allow us to continue making the replay service available for free, as well as fund the development of all the tools we are supporting.

The Premium offer is part of a greater release of features on HSReplay.net which we are calling HSReplay.net Statistics. It is a set of statistical analytics which includes a database of cards and decks with graphs on popularity and winrates over time, Mulligan statistics, loads of dedicated graphs such as top Discover picks, top targets for cards, and much more. Premium offers some extra graphs and features, such as matchup filters, granular time ranges and personal card statistics.

We are aiming to launch publicly within the next couple of weeks. It’s been a very active half-year, I have barely touched the tip of the iceberg of cool stuff we worked on during that time.

I’m looking forward to the other half.