K Skold vs K Arnstam, 1963

click for larger view

28 ♖b3-g3+! <pins against squares: g4-sq>

click for larger view

J L Hammer vs Carlsen, 2003

17 ... ?

click for larger view

17 ... ♕b5-h5+! 0-1 <pins against squares: h4-sq>

click for larger view

True, although in this case Sköld was not shield but sword. :-)

It seems to me that we have been gradually approaching an ultimate answer on this question, and it seems to be that, with best play, every game *should* end in a draw.>

I question that but don't have enough evidence to dispute it yet. Connect 4 has been solved to be a win, with perfect play, for the side that moves second. Checkers has been solved to be a win, with perfect play, for the side that moves second. Tic-Tac-Toe has been solved to be a draw with perfect play.

I grant that it is possible.

However, I note that the more complex a game becomes, the less significant the first move in it appears to become. And at least to date, all investigations show almost every "reasonable" opening sooner or later resolving to equality (although some disappear into intermittent clouds of ill repute, only to emerge when someone's analysis discloses an improvement).

White begins with a pull, but it weakens over time. (Perhaps, from his perspective, chess is dogged by inertia and will end in entropy's dread and icy grip. ;-) )

This is not conclusive, but it does seem to corroborate the idea of chess ultimately resolving to a draw.

(Interestingly, though, I note that the two examples you gave that were decisive, worked out in favor of the second player. Would this not imply that we should be looking to prove chess a win for *Black*?)

graded first mooted point sliding be frees right,

rook over hoofed queen be effected 28.rg3+ bed e4,

bades on good will since doff my cap c6 thrice at,

engender a foilable mission concorded rook eli see,

tricky delve fetch rook bet farm red cattle grid eg,

boogaloo 28.rg3+ f4 he find must re free in,

generate ford dynamo at or quec6 could wrench c1,

off in kingh2 light you repatriate will loosen both,

rooks we demonstrate am included 28.Rxf4 ignoble,

the feat sledged in c1 wiggle you feel it is old in,

decisive 28...Qc1+ eyes h2 g1 and alf f4 point be,

to shared I see no debatable 29.Kh2 qxf4+ look you,

set he forced in rg3+ when a huffed queen cook over,

flow 30...Qxg3+ leading to equality rag keep going undone knight e6 hook g7 cuffed or chest queenh6 light having uphill struggle am good be duelling boost f6 core approval edict it is hang hive king a net not bothered by go nearer g4+ etc look aiming,

glorious again light draw I f4 kind walled ratchet back it space in spin off banger bill a good while,

living blind free inch fluffed refer cages in fact,

rook basking the lime light fringe benefits moot he,

bopped ha good lift acrossing a dilemma in eg it is her bet c6? happened 28.rg3+ pawn f4 in badge eg,

honoured 28...fxg3 a foot in he has lock finger,

true delight of education 29.rg4+ king has to be,

protected 29...ng7 firm in glow a moments pause de,

"Perfect play by both sides leads to a draw."

"With perfect play, the first player can force a win by starting in the middle column. By starting in the two adjacent columns, the first player allows the second player to reach a draw; by starting with the four outer columns, the first player allows the second player to force a win."

I think we would need several conditions to be met for this to be true. First, we would need to have "solved" the game of chess to prove without a shadow of a doubt that it is a forced win for one player or a forced draw.

Then we would need to find two "players" each with a perfect eidetic memory who could memorise all of the winning/ drawing lines to be able to reproduce the theoretically ideal outcome in every game. It is, I suppose, just about possible that there might be a solution which involves a winning technique flawlessly rather than winning lines, so the requirement could also be that both players must be able carry out the winning technique instead of memorising every variation.

Then each player would need to have total confidence in their opponent's ability. Chess could be reducing to the equivalent of a pi-recital competition, where the only way to win is to have a better memory than your opponent.

Finally, having somehow arranged all that, we would need to find at least one of the two players who was simultaneously omniscient and daft enough to sit down to play a game they could not win.

That's what I get for quoting from memory.

Hopefully, I got the Tic-Tac-Toe outcome correct. :-)

The result of perfect play for chess is still to be determined...

I miss quoted those examples.

Chess is much more dynamic than these games. Most of the pieces are not quite as limited in scope. So I don't think it will be solved as a win for black.

