Well, going back to the title. I don't think that fortress has a practical value. Playing against computers there will be a fortress once in blue moon, you will simply never reach it. Amongst humans probably more often because one fails to win more often and more positions look like fortresses.
And we're talking of winning - ok, the one or other draw might be possible, agreed.
In my opinion if humans play assisted matches like corr. games meanwhile the engines becomes often worse than playing alone. These are the days. No one wants to hears this, I cannot help.
What's the story with this renowned corr GM who lost against the machines without human influence although he was machine-assisted?
Computer engines only play against each other nowadays for rating points. We will see the better programs ratings continue to sky-rocket as inflation takes over. With the base rating starting so high, no doubt a 4000 rating will be seen by the top engine in short time, as it beats up on the lowly 3200-3400 engines.
Going back to the title one has to admit that the question is loaded in itself, because it already implicates that this is so. Here are some similar questions to "discuss":
Why are males always sexual predators?
Why are women inferior to men?
Why are the Jews responsible for the bad weather?
Why was the moon-landing faked?
Why is the earth 6000 years old?
Why is \pi exactly 3.2?*)
etc., etc.. Just pick your favourite prejudice/conspiracy theory/hoax and put it into a question. Don't be surprised, though, that the discussion arising will not be exactly about what you intended.
*) Random fact: for the last one there was even an attempt to put it into a law in 1897 in Indiana.
In the year 2525... (heard that song?)
Anyway, the top chess playing engine, Zeus, will sit upon Mt. Olympus having achieved the maximum rating possible of 10,000. Zeus has not lost a game nor has any program achieved a draw in over 200 years. Zeus has won all his games with an average of 151 moves, and incredible mark as the average games last well past 200 moves for the lesser ranked programs. Zeus refuses to play bullet chess, he says the only real test lies in a complete game of 3 seconds or less with no increment.
By the way, I played 2004 a 2-game show match against Shredder on decent hardware, classical time control. At that time it was one of the best engines on the market.
So, two long games, one was lost „narrowly“ (at least people thought that), the other one was a fortress/bug/horizon effect, draw in a probably won king-pawn endgame.
In the year 2525...
The top chess engines refuse to play against the World Champion.
They say that games against inferior species are not accepted.
Hey ! Come on now.. give us humans some credit !
We are smart enough, (well, can't speak for everybody) to know not to challenge the mighty machines. We may be inferior, but we're not stupid :-)
"A computer once beat me at chess, but it was no match for me at kick boxing".
Because I've never played one yet.
> Of course, the newer versions crushed the older, that is what it's about, getting faster and capable of storing more memory.
I don't think you read the post...
1. All the engines played on period hardware similar to that used in the Brains in Bahrain match
2. The same version of Fritz that tied Kramnik in a 6-game match was utterly annihilated by a modern engine
3. Ergo, we have good reason to believe Kramnik and players around his level would be annihilated in a long time control match against Houdini 3 (or the even more modern and stronger engines out now like Sugar, Houdini 6, Komodo etc)
If you don't accept this you're just delusional