Well, I used 10m by 10m by 5m blocks. Inverse distance squared...

Well, I used 10m by 10m by 5m blocks. Inverse distance squared to a power of 3 for interpolation. Two search expansions using a search ellipse that is 30m by 20m by 2m. Long axis is horizontal with a bearing of N60°E, short axis vertical, and intermediate axis dipping northwest at -40 degrees. So, it is a stretch to project mineralisation between holes that are 100m apart when the hits are maybe just 15m long, so there are gaps between some holes with <0.5% Li₂O and I left those blocks out since I used a 0.5% cutoff. Sometimes there are holes with no mineralisation at all between holes that do have decent hits, and the interpolation takes that into account, and leaves a dead spot there. I only showed blocks that are >0.5% Li₂O. I suspect the models that other posters have put up are running numbers way far up and down the drill holes and too far between them. For example, it is very unlikely that a 15m interval can be stretched out horizontally for 100m with any degree of confidence. Also, intercepts should have hard vertical constraints. Ask the creators of other estimates for the specifics of their models and how they came up with such big numbers. The wireframe in the Canaccord piece with a link on the AZS that claims big numbers is of the entire pegamtite body, and that is given as a target, not a mineralised body. I suspect that exploration has moved on to T2 since the good stuff in T1 is way deep and does not extend very far between holes.