When will the chess programmers write an engine that plans ?

[BrendanJNorman]

Why not get a human sparring partner, anyway? All you need is Internet and Lichess, the site has gotten so popular that I can basically get a sparring partner for any time control I want of any strength I want. And their blunders will be as human-like as possible.

Because if one wants to play long time control games on LiChess (or any site), you can be sure that a certain percentage (in my rating bracket, this might not be a problem at below say 2100 level) will be using an engine.

At least they'll be "blunder checking" key moves.

And If I'm going to be playing against an engine, I want to at least choose which one.

[Ovyron]

Because if one wants to play long time control games on LiChess (or any site), you can be sure that a certain percentage (in my rating bracket, this might not be a problem at below say 2100 level) will be using an engine.

At least they'll be "blunder checking" key moves.

And If I'm going to be playing against an engine, I want to at least choose which one.

It's very easy to avoid engine users. 99% of my Lichess games are against people not using an engine, and from the remaining 1% I get back my rating points after they get banned (all the people I've reported have gotten banned.)

When was the last time you tried this? It seems cheaters are banned quick so just avoid playing new accounts.

[corres]

I mean that creates a plan and develops a main line that leads to something.
Not the usual engines we have today. That play chess within a Horizont of search depth.

For this the chess programmers need to write a human-like robot what can think like a man and can play chess too.
So simple chess engines can make real plan never.

[mclane]

"Planning" is a human construction.

Humans have to do it because all the future outcomes can only happen in their imagination, and they falter because their imagination didn't match reality (specially when they imagine that a future position has an evaluation favorable to them, yet it's losing.)

Machines instead build a tree of all variations that they can and see in reality the positions those moves lead to (unlike humans that might glitch and imagine a position with a knight in the wrong place), and play moves that lead the the positions they like the most. If humans could do that they'd do it instead of "planning", and would play at least 300 elo better than they do now, showing how the construction is useless.

Humans don't understand what chess is about, that's why machines have to give them big handicaps to stand a chance with their "planning."

(except I'm human, but these statements sound better in fourth person)

Machines do not plan in chess. They emulate chess. But they have no clue what chess is all about.

To be honest all that has changed from the origins of chess until today is the hardware.
The hardware has made so much progress from 4 MHz XT to the hardware we use now that it looks like the programs could play chess. When in fact all that happens is that they now come deeper in the tree then in earlier stage.
The ram that can be used is way bigger.

So in other words nobody tries because all of the programmers believe that the level we reached is perfect.

All are satisfied with the hardware progress.

[BrendanJNorman]

It's very easy to avoid engine users. 99% of my Lichess games are against people not using an engine, and from the remaining 1% I get back my rating points after they get banned (all the people I've reported have gotten banned.)

When was the last time you tried this? It seems cheaters are banned quick so just avoid playing new accounts.

No offense man, but we are playing in different classes.

You're in the 1500s, I'm in the 2300s - so based on the strength of average opponents - I am obviously facing FAAR more cheats than you.

So many at LTC that it makes it not worth playing.

How difficult is it in say a 15+5 training game, for an opponent who is about to sacrifice, to quickly "check" it with the engine?

Maybe he plays the rest of the game, except for this critical moment, by himself.

Their anti-cheating algorithm will NOT catch this guy.

Of course, unsophisticated (the 1200 patzer who merely relays ALL engine moves to the server) cheats will get caught, but sophisticated guys (who are decent players, understand computer play and only check a few key moves) are IMPOSSIBLE to catch.

And there are far more of these types than you think.

That's why I started playing LTC against engines in the first place.

I have played online chess for 20 years (and have a 10,000 game database to prove it!), but haven't played LTC online since it became ridiculous.

Online blitz is still fun though.

[dkappe]

I believe this should be the real challenge for programmers now, adding Elo is getting boring (although necessary for some engines).

Leela has done something on this front. Small nets and small search result in a very different playing style than a crippled Thinker.

https://lichess.org/@/leela2200
https://lichess.org/@/MiniHuman

[Rebel]

I mean that creates a plan and develops a main line that leads to something.

What you need to consider first is the question: WHY, do humans need a plan?

Answer, because their inability to flawlessly look ahead for 100 moves. If humans were able they would need no plan, not even chess knowledge and they would play for 99.99999% the best move.

Not the usual engines we have today. That play chess within a Horizont of search depth.

Because chess essentially is a search game.

And now I have spoiled your thread

[D Sceviour]

I mean that creates a plan and develops a main line that leads to something.
Not the usual engines we have today. That play chess within a Horizont of search depth.

