GoStyle Web App

Upload your Go game records and get a list of professionals with similar playing style, relevant book recommendations and typical bad shapes and other aspects of your play you should concentrate on to become stronger.

The games are processed using sophisticated machine-learning algorithms, please have a look at the research section of our homepage if you want to learn more.

Example Output

You can have look at the following example results, for the following KGS players:

pachi — the Go playing bot
bronislav — me :-)

Privacy

By submitting any data, you agree that we can make further use of it. All the data we publish (here or in any future papers) will be anonymized, which means that no information about any particular user will be published, or ever given to a third party.

Lets try it then!

Upload an archive (zip, rar, tar, bzip2, gzip, ..., archive size limit is 0.5 MB) with .sgf games of one player.
Around 40 games (on 19 × 19 goban, no handicap stones) should suffice:
The sgf files should have the name field (PW, PB) filled (in most cases, this is done). The player of interest is considered to be the one whose name appears most times. We only use at most 40 games from the archive, you can upload slightly more, in case some of them were not readable, had handicap stones, or were skipped for whatever reason.

Or, specify your KGS nickname and we will download some of your latest games for you:
This field is ignored, since you chose to upload an archive.
Be patient, fetching the KGS archives might take some time since we want to be nice about KGS's quota.

You should also have a HTML5 browser such as Google Chrome, Opera or Mozilla Firefox

News

Have been moved to the GoStyle homepage.