Maybe the site supporters should get to vote on the issues that they care most about.
3 Likes
Yeah, similar to user voice (is that still active?), but I meant to specifically build off of @Mulsiphix1’s idea, so I was envisioning a system where the site supporters in particular get that privilege.
1 Like
I am pretty sure it’s nicer to work with supporters than it would be to work for stakeholders. 
3 Likes
Surely it is.
The obstacle to there being more developers, and thereby reducing the Bus Factor Risk, is access to the back end code.
As I understand it, the obstacle to that is the amount of training required to bring in a new back end developer.
If this were to be converted from “pertinent conern” to “mitigating action”, @anoek would have to be persuaded that it’s worth the effort, and the person/persons chosen to ben invested in this way rather carefully…
GaJ
2 Likes
Is the concern here that somebody could wreak havoc? I’m not sure how Open Source projects prevent such situations from arising, but could making OGS open source be possible? I know you can still make money from going open source. It might be an easier way to draw in other developers and motivate programming capable users to add their own features.
2 Likes
Being open source does not mean that any code is accepted immediately. The project manager(s) still decide manually whether or not they accept the addition.
OGS frontend already is opensource. GitHub - online-go/online-go.com: Source code for the Online-Go.com web interface I don’t think there are currently plans to opensource the backend as well. Obviously there would be some advantages to going opensource, but obviously there would be disadvantages as well. Let’s probably leave the decision to the man who created the OGS code itself and surely knows more about it than we do 
3 Likes
Of course, of course 
. I’m just spit balling here. My intention is to see anoek get some help. Whether help is need or even desired or not, from anoek’s viewpoint, is another question entirely 
.
2 Likes
Just as a side note, integer komi is a wet dream of mine and if draws were properly handled, I’d create all my unranked game requests with integer komi and all my ranked games with NZ rules (I currently use AGA). This is the one issue that I want implemented the most.
1 Like
The point of allowing and supporting integer komi is to allow draws as a possible common outcome in the game. Since the typical practice is to avoid draws with non-integer komi, this does change the game in a major way, perhaps quite unfamiliar for many go players (myself included). Instead of (nearly) always having a decisive winner (even if having to be decided by the thinnest of margins), the possibility of draws allows for a third outcome that indicates that neither side was able to decisively outplay the other.
I think the difference in mentality between requiring a decisive winner vs allowing draws can also be seen when comparing some popular worldwide sports like soccer and chess (where draws are very common) vs the popular sports in the USA (i.e., American football, baseball, basketball, hockey, where draws are made at least quite rare to virtually non-existent).
In games where draws are possible, ideally we can hope that wins are earned by decisively and significantly outplaying your opponent, rather than just winning by default (since someone has to) for just slightly edging them out.
However, integer komi go still does have a pretty thin margin covering the possibility for a draw. Another interesting variation would be to consider go where a player has to win by a margin of X points (after factoring in komi), otherwise the game is declared a draw.
5 Likes
Thank you for further clarification @yebellz. I understand why some people could find this feature really useful, and by all means, it is a wonderful idea that others would like to see happen. That is what really matters. 
1 Like
I also very much like integer Komi to allow draws. IRL tournaments allow it (at least in Europe) but I don’t know how it’s treated in the rating algorithm. Hopefully not the same as both players losing.
3 Likes
As do I.
This is an interesting idea.It seems feasible that draws could potentially become commonplace, for players of ranking/skill level Dan/P Dan AND/OR with matches containing players of relatively equal skill. I’m curious how a lot of draws might affect the ranking system. I’ve tried looking for the answer myself:
- OGS has a new Glicko-2 based rating system!
- Glicko Rating System
- Example of the Glicko-2 system (excerpt, page 2)
We now want to update the rating of a player with (Glicko-2) rating µ, rating deviation
φ, and volatility σ. He plays against m opponents with ratings µ1; : : : ; µm, rating
deviations φ1; : : : ; φm. Let s1; : : : ; sm be the scores against each opponent (0 for a loss,
0.5 for a draw, and 1 for a win). The opponents’ volatilities are not relevant in the
calculations.
I assume this means that a draw would count as a half win for each player in a match that ends in a draw? But I have asked for clarification.
2 Likes
Maybe I should have added that I would be against the idea that a range of winning scores could be declared draws. My feeling about allowing draws via integer Komi is that it’s more natural. There is no way of scoring half a point in Go so introducing half a point feels artificial to me. And then if the score would otherwise be equal why should one side be declared the winner?
I would hope that draws would still be unusual. (Someone will doubtless be able to tell us what ratio of games were decided by half a point!) I think it’s one of the strengths of Go that draws are unlikely. But having the (slim?) possibility I think is good. A one point win or loss would be as exciting as a half point one currently is wouldn’t it?
2 Likes
No. In the event of a draw, and the stronger player wins, the weaker player will gain rating points, and the stronger player will lose rating points. If both players have the same rating, then both their ratings would remain unchanged.
You can see this in one of the links you gave. The change in rating is proportional to s-E where s is the result (1=win, 0.5=draw, 0=loss) ans E is the expectation value. The expectation value is 0.5 if both players have the same rating.
So if the expectation value is 0.75, then the stronger player would lose for a draw the same amount of rating points that they would have gained for a win.
4 Likes
I find the idea interesting, but I too would be against seeing it implemented on an official basis. It might be interesting if contained to specific scenarios, like a tournament. Playing with different settings or slight changes in rules is always fun in a tournament setting.
I wholeheartedly agree 
.
Based on how @opuss and @flovo (link) describe draw’s being treated when players have different ranks, I am not a fan of this either. It takes the notion of a tie and throws it out the window. Even in a draw, there is still clearly a winner. That doesn’t seem fair to me. I can see how a weaker player might be rewarded some points in a ranking system, to reflect that they were able to achieve a tie against a higher ranked player.
But for the stronger ranked player, it feels unsportsmanlike to punish them for a draw. I can wrap my head around why it is done inside the ranking system, but I personally wish that the stronger player’s points were not affected in this scenario.
I am dying to know this information. I wish OGS had a stats page that broke down different statistics for the platform. Something similar to the same treatment members get from the G0t Stats? Analystics Tool. I wonder who might even know such a thing 
.
Yes. Yes it would 
.
2 Likes
Generally speaking, ratings algorithms would treat a draw as a data point that supports the hypothesis that the two players are roughly equal in strength. On the other hand, a win is treated as a data point that supports the hypothesis that the winner is stronger than the loser. Each player’s rating is somehow calculated by taking many data points (game results) into consideration (while typically giving more weight to more recent games), with the actual ratings being estimated as those that best fit the data. The maths behind these calculations can be a bit complicated and varies significantly between different algorithms, but they all follow the same general philosophy outlined above.
A draw is not the same as both players losing, nor is it the same as both players winning, but it does tend to have the effect that the higher rated player will lose some rating points and the lower rated players will gain some rating points, bringing them closer to parity.
3 Likes
I’d quite to like to see draws with integer komi handled properly on OGS. Considering we’ve now got night mode, zen mode, a selection of stone and board textures, automatic tournaments etc. etc. it does seem like the developers have time to get on to it. But ofc I’m not going to presume.
1 Like
Just noting that all the nice things you mentioned are “front end” (cosmetic) and have a pool of developers working on them (thanks to it being open source).
Game mechanics is back end and has only one person working on it (because only one person has access).
4 Likes
I did not know this bit of information. I find it interesting and am glad you shared it 
.
3 Likes