Proposal of new WCC Match Format

They should just remove the kings in the final decisive 13th game. Player without pieces lose. Preferably as a surprise, so the competitors have to improvise, for the entertainment of the masses.

I like the idea of having a variety of time controls from blitz to classical. All games add up. I think that might create a more "balanced" champ (not saying Carlsen still wouldn't be the champion). I'm on board with the multiple time controls.  

Clever - but I don't think you are taking the subtleties of 2750-level chess into account. In games between elite players, Black always plays for equality above all else, and is more than satisfied with a draw. It's the player with White, first and foremost, who determines how high the likelihood of a draw will be.

You have to give something to get something, though. Even with White, one cannot realistically expect to defeat a top-20 player without any sort of risk being taken. And unfortunately, many at that level do not believe such risks are justified, given the ever-increasing difficulty of "winning on call" (mostly due to the depth of computer-assisted preparation) in addition to the obviously high cost of losing as White.

Of course, no format change will make winning any easier. What "5-point scoring" does do, however, is make winning with White more desirable while making failed gambles with White less costly.

I see - in that case, your idea is definitely a major improvement over what took place in this year's match. I have to ask, though: is speed chess really preferable to classical, even for the purposes of the tiebreak? Back in the 1st post, I outlined a fair and feasible solution that never deviates from long time controls.

I do think the 3 Armageddon games I proposed would be overkill, but the idea combines rather well with what you've suggested. The speed chess playoff could easily be replaced by a single round of classical Armageddon. After that, the playoff would end with 2 classical games as per your description.

Here's a format for ya:

Both sides get 100 Hours Time control.  Wins are 1, losses are 0, draws are 1/2.  Whatever time you don't use each game is what you have left for the match.

Match continues until one player has run out of time.  At that point in time, he loses that game on time, and for subsequent games, the other player keeps his remaining time, the player that ran out of time gets 1 minute per game.

There is no increment or delay.  So having the lead is not incentive to drag your own clock down to 0 as you'll have to play a complete series of games with 1 minute while your opponent has his remaining hours.  If he's got 25 1/2 hours left, even a 10-point lead won't be safe for you!

Higher score when all 200 hours are exhausted (the 100 by both players) wins!

Classical Armageddon: interesting; I'm not sure about the time control but the bidding at least keeps it fair.

Scoring system: makes it possible for a player to win the match with a minus score.

Is this some sort of IQ test, where you wait to see how long it takes people to figure out that such a format would only be suitable for World Bullet Chess Championships? Because every game will effectively be bullet, after all - spend so much as 1s more than a fast-moving opponent per game, and it's curtains for you once your clock runs out!

Wow, thanks for reading all that! (the 1st post, I mean) I'm well aware that it was horribly long.

I think the "minus score" issue is something which is interesting to debate, rather than a fatal flaw. As a former mathematics major, I don't necessarily view "5-point scoring" as a system which may award victory to the inferior player. I see two systems ("balanced scoring" and "5-point scoring") which are not identical, and so by definition will not produce the same result in every situation.

For instance, suppose that in this year's match, Carlsen had wound up losing Game 5 (which was arguably the most aggressively played game of the match, based on the choices made in the game's early stages). If all remaining games had played out in the way that they did, Karjakin would have won the match outright. With "5-point scoring", however, Carlsen would have won the match (without tiebreaks), both in that situation and in the one that was reached at the end of the actual match. While one may argue such an outcome would not be at all fair in terms of simple win-draw-loss numbers, I honestly believe the nature of high-stakes, high-level chess to be a bit more complex than that. In my opinion, the unbalanced scoring is a more accurate reflection of what is really going on during a WCC.

Let's take a look at this points system.

White win = 5 points
White draw = 2 points
White loss = 1 point
Black win = 4 points
Black draw = 3 points
Black loss = 0 points

Wins and losses with white score more than wins and losses with black so decisive results favour white. Draws score more for black so draws favour black. Let's look at the extreme cases.

Example 1:

Player 1 scores 3 wins and 5 losses with white, and 8 draws with black.
Player 2 scores 8 draws with white and 5 wins and 3 losses with black.
Player 1 wins the match 44-36.

Example 2:
Player 1 scores 2 wins and 6 losses with white, and 8 draws with black.
Player 2 scores 8 draws with white, and 6 wins and 2 losses with black.
The match is tied 40-40.

Why should a player who scores +6 =8 -2 over a 16 game world chess championship match have to play a series of tiebreak games? Should he not become the world chess champion?

