Position's sharpness indicator
Sort:
I've been researching a little bit, and I haven't found any analysis board that says how sharp a position is. Here I'm asking 2 questions: Do you know any analysis board that does that? and if not, how could it be made?, I mean, maybe by evaluating things like how open the kings are, how many pieces are pointing in one way or how much does the eval changes if you dont make the best move.
Nope not really, just which side the engine thinks it is in favor of, but thats an amazing idea! I know some coding stuff so I might think of making that (albeit that would be quite difficult)
A sharp position has is a large difference in evaluation between the top move and the alternatives.
Oh thank you very much. If you ever do it, please tell me about it
theswooze wrote:
tygxc wrote:
A sharp position has is a large difference in evaluation between the top move and the alternatives.
why dont you just look at the relative imbalances between the best move and the other moves? wouldnt that tell you what you need to know?
Yep. It would pretty hard to tell a computer that though.
theswooze wrote:
CaroKannEnjoyer02 wrote:
theswooze wrote:
tygxc wrote:
A sharp position has is a large difference in evaluation between the top move and the alternatives.
why dont you just look at the relative imbalances between the best move and the other moves? wouldnt that tell you what you need to know?
Yep. It would pretty hard to tell a computer that though.
why do you need to tell a computer? the computer doesnt care how sharp the position is.
I was talking about if I were to code something to look at the sharpness of the position, referencing the first comment I made on this forum.
theswooze wrote:
CaroKannEnjoyer02 wrote:
theswooze wrote:
CaroKannEnjoyer02 wrote:
theswooze wrote:
tygxc wrote:
A sharp position has is a large difference in evaluation between the top move and the alternatives.
why dont you just look at the relative imbalances between the best move and the other moves? wouldnt that tell you what you need to know?
Yep. It would pretty hard to tell a computer that though.
why do you need to tell a computer? the computer doesnt care how sharp the position is.
I was talking about if I were to code something to look at the sharpness of the position, referencing the first comment I made on this forum.
why do you need to code anything? just look at the relative strength of the top vs secondary moves. what more do you need than that?
Eh, could do that too.
Nov 13, 2023
0
#8
tygxc wrote:
A sharp position has is a large difference in evaluation between the top move and the alternatives.
The concept is not really that simple. That would make every position where there is a boring recapture "sharp"!
arsejemwhittier wrote:
I've been researching a little bit, and I haven't found any analysis board that says how sharp a position is. Here I'm asking 2 questions: Do you know any analysis board that does that? and if not, how could it be made?, I mean, maybe by evaluating things like how open the kings are, how many pieces are pointing in one way or how much does the eval changes if you dont make the best move.
How about defining a function of material difference and engine assessment of a position? If there's a huge material advantage, that position isn't "sharp" but just "won". But if the engine assessment is quite a bit larger than the material difference, perhaps it's "sharp".
A sharp position places a high evaluation on immediate tactical chances. I would suggest that a position can be "sharp" if one side has attacking chances or if both sides do. A position may be sharp even though the engine assessment is level, if the chances for each side are all defensible.
Certainly, a position is "sharp" if one side has a big choice of viable, tactical plans, provided that there isn't an overhelming win and that tactical resources are also available, in each possible sequence, equally for either side.
So maybe we need to junk the bit about engine assessment being larger than material advantage as a measure of sharpness? I think I'm coming down to the basic idea that a sharp position is one that is fairly level but which has a range of tactical chances on offer? What do you think?
Further from that, an indicator could measure the number of surprising, forcing sequences available in a position, maybe? That actually shouldn't be too difficult to program. I don't mean "surprising" in an engine assessment scenario. A sharpness indicator has to recognise positions that are difficult for humans to assess. I imagine someone's already pointed that out.
@11
Yes that is right. So maybe:
A sharp position is a position where there is a large difference in evaluation between the top move and the alternatives, except when the top move restores material balance.
There is such a thing in chessbase: it´s a function of the ´Buddy Engine´ called visual evaluation. They are using flames to indicate sharpness.
