How to play like an engine !

Sort:
SebLeb0210
BMeck wrote:

Enignes do not need to remember either lol. Its a program

They need to remember there openings and there variations and ther lines ect.

BMeck
buldermar wrote:
BMeck wrote:

Enignes do not need to remember either lol. Its a program

That really depends on how you conceptualize memory retention. For instance, I'd argue that any engine based on an adaptive algorithm that stores previously played positions and adjust accordingly at least to some degree "remembers" by the definition of the word.

I see what you mean but I do not think adaptive engines exist. Hence why centaur chess is played. I am not 100% sure now but in the past i.e. one or two years ago people could automatically draw and beat engines by playing specific lines. Excuse my ignorance with computer programming, but do adaptive algorithms even exist? And if they do are they accurate?

BMeck
SebLeb0210 wrote:
BMeck wrote:

Enignes do not need to remember either lol. Its a program

They need to remember there openings and there variations and ther lines ect.

Those are programed in via an opening book

sisu

Let's make it happen!

sisu

Let's make it happen!

sisu

Let's make it happen!

SebLeb0210
BMeck wrote:
SebLeb0210 wrote:
BMeck wrote:

Enignes do not need to remember either lol. Its a program

They need to remember there openings and there variations and ther lines ect.

Those are programed in via an opening book

Yes but they need to remember that book. ex: usb key.

buldermar
BMeck wrote:
buldermar wrote:
BMeck wrote:

Enignes do not need to remember either lol. Its a program

That really depends on how you conceptualize memory retention. For instance, I'd argue that any engine based on an adaptive algorithm that stores previously played positions and adjust accordingly at least to some degree "remembers" by the definition of the word.

I see what you mean but I do not think adaptive engines exist. Hence why centaur chess is played. I am not 100% sure now but in the past i.e. one or two years ago people could automatically draw and beat engines by playing specific lines. Excuse my ignorance with computer programming, but do adaptive algorithms even exist? And if they do are they accurate?


Adaptive engines do exist. Here is a research paper about one: http://www.sciencedirect.com/science/article/pii/S0167865507002127

If people can continually draw or beat an engine by playing specific lines, as you have suggested, that engine is by definition maladaptive (i.e. it does not adapt according to its previously played games).

With respect to adaptive algorithms: they too do exist. In fact, I'm currently part of a team developing a game that utilizes an adaptive algorithm to adjusts each of five different and distinct difficulty dimensions so that it optimally fits and challenges the unique player . We hope that this will help facilitate post-traumatic brain rehabilitation, and we will be comparing the cognitive effects of training on a game based on this type of algorithm to the same game based on a static algorithm soon.

Are chess engines based on an adaptive algorithm accurate? I don't know. I'm not even sure that they are necessarily superior to nonadaptive chess engines. Chess is a game of complete available information (contrary to a game such as Poker, where the information available to the player is incomplete). As such, there is always a "best move" to be identified (or multiple equally good "best moves"). Therefore, the ideal engine - namely one that can always identify the best move from any position - does not actually need to adapt to its opponent; it can merely look at the position in isolation.

TheGreatOogieBoogie

I want to beat the engine on hardest someday best I did was against only 2100.  Don't hash tables make them adaptive? 

hakim2005

if this man okslo beat houdini he should be the chess world champion

BMeck
buldermar wrote:
BMeck wrote:
buldermar wrote:
BMeck wrote:

Enignes do not need to remember either lol. Its a program

That really depends on how you conceptualize memory retention. For instance, I'd argue that any engine based on an adaptive algorithm that stores previously played positions and adjust accordingly at least to some degree "remembers" by the definition of the word.

I see what you mean but I do not think adaptive engines exist. Hence why centaur chess is played. I am not 100% sure now but in the past i.e. one or two years ago people could automatically draw and beat engines by playing specific lines. Excuse my ignorance with computer programming, but do adaptive algorithms even exist? And if they do are they accurate?


