# Queen vs Rook neighborhood

Queen and rook neighborhoods are two common ways to calculate statistics for a focal cell. They are also known an Moore and (von) Neumann neighborhoods. I think a "bishop version", i.e. from lower left to upper right corner also exists. But why are these two methods still implemented in software packages?

Source: Lloyd, C. (2010). Spatial data analysis: an introduction for GIS users. Oxford university press.

I always thought that rook was used in the "old" days when computation power was low and that queen´s case should be used in GIS analysis wherever possible. I wonder whether there is any advantage of rook vs queen?

• this doc can help you loo at Defining Contiguity Commented Dec 7, 2015 at 20:57
• @GeoStoneMarten in fact now I am even more confused. It says "Of these three the rooks case is the most commonly used and most programs only will compute this particular case." But...why? The Queens case seems more advanced to me. so...the question is still open.
– Jens
Commented Dec 7, 2015 at 21:29
• comparing result: in vector with queen case you can get all in contact of a center entities. In rook you can filter entities with only one point junction. That is just an exemple Commented Dec 7, 2015 at 22:08
• My guess that conversion of Rook groups into polygons will result in a single part polygon, pleasure to work with. 8 neighbours might result in multipart polygon, b-r-r-r Commented Dec 7, 2015 at 22:30
• @Jens all depend of the regularity of the matrix geometry and shape type. Exemple uniform matrix (square, triangle, square with 45 ° rotation ) non uniform (admin boundary) That also depends of the propagation (you can also create a specific propagation in zigzag if it's really your case.) That is just a method. You need choice the best for your study case Commented Dec 8, 2015 at 10:08