There are a lot of plans that programs follow. First:

(1) To checkmate the other side. More and more, engines are announcing mate even in the middle game.
(2) If there is no checkmate on the horizon, then the plan is to queen a pawn to gain material.
(3) Before queening a pawn, the engines try to look for ways to gain material with combinations. Programs do this better than humans with multiple threats all over the board.
(4) The engines also demonstrate plans for openings and good piece development.

What plan are you looking for?

Firstly, I agree with these 4, but I want to add...

Your 4 ''plans" are mostly related to gaining material, not checkmate.

"No checkmate on the horizon" (mentioned in point (2) ) misses a key point:

Checkmate can only REACH the horizon if you PLAN to reduce king safety by a certain margin.

To do this, engines (especially when I tweak the hell out of them ) can:

(5) Exchange pieces close to the enemy king, so as to reduce king safety (simple examples being a white knight from c3 exchanging for black knight on f6 via a well-timed Ne4 or a thematic ...Rxc3 sacrifice in an opposite castling Sicilian)

(6) Accumulation of force. That is, transferring pieces to the king's vicinity (I believe you guys call this tropism or something, right? ) via rook lifts (Re3-g3/h3), Qh5, Ne5, Bg5 type moves - such plans can be done in isolation or combined with pawn storms.

(7) When the above two strategies are accomplished (king safety reduced and piece forces accumulated near king), usually a sacrificial possibility will open up, which even if it doesn't lead to mate, will lead to such forced defensive measures that the defending side will need to give overwhelming material to avoid mate.

(8) Once those 3 outcomes are successful, your point (2) may come into the picture.

IMO the more smoothly one can make an engine walk through this process, the more "human-like" the play will appear to be.

There are other things, but this thread relates to planning only.

Most of what you suggest is called tropism - that is, to gravitate the pieces towards the friendly king for safety, and the enemy king for attacks. More recently, I have not used tropism calculation in Schooner because there is no demonstration of gain in elo. It is better for the engine to find these things out for itself, with more primitive considerations dictating the plans. Attack lines to the king are more important than piece placement.

There was a Schooner game recently that Graham Banks and I were observing. It looked as though Schooner (playing black) could not castle safely. Black castled queen side anyway and then began to gravitate all of its pieces towards its own king. Only after the king was safe, black began operations on the other wing. The whole plan came together with smashing success, and without tropism calculations.

Unfortunately I did not take note of which game it was. Graham Banks has been using Schooner a lot lately for his tournaments.

[mclane]

I mean that creates a plan and develops a main line that leads to something.

What you need to consider first is the question: WHY, do humans need a plan?

Answer, because their inability to flawlessly look ahead for 100 moves. If humans were able they would need no plan, not even chess knowledge and they would play for 99.99999% the best move.

Not the usual engines we have today. That play chess within a Horizont of search depth.

Because chess essentially is a search game.

And now I have spoiled your thread

I don’t think so Ed.


I do think of we would have algorithms that create planning, this would also help the normal tree search based chess engines.
Because any chess programs comes in the situation that the tree search comes to an end.

Maybe we have to create or found a new subgroup.

Where Engines exist that do no tree search but only analysing of the position and creating a main line that was not founded the usual way.

Some say that the humans have to play this way because they lack hardware power in the brain.
So the programmers like to stay in the old paradigms. And continue stupid programs as they did in the last 50 years.

Ok. I can live with this.
But it is not what I am interested in.

I am interested in the idea to solve the problem without this huge search tree.

I am NOT interested to see what a chess engine CAN DO if it computes millions of positions. That is boring.

I want to see if it can analyse the position and come with a senseful idea or plan because it knows about chess.

We had human beings on the moon. But no chess engine that can play chess.
Ok we have Stockfish or komodo or others who play chess by going deep in the search tree. But when my blind grandma who can only play
Bingo gets hundred keys and tries them out, one of them will maybe open the door of the garage. That’s not understanding chess,
That’s using hardware progress to open the door.

The results of AI in computerchess are minimal.
The progress is minimal.

One could summarise it that chess engines have no idea what chess is and what the goal of chess is.

But they know how to reach 40 plies in 3 seconds.

Look I searched for this goal a long long time,
When you appeared in cologne Ed together with Jan and the hair dryer on stage , I was there.
I was there when thomas nitsche held his speech about computerchess progress in intelligent search strategy in 1986 .
Ossi and richard had a very weak mephisto cologne.
Levy tried with a motorola 68K Hardware in cologne. I think even mike johnson was there in cologne with motorola 68K Hardware at that Time.


At this time I still had hope and thought: yes there will be one day we will have intelligent chess engines.


But today I am watching the “progress” and I get the feeling that the development ends.

[mclane]

I had an atari ST and bought a chess program made by an eccentric guy from Oxford. It had a very low search depth and I was fascinated.
I thought : this is it.

And followed his work over the years to come.

The guy has birthday today.
He is now 71.