No black sometimes play for the win, and surely does so if his opponent makes innacuracy/Gambit that allows it. Black wins happens and are at least as interesting, they also should be encouraged. A 5-0 2-3 0-5 would still encourage white to win, because it can be presented as the equivalent 6-0 3-3 1-5, where white still takes less risk than black playing for win, but at least, you can't lose a match by winning three times more games than your opponent. So I'm still not convinced by your scoring system.

By the way, (in a duel) your scoring system is equivalent to 4-0 1-3 0-4, which clearly gives white a much greater incentive to win but is presented in a way where winning once as white and losing two times is bad. In comes from the fact that you are giving free points to white, but players will also play black eventually. That also means that your system is awfull for anything else than duel with equal number of games with both color.

The way the scoring system is presented really matters, your scoring system (as well as most example I presented) is really badly presented because free points are given to white. Reasoning like winning X times and losing Y times as White/Black is inherently flawed, because dependent of the way of presenting the scoring system. A very good way to find the right offset for the scoring system to be well presented would be to set it so that each side has the same expected points gain (eg 5/2 for your system). 33/33/33 is a good approximation to keep the numbers round (at least, the number in the database say something like 35/35/30, although I think it's at master level result distribution, not at 2750+).

In the case of your 5-0 2-3 1-4 system, presenting it 4-0 1-3 0-4 gives a small fraction (1/3) of free points to black per games, compared to a 2/3 fraction to white, assuming a 33/33/33 distribution of results. Assuming 35/30/30 makes things even more favorable to the 4-0 1-3 0-4 presentation of the system, 1/5 for black per game compared to almost a full points (4/5) given to white per game with your scoring proposal.

That being said, I really love your armageddon idea.

I'm going to reply to the above two comments in separate posts (my replies are too long for me to include in this comment), but would first like to thank everyone who has been contributing to this discussion. I appreciate ALL comments posted in the thread, whether they contained positive feedback (related to the changes I felt may improve the quality of future WCCs), negative feedback, or were related to interesting alternative WCC proposals.

I guess I should especially thank 4xel and JoeFarnarkle for their willingness to discuss some of these issues in much greater depth. Though the more people who weigh in on the conversation, the better. I know there tend to be many widely varying opinions on such topics, and my ultimate goal back in the 1st post was to come up with something that LOTS of people (maybe even the vast majority) would prefer over the current WCC format.

Yes - that is pretty much the sole motivation behind scoring a match in such unusual fashion.

For starters, it's important to really emphasize the players' ratings, and the impact those ratings have on the general match dynamic. If we were talking about a series of games between amateurs, or even ~2350-rated FMs, then it would of course be nothing short of ridiculous to keep score in the manner that I suggested.

Now, regarding your examples in the context of match play between world class GMs: I think it goes without saying that, if the final overall outcomes (in both situations) were only based on the games where Player 2 had the Black pieces, then Player 2 was either convincingly superior (in the 2nd scenario) or at least a bit stronger (in the 1st scenario) than Player 1. But it also goes without saying that the other half of the games (the 8 draws) must be considered as well. Thus, it becomes a half-subjective decision concerning just how much Player 2 should be punished for taking virtually no risks (based on results like those between two very high-rated GMs, it's a very safe assumption that Player 2 was far more cautious than Player 1, to put it mildly), and consequently getting absolutely nowhere with White.

If one player wants to take the same approach they would in a match with traditional scoring, they are welcome to do so. The whole point of my proposed system, though, is that (as your examples demonstrate) such an overly solid approach will leave them at a significant disadvantage in the long run.

In Swiss style events? Absolutely, as (statistically speaking) they have little choice, especially when paired vs. much lower-rated opposition. In round robins? Yes, though not as often as you might think (since it's not statistically justified in many cases). But in match play? Heck no! Not unless they are at the brink of defeat in the final game(s).

Note that Black will still occasionally play for a win under "5-point scoring" because, when playing aggressively in nearly every game (as White), one will sometimes blunder and wind up in a two-result position where the opponent can only either draw the game or win it.

 A 5-0 2-3 0-5 would still encourage white to win, because it can be presented as the equivalent 6-0 3-3 1-5, where white still takes less risk than black playing for win, but at least, you can't lose a match by winning three times more games than your opponent. So I'm still not convinced by your scoring system.

