Well, that’s a very good question. When I began this thread I figured it was a straightforward matter, but once again I have waded into an existensialist quagmire. Ah well, it’s a fun game all the same.
In all honesty, I really just wanted to write about Spinal Tap. lol
I’ll admit that I haven’t read the thread, and that these two points are minor, but still.
Firstly, it’s good to have draws, at least in non-tournament games. If there is a draw, that just means that the result reflects the reality of the game, and how can that be bad? Is that not the point of the result, ultimately?
Secondly, I would personally favour the tiebreak komi being represented differently. Instead of “6.5”, “6 + tb” in which “tb” stands for tiebreak, representing White’s privilege of winning ties. This seems better to me as in numerical terms, the tiebreak value could be any number in which 0 < x < 1. To format as 6.5 is to first distill a concept that is essentially non-numeric and then represent it with a semi-arbitrary value. But I don’t really mind. 
To pull on the logic of this statement a little 
If there is no draw, that also just reflects reality of the game. How can that be bad?
Really the question is if a komi that allows ties would be a more accurate representation of the skill involved.
I like this problem statement. But I’m not sure how would you go about delivering solution to it. Komi is not an absolute value - it’s there to compensate white for playing second. Logically for beginners value of the tempo is lower than for highly skilled players. So questioning correctness of a komi needs to be done within a context of specific skill level.
The best experiment for evaluating accuracy of komi levels would be to analyse slices of OGS games played with given komi by players of certain strength to check how far of the 50% equilibrium are they.
For example:
Given a skill range, say SDK you could check the stats for komi of 5.5, 6.5, 7.5, etc. If results would show that komi 6.5 is biased by 4% towards Black and komi 7.5 is biased by 6% towards White that would show that impact of applying even komi of 7 is potential improvement of up to 4% scores vs. komi of 6.5 and up to 6% for komi of 7.5.
I’d be curious to see such an analysis and komi recommendations for different skill levels. Realistically I’d expect that for amateur games the difference between komi of 6.5 and 7.5 would be negligible (like <0.1% of games), as many games end up won by time, resignation or margin of 10+ points. I don’t see why would someone even consider a major rework of rules for a game that’s thousands years old, given such a minimal gain.
Similarly you could argue for example that measuring 100m sprint up to 0.01s precision is unnecessary. Clearly if two runners can run 100m in under 9.7 seconds they are both top world class and should both stand on top of the olympic podium?
Actually, that’s not quite what I had in mind.
What I did imagine was that you test for the accuracy of komi via ratings. i.e.: if a certain komi or the allowance of ties affects the ratings in such a way that they are more accurate predictors of winners, the same way we tested for the Glicko-2 constants.
Actually, timing does not determine who wins the race; crossing the finish line first determines the winner. Timing is significant only for record keeping, and some of us do question its meaningfulness when measured in hundredths of a second. When the difference is that small, inconceivably small factors can become significant, and these factors may have nothing to do with the performer’s prowess. With records, or merely comparing performance in different races, you are dealing with entirely different conditions for each race. For example, while wind above a certain strength disqualifies a record, the strength affecting a hundredth of a second is absurdly small and raises questions about accurate wind measurement and how much it may vary in the course of a race. Also, the type and condition of the track becomes much more important. And then there is the question of whether the runner is in motion (but without leaving the blocks). Much less motion, much harder to detect, might well affect the result. (Once, as a line judge, I and my fellow judge disqualified a shuttle hurdle relay team because the second runner was in motion before the “hand-off.”) In short, records measured in hundredths of a second are ridiculous in track, and the difference of a hundredth of a second is meaningless when comparing two different races, in my opinion.
The margin of draw is large in chess - much larger than a couple of points in Go. There are common chess endgames in which one side is a whole piece ahead, but which can’t be won.
That’s more or less similar to declaring a Go game drawn if neither player is more than 3 points ahead.
With integer komi, allowing draws, draws would be rare in professional Go.