Adaptive engines do exist. Here is a research paper about one: http://www.sciencedirect.com/science/article/pii/S0167865507002127

If people can continually draw or beat an engine by playing specific lines, as you have suggested, that engine is by definition maladaptive (i.e. it does not adapt according to its previously played games).

With respect to adaptive algorithms: they too do exist. In fact, I'm currently part of a team developing a game that utilizes an adaptive algorithm to adjusts each of five different and distinct difficulty dimensions so that it optimally fits and challenges the unique player . We hope that this will help facilitate post-traumatic brain rehabilitation, and we will be comparing the cognitive effects of training on a game based on this type of algorithm to the same game based on a static algorithm soon.

Are chess engines based on an adaptive algorithm accurate? I don't know. I'm not even sure that they are necessarily superior to nonadaptive chess engines. Chess is a game of complete available information (contrary to a game such as Poker, where the information available to the player is incomplete). As such, there is always a "best move" to be identified (or multiple equally good "best moves"). Therefore, the ideal engine - namely one that can always identify the best move from any position - does not actually need to adapt to its opponent; it can merely look at the position in isolation.

Interesting. Thanks for letting me know!

BMeck
SebLeb0210 wrote:
BMeck wrote:
SebLeb0210 wrote:
BMeck wrote:

Enignes do not need to remember either lol. Its a program

They need to remember there openings and there variations and ther lines ect.

Those are programed in via an opening book

Yes but they need to remember that book. ex: usb key.

We take the definition of "remember" differently. In essence we are both right

buldermar
BMeck wrote:
buldermar wrote:
BMeck wrote:
buldermar wrote:
BMeck wrote:

Enignes do not need to remember either lol. Its a program

That really depends on how you conceptualize memory retention. For instance, I'd argue that any engine based on an adaptive algorithm that stores previously played positions and adjust accordingly at least to some degree "remembers" by the definition of the word.

I see what you mean but I do not think adaptive engines exist. Hence why centaur chess is played. I am not 100% sure now but in the past i.e. one or two years ago people could automatically draw and beat engines by playing specific lines. Excuse my ignorance with computer programming, but do adaptive algorithms even exist? And if they do are they accurate?


Adaptive engines do exist. Here is a research paper about one: http://www.sciencedirect.com/science/article/pii/S0167865507002127

If people can continually draw or beat an engine by playing specific lines, as you have suggested, that engine is by definition maladaptive (i.e. it does not adapt according to its previously played games).

With respect to adaptive algorithms: they too do exist. In fact, I'm currently part of a team developing a game that utilizes an adaptive algorithm to adjusts each of five different and distinct difficulty dimensions so that it optimally fits and challenges the unique player . We hope that this will help facilitate post-traumatic brain rehabilitation, and we will be comparing the cognitive effects of training on a game based on this type of algorithm to the same game based on a static algorithm soon.

Are chess engines based on an adaptive algorithm accurate? I don't know. I'm not even sure that they are necessarily superior to nonadaptive chess engines. Chess is a game of complete available information (contrary to a game such as Poker, where the information available to the player is incomplete). As such, there is always a "best move" to be identified (or multiple equally good "best moves"). Therefore, the ideal engine - namely one that can always identify the best move from any position - does not actually need to adapt to its opponent; it can merely look at the position in isolation.

Interesting. Thanks for letting me know!

You're most welcome!

Good luck with your chess ;) 

SebLeb0210
TheGreatOogieBoogie wrote:

I want to beat the engine on hardest someday best I did was against only 2100.  Don't hash tables make them adaptive? 

yes they do.

BMeck
VivianMery_Ressurect wrote:

According to survey. DeepFritz-13 vs Houdini have same strenght. OcKaSlo can defeat DeepFritz-13 without undo/take back. It means he also can beat houdini. I believe it because I've been saw with my eyes.

Does he beat it by playing the same line over and over again? If so I believe you. If not, then I know you are lying

SebLeb0210
VivianMery_Ressurect wrote:

According to survey. DeepFritz-13 vs Houdini have same strenght. OcKaSlo can defeat DeepFritz-13 without undo/take back. It means he also can beat houdini. I believe it because I've been saw with my eyes.

NICE !