Correct - this would be a more conservative attempt at accomplishing the same thing. It is an open question whether the added balance would be worth the reduced incentive. "3-point scoring" (3-0  1-2  0-3) is an interesting middle ground lying between your suggestion and mine. My opinion is that the only way to truly settle the debate would be to ask the world's top 50 players (especially those with match experience) about the "least extreme" version of modified scoring that would motivate them to play [a lot] more ambitiously with White.

By the way, (in a duel) your scoring system is equivalent to 4-0 1-3 0-4, which clearly gives white a much greater incentive to win but is presented in a way where winning once as white and losing two times is bad. In comes from the fact that you are giving free points to white, but players will also play black eventually. That also means that your system is awfull for anything else than duel with equal number of games with both color.

You're correct (again) regarding the equivalence of 4-0 1-3 0-4  scoring in match play, which is probably the most relevant context given the title of the thread. I'd also agree that, for technical reasons, some "presentations" of a scoring system are better than others. The estimates that follow (later in your post) appear to be incorrect, however. Before deciding on the precise scoring I wanted, I performed similar estimates using the same reasoning.

If we assume a 33-33-33 result distribution, then 5-0 2-3 1-4 scoring does give White (on average, using 4-0 2-2 0-4 as a baseline for "normal scoring") 2/3 of an extra point per game, but also gives Black 1/3 of an extra point per game. Meanwhile, 4-0 1-3 0-4 scoring does indeed give Black only 1/3 of an extra point per game, but also gives White -1/3 of an extra point (emphasis on the negative sign) per game. Finally, after dividing everything by 4 (to normalize results back down to familiar terms where 1 point = 1 game), we can conclude the following:

In a single game, White gains an average of 1/12 of a point relative to Black under 5-0 2-3 1-4 scoring, while Black gains an average of 1/6 of a point relative to White under 4-0 1-3 0-4 scoring.

 33/33/33 is a good approximation to keep the numbers round (at least, the number in the database say something like 35/35/30, although I think it's at master level result distribution, not at 2750+).

Yes, indeed! The draw rate would undoubtedly be higher among GMs rated 2750+; if we increase it to something resembling a 25-50-25 result distribution, then my choice of 5-0 2-3 1-4 is practically right on the money in terms of fairness (while the "equivalent" 4-0 1-3 0-4 system favors Black even more heavily).

That being said, I really love your armageddon idea.

On the fence here, but... I think I agree with this, too. 

Don't know how I missed this. Given the long-standing tradition of the "Champion's privilege", the concept of just hurling the reigning world champ right back into the fray for the next WCC cycle is quite radical - more so, I think, than the proposal I made to begin this thread. It would unquestionably be met with very serious resistance, though in fairness, the fact that Magnus Carlsen himself supports such a change could certainly be a big help in eventually making it happen.

It's interesting to consider the possible modifications to the existing championship cycle that would yield two challengers for the title match (doing away with the match format altogether would be overkill, imo) instead of one. The first thought that comes to mind would involve two Candidates tournaments (yes, you heard that right! ) - one would include the reigning WC, the previous cycle's runner-up, two qualifiers from the previous World Cup, two qualifiers from the previous Grand Prix series, and the highest placing competitor in each of the two previous Candidates tournaments who did not qualify by higher-priority means; the second would feature the top 8 rating qualifiers among those who did not participate in the 1st Candidates tournament.

Back when I was a mere hatchling on this site, I proposed that we do away with the entire World Championship contest. I was assailed as a troll. I was seerious. I don't think we need a WC now in the age when the best chess player is a gadget. Humans playing chess is an anachronism. Humans are now totally obsolete, physically and mentally. Chess is just one more thing ruined by computers.

Well, you'll just have to accept the fact that Armageddon games have already been established as the modern standard for tiebreaks in high-level competition.

http://www.chessdom.com/kamsky-wins-us-chess-championship-after-armageddon-game/

http://www.bakuworldcup2015.com/news/9/144

Btw, have either of you clowns even played this "game" before? I have - and I know from experience that it would be a dreadful way to decide any serious chess event. The problem is that the whole thing has virtually nothing to do with standard chess, or even with rapid or blitz chess. Not only do rounds/series of games often reduce to an inane bullet chess exercise, but the clock plays a truly disproportionate role throughout all the games in the series; it's not hard to abuse a noticeable time advantage.

