Good questions.@nordesmic has answered it very well, so i'll...

Good questions.
@nordesmic has answered it very well, so i'll just add around the edges.

Re: GDX
- They needed to reduce NST in the rebalance given our relative underperformance compared to the larger cap members of the in fund (this means that as we recover they will need to add back, given I expect us to appreciate much faster than NCM)
- NST was 3.77% of GDX as of 28/2 and has reduced to 3.1% as of 17/3.
- Comparing the latest VE holding number to the last SSN, shows they've sold around 1.4m shares in the past month.

There is a big momentum factor with ETF's given, the flow of $ into the funds will drive incremental buying of the constituents. This is accentuated, given macro funds tend to invest on a factor basis (not fundamentals), of which momentum is a key one, which will see the money flow one way a lot of the time, this benefited us well last year with strong flows into Gold funds and has hurt us this year with the outflows, GDX was 17% in the first 2 months of the year.

I agree with the sentiment that once all this finished is done, it should be positive (unless gold tanks and we get further negative fund flows), we're well and truly oversold and look like a bargain at the moment IMO.

The issue in the CBA example in the VE doc you shared is that it assumes a static weighting (20%) for rebalancing dates, GDX and GDXJ, are different in that the weightings shift and are biased towards the larger market cap constituents, i.e. the bigger get allocated more money, or the better performers get more $$, so it does form a bit of a positive feedback loop which is amplified when there is above system inflows into the ETF as well.