It is interesting both Alpha and Stockfish are still playing exclusively only open positions:
Seemingly, closed positional treatment is still taboo for them.
Another intersting thing is how frequently Alpha chooses 1.d4.
But then, it does not know 1...c5 draws easily.
*facepalm*
The obvious reason why closed positions are more difficult to play is that, when there are blocked structures, direct threats/attacks are almost lacking, which means that the tension could not be resolved easily by captures on both sides or progress on open lines. Instead, correct play there involves slow regrouping, sometimes on a large scale, including more pieces, which might take 10, 20 and even more than 30 moves.
30 moves are a lot of depth, so even the top engines will have difficulties to decide what to do, if their evaluation is suboptimal, which is frequently the case. That is the simple reason closed postions are much more complicated for both humans and engines than open ones: immeasurably bigger depth.
As seen in the Alpha-SF match, none of the 2 engines is still up to the task, they both tended towards more open play.
*facepalm* 2
The obvious reason why closed positions are more difficult to play is that, when there are blocked structures, direct threats/attacks are almost lacking, which means that the tension could not be resolved easily by captures on both sides or progress on open lines. Instead, correct play there involves slow regrouping, sometimes on a large scale, including more pieces, which might take 10, 20 and even more than 30 moves.
30 moves are a lot of depth, so even the top engines will have difficulties to decide what to do, if their evaluation is suboptimal, which is frequently the case. That is the simple reason closed postions are much more complicated for both humans and engines than open ones: immeasurably bigger depth.
As seen in the Alpha-SF match, none of the 2 engines is still up to the task, they both tended towards more open play.
Makes a lot of sense. Thanks for the insights. Now I just wish I had the skills to utilize them! 
See what a closed position means:
SF thinks black is some 200cps better initially, when, in actual fact, white is winning that.
How could it see that the a5 pawn will fall, then a4 will be pushed to a6, and then some lines will be opened and white will penetrate?
With all of the regrouping manoeuvers, and each step taking some 15-20 moves on average, a decisive white advantage might come only after some 50-70 moves or so. That makes more than 120 plies.
No engine on Earth could see that, unless it has tremendous evaluation.
You think Alpha would perform better here?
@Lyudmil_Tsvetkov: I beg to differ. I think playing closed positions is easier, as long as the position remains closed, simply because there’s nothing to calculate.
I myself plan on releasing a second part at some point, that will include a large sample of world champion and top engine games, explaining in detail the different terms. Also, getting rid of numbers and substituting them for human interpretations.
Why not give both? Most people are used to the numbers
This guy should be banned. He has multiple accounts (GWTR is one of them). And he is liar.
One more thing - he thinks he can beat stockfish. This even might be true, because stockfish strenght depens on hardware you're using.
If he was great he could easily prove it. And would prove it instanly, it would be best recommendation for his book!
The truth is he is not great and he has to pretend he is great.
Personally I wouldnt buy a book from a guy who isnt gm and there is no evidence of his games at all. And who dont want to prove his abilities.
This is all bullshit and hope chess.com will ban him soon.
This guy should be banned. He has multiple accounts (GWTR is one of them). And he is liar.
One more thing - he thinks he can beat stockfish. This even might be true, because stockfish strenght depens on hardware you're using.
If he was great he could easily prove it. And would prove it instanly, it would be best recommendation for his book!
The truth is he is not great and he has to pretend he is great.
Personally I wouldnt buy a book from a guy who isnt gm and there is no evidence of his games at all. And who dont want to prove his abilities.
This is all bullshit and hope chess.com will ban him soon.
I am not him. Please relax.
But not very well 
I thought that as well.
Mods, please move.
This guy should be banned. He has multiple accounts (GWTR is one of them). And he is liar.
One more thing - he thinks he can beat stockfish. This even might be true, because stockfish strenght depens on hardware you're using.
If he was great he could easily prove it. And would prove it instanly, it would be best recommendation for his book!
The truth is he is not great and he has to pretend he is great.
Personally I wouldnt buy a book from a guy who isnt gm and there is no evidence of his games at all. And who dont want to prove his abilities.
This is all bullshit and hope chess.com will ban him soon.
I am not him. Please relax.
thisthis is really funny stuff to read.
@Lyudmil_Tsvetkov: I beg to differ. I think playing closed positions is easier, as long as the position remains closed, simply because there’s nothing to calculate.
Unless it is an obvious fortress, I guess it is the other way round.
Nothing to calculate, but a lot to evaluate positionally.
Why do you think precise positional evaluation is easier than calculation/tactics?
I myself plan on releasing a second part at some point, that will include a large sample of world champion and top engine games, explaining in detail the different terms. Also, getting rid of numbers and substituting them for human interpretations.
Why not give both? Most people are used to the numbers
Indeed, this is the better way to proceed, but most people, including Mr. Smerdon, sometimes find such numeric assessments weird.
This guy should be banned. He has multiple accounts (GWTR is one of them). And he is liar.
One more thing - he thinks he can beat stockfish. This even might be true, because stockfish strenght depens on hardware you're using.
If he was great he could easily prove it. And would prove it instanly, it would be best recommendation for his book!
The truth is he is not great and he has to pretend he is great.
Personally I wouldnt buy a book from a guy who isnt gm and there is no evidence of his games at all. And who dont want to prove his abilities.
This is all bullshit and hope chess.com will ban him soon.
You know some GMs buy their titles, don't you?
I did not play a single game here.
Maybe someone else played for me.
I've probably challenged Lyudmil as much as anyone about his bold claims not supported by independent evidence. But I haven't called him a liar, a fake or a fraud. I haven't insulted him and he hasn't insulted me. In another forum he made a statement about the Fischer-Spassky match of 1972 that I knew was incorrect and told him. He stuck by his guns but I asked him to provide evidence to the contrary. He researched it and and admitted he was wrong. When he learned he was wrong, he could've ignored it but he admitted it publicly. That shows some character. I'm skeptical about his claims but I don't think the man is a liar or a fraud. I think he genuinely believes in his theories and has presented them again and again despite being ridiculed by many. I've got to give the man credit for brushing off all the insults and continuing to make the case for his book. Calling for him to be banned from the site is ridiculous. He has said or done nothing to warrant that.
This already is a matter of discretion.
And will depend on where the focus of the discussion is.
I wonder why so many people care so much for superficial details.
So far, I read 5 or 6 posts that do nothing apart from making personal remarks. And you want to convince me this is the better way to proceed? I am used to posting chess games and diagrams, and comments on them, that is what makes the thread meaningful and the discussion productive.
Really weird. Where are your games and diagrams?
Something positive to say, or in substance?