# Points raycast method doesn't match point rendering

**URL:** <https://discourse.threejs.org/t/points-raycast-method-doesnt-match-point-rendering/27228>\
**Category:** Questions\
**Tags:** points, raycaster\
**Created:** [June 14, 2021, 11:37pm UTC](https://discourse.threejs.org/t/points-raycast-method-doesnt-match-point-rendering/27228 "2021-06-14T23:37:07Z")\
**Posts on this page:** 1\
**Page:** 1

<div class="post-metadata">

**Author:** ![acthp](https://yyz1.discourse-cdn.com/flex035/user_avatar/discourse.threejs.org/acthp/32/18924_2.png) [@acthp](https://discourse.threejs.org/u/acthp)\
**Post date:** [June 14, 2021, 11:37pm UTC](https://discourse.threejs.org/t/points-raycast-method-doesnt-match-point-rendering/27228/1 "2021-06-14T23:37:07Z")

</div>

Using raycast to allow mouse selection of points is extremely inaccurate. Looking at the code, the Points raycast method models a point as a sphere, but the Points rendering seems to be more like a sprite: a point is effectively a flat object, always facing the camera.

More specifically, a sphere mesh near the edge of the near plane is elongated rather than exactly round, but a Point with a disc map in the same location is exactly round. So a raycast near but not on a Point will return an intersect with the Point due to the elongation. With many overlapping points, it is unusable.

This seems like a bug. Is there some reason the raycast method models a point as a sphere? Seems like it should instead model a point more like a circle mesh.
