I looked at this over the weekend, and thought the following migh be of interest - only my personal look mind.

The first chart is for the commercial COT gold positions since 2001 (in 1000s), for their max and min short positions, their maximum long position, and then, how much they moved their short position from min to max. I've added a 3polynomial trend line to each one.

In 2001, at the start of gold's bull run, max short & long positions were in the same ballpark, subsequent to which the bias has grown strongly and steadily in favour of a large short position.

To the mid 2000s, they could expand their short positions over longs by 3, 4 or 5x (up to 500% in 2003). But since then their minimum short position level has steadily increased (through inabaility to cover in a bull market?), seeming to increasingly limit their ability to pile on large percentage increases - in 2009 they were limited to a 73% increase bottom to top (brown line, or difference between the red and teh orange lines - 2010 in progress).

At present the limiting factors seem to be a ratio of 4.5:1 max short:long position (second chart) and a total short position of around 400,000 contracts.

If this present reduction in positions is stalling out at around 300k, and the chart trends hold, then the next big short position might be expected to reach 500,000 to 600,000 (75-100% increase). To get the 300,000 down, wouldn't they have to kill the price even more, which might be difficult?

It does look like this chart is heading towards a squeeze if a big upward price move comes along, especially if position limits come in (March meeting?) and they can't go higher with their short/long ratios.

One thing that seems to show on this chart, is that the commercial longs might also be used to help nudge price down, ie the second green peak and max short was acheived within the 4.5 ratio, but only by increasing the longs, which were reduced through the price toppling. Also, the increase in longs early last year, with a sharp subsequent removal, seemed to get a little bit more out of the larger downmove. I was thinking, an increase in their longs now, and timely removal, might get the price a little lower, but i don't know. I wanted to contstruct chart 2 for the whole decade, but couldn't easily get the data - and it would take too long to manually enter it.