By contrast, classical Armageddon is nearly indistinguishable from pure Classical Chess; plenty of time for White, reasonable time for Black, and White faces a must-win situation (which in fact occurs naturally in run-of-the-mill tournaments from time to time). The distraction of bidding can readily be handled weeks in advance, courtesy of the players' teams/seconds.

Of course it's nonsense to you. You aren't capable of understanding the straightforward math being discussed, or even the simple scoring system you are feebly trying to mock. I guess I can give you one last chance to prove me wrong on this:

1. What would be the most extreme possible version of my "unbalanced scoring" concept?
2. What would be its least extreme version (besides just leaving the traditional scoring as is)?
3. Why didn't I make wins as Black most valuable, instead of the other way around?

If you correctly answer the 3 questions above, I'll at least acknowledge that not ALL your comments are truthless, factless heaps of garbage...

...but if you fail miserably, then there is nothing worthwhile that you could possibly contribute here, so get lost. Or better yet, go troll some other forum on a different website - pitiful attempts at insults and low-IQ commentary will blend in much better elsewhere.

I don't see how any of this is relevant to the subject at hand (namely, feasible improvements to current WCC methodology). It's decent material for a philosophical debate, but does not exactly pertain to the real world. Competitive chess, even at the correspondence level, isn't going anywhere in the foreseeable future.

From a theoretical standpoint, the game is still prohibitively far from being truly "solved" by any reasonably strong definition of the word. Practically speaking, meanwhile, computers are no more useful in serious OTB play than they were 50 years ago. If you plan on living to see your prediction actually come to pass, then you'd better hope for some major medical breakthroughs in the coming decades! 

Since shorter time controls seem to be quite a popular recommendation among many players and organizers who are not very optimistic about the future of classical chess, I decided to come up with a 2nd possible WCC match format alternative. Unlike what was described in the 1st post, this layout will be based on something you might call "neoclassical chess" ():

• The players get every 3rd day off, just like they do under current championship match regulations. On non-rest days, however, they must play pairs of games at a 60+30 time control, with just a 1-hour break in between. Thus, their rest days are hard-earned, indeed.
• The full match comprises 24 total games at the aforementioned time control, unless a 36-36 tie occurs, or one player reaches a score of 37 points or more. 3-1-0 scoring applies to White in each game, while [consequently] 3-2-0 scoring applies to Black. 
• If a playoff is needed, there are two basic options (further details below), with the superior choice depending upon the preference of the parties involved. 

1. The fairer method would begin with a 1-point Armageddon game (60 minutes for White, draw odds for Black, time bidding to determine who plays as Black and how much time they receive, and no 30-second increment until move 61), following a rest day after games 23 and 24. The playoff would then continue with 2 60+30 games on the same day and end with 2 additional 60+30 games on the next. The "3-point scoring" from the regular match would remain throughout these last 4 games.
2. If a two-day playoff (like the one above) is not acceptable due to cost-efficiency and logistical considerations, a "zero-day playoff" is a half-decent solution. Simply schedule a potential Armageddon game to start 30 minutes after Game 24 officially ends.

Some comments on the newer proposal are probably in order. I'll list a few here, in roughly increasing order of significance:

-- Many color alternation schemes exist, but personally, I'd suggest the simplest which at least tries to approximate fairness to some degree. This would mean that colors alternate after every round, with the exception of repeated colors in rounds 12 and 13.

-- Press conferences could only be held after even-numbered rounds, for obvious reasons. In fact, the best approach would likely be to schedule them for the following day, at one hour earlier than the normal start time for odd-numbered games.

-- Extra rest days (compared to my original proposal where only 2 rest days were included) compensate for the fact that 2 60+30 games in 1 day are a lot more demanding/exhausting than a single classical game with the usual [longer] time control.

-- Both playoff options are intended to be especially grueling for the players, so they are less apt to play directly for a tied match in the final rounds.

-- My choice of "3-point scoring" here, as opposed to "5-point scoring", can indeed be interpretted as an admission that the latter system might be too extreme/unbalanced. The middle ground approach is also befitting of the overall format, which uses time controls that are a compromise in and of themselves.

-- I suppose this may go without saying, but the major Achilles heel of such a proposal (despite its many virtues) is, inescapably, the revolutionary time control it seeks to employ. Contrary to what I put forward back on the 1st page of this thread (which was meant as a realistic replacement candidate for the current match structure), the idea of completely changing the nature [and quality] of classical chess itself has little to no chance of being seriously considered anytime soon. In fact, I myself would rather preserve the tension that only a 7-hour game can provide.