Perhaps the opening statistics are an indicator of the likely outcome. The Kings Indian Attack for example: King's Indian Attack (A07) With more than 7000 games, white wins more often than black or a draw.

<Once: ... a massive artificiality ... the concept of taking turns.> In Avalon Hill's Third Reich, the Axis player moves first in Fall '39, and thereafter whichever side starts the turn with more BRPs moves first in that turn. This promptly produced the hallowed, dreaded, and two-edged meta-game weapon of the "flip-flop", in which the Axis player <deliberately> spends his BRPs down below the unwary Allies player, sacrificing the initiative next turn so that he can hoard BRPs and take <two turns in a row> at some point thereafter. This adds a devious economic war on top of the hex-board war, and throws a royal monkey wrench into both sides' defenses. It's hard enough to devise an interlocking line of armor units whose zones-of-control can withstand any one turn's attack-breakthrough-exploitation (which is already two attack-move sequences within one turn, as long as you win the 1st one). But try to plan against a possible <two turns> of attacks in a row, and neither side ever has enough units to fill all of the gaps. So both sides expend 2/3 of their mental stress worrying about and preparing against it, and in most games, neither side dares to actually do it. The richness is truly all in the analysis!

Not only does chess impose game turns, it also restricts each side to move one piece per turn (except castling). Relaxing that 2nd constraint gives you Avalon Hill's Feudal: still turn-based, but you move all of your pieces each turn, as in a hex boardgame. This leads towards the joys/evils of movement points, terrain effects, and other ways to distinguish mobility.

As an armchair game designer, I wonder: what chaos-prevention mechanism could you feasibly use <other> than turn-based? Computers can provide bookkeeping support in the role of a referee who never gets bored, which leads to <real-time> (of course), or <concurrent> (which makes interceptions, rendezvous, and other middle-of-turn engagements difficult to manage). I've always chafed at the notion that these few schemes exhaust the game space: surely there's an innovation left to invent.

Afterwards I thought about 28, Rxf4, but 28...Qc1+ 29. Kh2 Qxf4 30. Qxf4 Nxf4

28.Rxe6(?) fxe6
29.Qg5+ Kf8
30.Qg7+?? Rxg7 >

Well, chess is <currently> a pure turn-based game. On each turn, each player can (and must) make precisely one move. But it wasn't always so. The game used to be played with dice, like backgammon. This meant that one player would be able to move more on his turn than his opponent.

In some odds games, the stronger player would give his opponent one or more moves start.

There are some games where you can prevent your opponent from taking his turn (eg backgammon, dominoes).

If we venture into the world of war-gaming, we find warfare simulations (eg Warhammer) where every unit in an army can move on a turn or round. This means that a bigger army gets more turns per round than a smaller army.

There are some chess variants where you make more than one move at a time (eg progressive chess).

Chess could also relax its insistence on turn based moves. For example, it has never felt quite right that stalemate should be a draw. If I have tied by opponent down so much that he can't move anything, then surely I should win? Why is it so important that he should be allowed to move?

Or for that matter, why can't we pass a move? Admittedly, that would deprive us of the delights of the zugzwang. But wouldn't it be more realistic? Armies don't feel compelled to move all the time. Sometimes they just hunker down in a defensive posture(hunker in a bunker?)

Then I looked at Rg3, and saw it would be a win if not for the pesky pawn... maybe I'm trying to remove the wrong defender?

So then I looked at RxP, and saw if NxP you have a win with either the Queen & Bishop, or the Rook on the column.

Then we would need to find two "players" each with a perfect eidetic memory who could memorise all of the winning/ drawing lines to be able to reproduce the theoretically ideal outcome....>

I would hazard that when <Nick46> wrote of "two all-knowing players," he had principle rather than practice in mind. His point was that, with perfect play, White should always win.

As you have seen, I do not agree, but I think <Nick46> never meant his thought experiment to be translated to an actual case of two omniscient players. (Which is just as well, because if even one such player should ever appear, chess will be forever stripped of its mysteries, and with them all the joy of playing.)

I don't either. Perhaps I should have borne Poe's Law firmly in mind and appended an emoticon.

As I previously affirmed: We do not yet know how chess games should end if perfectly played, but the clear trend in opening theory does seem to suggest that White's initiative melts away over time, and games do seem to trend toward equality.

You can read it here -> DcGentle chessforum