The new information content of the MRE seems relatively low,...

The new information content of the MRE seems relatively low, given the results already published - a few minutes of arithmetic gives a crude estimate, tallying previous releases, see e.g. the footnotes of this post, which is a bit outdated now, and was really just intended a sanity check for the peer comparison with CBMM. Even more crudely, one can take, say, 1.5km (length) * 1km (width) * 40m (depth) * 2.8 t / m^3 (s.g.) ~ 150MT @ ~1.5% grade. Obviously the error bars are suitably large.

As for timeline of mine development, it's an important point, but the 1-2 years at the margin are lower risk than the upfront risk, and so, my subjective discount rate for those is not especially punitive - let's say it gives an additional +/- 10-15% uncertainty on NPV.

To me, the primary outstanding risk item is clearly mineralogy and processing. I have posted previously on what I understand to have been wrong with Niobium @ Mt Weld and REE @ Araxa - the main items being particle size distribution (too much of the desirable minerals in the fines), mineral texture (poor liberation properties), and alteration processes moving the desired elements into other host minerals (e.g. niobium in crandellites at Mt Weld). This appears to be a typical risk with the weathered, enriched deposits we are interested in. @ozblue has said that thorium content at Araxa is a significant problem, and @Naydoe has said that thorium is prohibitive at Mt Weld, but does someone have a source for this, or more information? The two papers I cite above analysing the viability of these two processes make no mention of thorium, and in this old Araxa operational paper, 2% thorium is just described as being removed as slag in the smelting stage (ferroniobium synthesis), with no mention of further complications.

I have become less concerned about the possibility of comparable mineralogical issues to the above at Luni, as I interpret the liberation numbers given in the preliminary mineralogical report (>85% of niobium-bearing particles being "well-liberated" or "high-grade middlings" in the -150+38 micron fraction) as constraining the particle size distribution, i.e. it would not be possible to achieve those liberation statistics if the majority of pyrochlore particles were, say, <30 microns, and similarly I conclude that the mineral texture should not be a major concern (perhaps a geologist can confirm or refute this).

Lastly, @galliumaslan, I think you misinterpreted my previous post. Both DCF and least squares regression are useful, elementary tools, but only when used to calculate numbers, and only within their domain of applicability. The number I had in mind with my original question was as follows: the explanatory power of forward-looking value estimating factors, using DCF or your calculation of choice, over subsequently realized 1y returns, is optimistically ~.5% across a large cross-section of stocks. There are many other important considerations in estimating future prices, even over the long term, e.g. currently WA1's stock price sees a lot of prospective tailwinds in relation to index and fund inclusion. As another example: when Baosteel, Nippon steel and friends bought 15% stakes in CBMM in 2011 @ ~$2b a pop, do you think it was on the basis of DCF? Almost certainly not: they paid whatever was necessary to keep their seat at the table. Btw, one last point on least squares and DCF - Jim Simons sadly passed away a couple of weeks ago: what fraction of RenTec's forecasting infrastructure do you think was dependent on a) linear regression, and b) DCF. Anyway, happy to have you here if you are looking for serious discussion, and I am not a fan of the cultish mentality or silly price targets either.