Re:Researchers create new classification of chess openings
by jfengel on Tuesday April 04, 2023 @12:14PM
(#35441960)
They've defined an idea of "similarity" (based on who plays what openings), and place the dots so that similar openings map to closer dots.
That means there isn't a meaning to the exact position. It's just an attempt to show the similarity visually. Usually, you do that by laying them out randomly, and then having closer dots drift towards each other until you get something more or less stable.
You can think of it as trying to minimize the energy or stress in the graph, though there's no guarantee it will be the absolute minimum.
Their "similar" metric is "different players tend to play both of these openings". Not sure how meaningful that is.
That means there isn't a meaning to the exact position. It's just an attempt to show the similarity visually. Usually, you do that by laying them out randomly, and then having closer dots drift towards each other until you get something more or less stable.
You can think of it as trying to minimize the energy or stress in the graph, though there's no guarantee it will be the absolute minimum.
Their "similar" metric is "different players tend to play both of these openings". Not sure how meaningful that is.