This is a list of issues that stakeholders may need to understand in order to make decisions about the suitability of this scheme or to contribute to its development. They are not all, necessarily, problems.

Record Issues

1. The aliases file is presumably both incomplete and inaccurate. It would be appreciated if knowledgeable players were to assist with this. My thanks to those who have.
2. Records are incomplete after 1994 and non-existent before that date. If anyone has records from tournaments, especially recent ones, that had non-handicap games in them and were never passed on to Neville Smythe then I'd be grateful to see them.

Mathematical Issues

1. As is true for basically every rating system, there's one degree of freedom left undetermined. In this case, only ratios between the s_i are determined. This might be addressed by anchoring AGA ratings to those of some other rating scheme, using players in common.
2. The justification for the discounting rate is unrigorous and somewhat circular.
3. The justification for the rank interpretation requires more thought.
4. No allowance is made for players improving. Suppose you play someone when they are a beginner, then they get better faster than everyone else and beat some strong opponents. Your rank will be higher than your strength would normally merit. Discounting for years since the game and the restriction to dan-level players should mitigate this.

Teleological Issues

1. Forfeits and byes are ignored: a player is not penalised for a forfeit. The program is trying to estimate a player's strength to the best of its ability, not reward or punish players for positive or negative behaviour.
2. Handicap games and kyu players generally are ignored. Even games from handicap tournaments between (equally-ranked) dan players have been included.
3. The rating determined is not affected by the uncertainty. So a player may have a very high or low rating which is the result of a few lucky wins or unlucky losses.

Software Issues

1. The software is beta test and needs verification.
2. The software runs on Microsoft Windows only.

Organisational Issues

1. The Australian Go Association has not authorised this scheme.
2. The scheme and associated software lack a witty name, though I'm considering “Gratis” for “go rating software”.

I welcome feedback at David.Bofinger@dsto.defenceSpamProofing.gov.au (delete the spamproofing). This page is hosted by GeoCities, in return for carrying their advertising they will give you a free home page much like mine. Everything on this site varies without notice, especially after I get feedback.