.
GMTonyKosten wrote on 09/22/13 at 23:34:32:
... 30 seconds [increments] - by the time you've written down the moves and refocused you've lost a good part of that anyway.
I strongly agree.
REfocusing after moving the eyes and attention away from the board, to the score sheet, is very disruptive. 30 seconds is not enough time to then refocus and think clearly.
But the bottom line is that the Tournament Director must have either adjournments (yuk) or a predictable and reasonable time by which all games of the round will be finished.
I think the best solution begins by acknowledging that all chess clocks available for purchase today are crummy in terms of their flexibility. If we have truly programmable chess clocks with granular configurations for all concepts and all implied concepts...
Divide the "fat" first time-control segment into two shorter or
*thinner* segments. Each might be 20_moves/50_minutes + delay_30_seconds.
Further, any time remaining at the end of segment is only partially carried forward to the next segment! For example, perhaps only 50% of the remaining time is added at the start of the next segment. Phrased slightly differently, any time remaining at the end of segment results in only a
*partial-accumulation* of the prior segment's remaining time.
That way players get the time they need, but they cannot save all unused time and in so doing bloat the maximum duration of the game and thus of the round.
Under these two basic ideas (thinner segments, partial-accumulation of prior time), the players no longer have all the usual *guesswork* about how to manage their limited time. At the start of the game, neither player has any way of knowing whether the game will last of 33 move-pairs or 99 move-pairs; thus they must guess about how much time to save early in the game.
Very few other sports, either analog or digital sports (like chess), have this element of guesswork about time.
By having thinner segments, none longer than 20 move-pairs, the endgame has time if it needs time. But if it turns out that the endgame does not need time, then the endgame's time is not available for excessive and perhaps self-destructive consumption by the earlier segments of the time-control. Mathematically it is impossible to reliably know that the round's longest game will complete within 5 hours if the time-control allots, to each player, on average 3 minutes for every move of a 90 move-pair game (2 * 3 * 90 = 540 minutes = 9 hours). So we must pick our poison.
The thinner and partial-accumulation principles give us more ways to make the poison as palatable as possible.
. .