I don’t really understand the mentality of people who insist “there has to be a winner” even when both players have played perfectly, but if they need that, they can hold events where the “winner” of a drawn game is decided by tossing a coin.
There are many competitions where the win margins are so tight that the end results is not just effect of skill, but a combination of many conditions which you can summarise as ‘luck’. But for some reason people would not try to rationalise all these sports and make draws more prevailing, rewarding the top performance with a shared top spot. Personally, I would see such a change as healthy mindset switch given social obsession with competition is a little crazy, realistically though, the very nature of competition will stop any attempts to blunt the ‘edge’ simply because of the risk of making those disciplines more ‘boring’.
In essence it seems that the debate about draws is mostly philosophical debate about the nature of competition. And this forum is a great platform for such discussion, as no matter how long we drag it, the discussion will remain academic with no real impact on anything 
Let me clarify what I was trying to say.
If there is a jigo (an equal score by both players), then a drawn result reflects the reality of that particular game whereas a won result for either player caused by a tiebreak komi does not. I didn’t mean to make any point about “the game of Go” in general.
I really don’t think you can equate the .5 at the end of the komi to a coin toss because the result is absolutely not random in any way. It sets a defined line that black must cross in order to win, and if not then black loses. This is not a random result, and I would have absolutely no respect at all for anybody playing black that decided to whine about losing by a half point. You know what the win condition is when you accept the game.
I disagree. If the rules define that a tie is impossible, as per the half point komi, then both players know exactly what the win condition is and neither player has any ground to stand on if they start whining about how “it should have been a tie”.
It is true that X.5 komi is a valid and non-random win condition. However, it moves the win condition of the game further away from the more fundamental relation between the territory and prisoners and the amount of points held by each player, and the effect of those points on the game’s result. Just as komi itself moves away from the emergent principles of Go and towards a prescriptive constructed form, so does tiebreak komi on top of that.
I’m not sure which game are you referring to. If we are talking about Go, usual rulesets do not allow for a draw.
It doesn’t. It just makes the gameplay equally challenging for both sides.
Nobody was “whining” about anything. If you can’t be courteous when you comment, then don’t comment.
You seemed to suggest that having a .5 at the end of the komi is equivalent to making players flip a coin to decide the winner of a tie. I think that’s ridiculous. If you don’t like how I worded it, too bad, get over it.
I’m not sure why old points are being brought up again (especially the ‘perfect play’ one), but I’m going to state my opinion on what my opinion the matter is:
Ties in elimination tournaments or in title matches are problematic and should definitely be disallowed.
Ties outside of this format are OK, but there should definitely be some study of whether or not this reflects the strength of the players involved being roughly equal. (Thus my experiment with ratings).
only at OGS can you get really heated arguments about half a point of scoring (don’t be mean guys, I am sure neither meant any harm )
At first I understood this thread only as a friendly chat about. But just to be sure if some of you would want possible ties as a default settings, then as far as I am concerned as long as having the half point is a worldwide standard OGS should reflect this as default.
OK, so here is my opinion: if the rules (Japanese or Chinese) avoid ties by giving x.5 komi, then that is how the game should be played.
I have had perhaps two games decided by 0.5 points (if I recall correctly, I lost them both). So what? The person who lost simply should have played that little bit better. Is that so terrible? 
Don’t complain about how much time/effort went into your “lost-by-0.5-point” game. Be happy that you had an exciting game and try to learn something from it. You might start complaining about wasted time/effort when you get defeated by, say, 15 or 20 points (especially when you learn nothing from the game). 
– Musash1
Look, I’m not against .5 point komi. As a competitive player, I like that there are no ties.
But as someone who would like results to reflect ability, I think it might be worth researching what the optimal komi is, and whether or not that involves adding the .5 point tiebreak to it.
Also I would like to contest this idea. The rules once called for no komi, but people eventually agreed that it was not how it should be played. As a community, we can make that decision once again on whether or not there should be a tiebreaker, make a new ruleset, and change it.