In most chess games someone has a winning advantage by the 15th move

Ponz, I appreciate what you are saying, but to rigorously prove a position is winning, one can not just throw it into a chess engine.  One must have a set of moves for one side which offer no chance of escape for the other side. It is easy to make claims about the true state of Chess without actually knowing.  One can not be 100% sure of a claim without checking 100% of the variations referred to. Chess engines have various ways of *estimating* when a position is won or lost, but (perhaps I am just not informed completely about how well analyzed Houdini is) the logic (as far as I know) is missing that *logically* relates the estimates to the true state, or complete evaluation, of the game tree stemming from the position.

In the first position we can note:

1.Black is cramped, weak on the dark squares, but may have some fight left.

2.White meanwhile has doubled isolated c-pawns, but this doesn't mean much as white has more space, better development, the bishop pair, and Bf4 and the queen is threatened and the white queen and bishop converge on the terrible d6 square.  White is much better I think.

Bf4,e5 doesn't work, and f4 (with the idea of e5) leads to Qb6+.

Ba3 looks worth considering, it covers the d6 square with a threat, and said threatened piece can't be blunted with e5.  If there's nothing better I'd recommend Ba3 but looks best to me personally. 

Cavatine  I have far more "faith" in the strongest chess engines than you do.

Also, from my own experience I would expect to win any position where I had a half pawn lead.  However, I realize this is not enough to force a win and it would probably take a one pawn lead to force a win.

Snowyqueen   You are talking about a pawn advantage at the end of the game and only king and one pawn vs king. And you are also talking about games with opposite colored bishops in the endgame.

But before there is an endgame you will have the opening and middle game and these situations that you described can easily be avoided when you have a pawn advantage say in the late opening or early middlegame or say on the 15th move.

I probably could show you a pawn endgame where one side is up several pawns and cannot win.  In fact I had one such endgame recently.

I say between move 15-20.

chessmaster, remember we include people who have just learned the moves in this group of players which brings down the move number where one side is winning.

I don't think I could say I could win confidentally a position where i have a 1.25, or 1.5 advantage, I'm just too erratic with keeping my advantages. I'm confident that I could convert a general position where white is just up 2 clear pawns.

Infinite   The hypothesis was someone has a winning advantage by the 15th move. Not that any particular person could win from the winning position.

I always thought I could win with a 1/2 point advantage but I was an optimist on this.

Lou-for-you  I like your statement.  "A new type of chess. A ponzi game : after 15 moves a winner is declared."

Another option is the doubling cube.  After 15 moves you are playing for double stakes if you do not resign.  So if you are losing at the 15 move mark you better resign or you will be playing for double stakes.  [double staakes if you are playing for money  or you win or lose two games rather than one]

of course if you get doubled and improve and now think you can win you can redouble!

[doubling cube is great for money chess]

hessmaster,  I was fortunate to quit a few years before there was cheating.

In the years before chess engines got to expert or master level.

[I had to quit because of my health]

Thinking back, however, in correspondence chess they may not have been cheating as later it was unclear if chess engine help was allowed?

qablo  You do not agree that I expect to win when I have a half pawn advantage?  

I always expected to win when I had such an advantage.  However, with chess engines being used may be I could not.  Guess I was thinking about the "olden days" where my statement was true.

Magnus is a genius for sure. 

well white is winning after 1.e4, so with perfect play, white will be ahead after 15 moves 

qablo

Using drawn endgames king and pawn vs king does not imply that one pawn  advantage  is not enough to win on the 15th move as players know all these drawn postions and usually they can be easily avoided.

By your anology even being up a bishop and a pawn at the 15th move would  not be enough to win as there is the well known position with king vs king and bishop and pawn which is a draw.

Also when I say the superior has enough to win, I mean with perfect play by White.  Of course in a tournament where there are many factors which affect humans there will not be perfect play.

I was once 3 pawns down vs a grandmaster and won. This does not mean that a 3 pawn advantage is not enough to win.

A couple years ago, NM @Reb said much the same thing in one of these forums.  

He indicated he was simply too tired playing against (so many) engine users at his level. And was considering giving up online play, and playing only OTB.

It's why opening study is important.