The formula for plastic tipped bullets was also developed by Don Miller, who developed the formula for untipped bullets that the Berger calculator uses. I have found it to be a reliable indicator regarding twist rate requirements. By comparison, the Berger calculator predicts a lot of twist rates will not work with tipped bullets, when they really work great.
I suspect there is another error inducing factor around too for bullets like light 22s, sixes and suchlike. The Berger stability calculator is for boat-tail bullets and it states specifically that it is overly pessimistic with flat base types. Comparing Miller/Berger against the twist rates Berger quotes for its various FB types, they're out by 1-inch or more.
Although the 55gn NBT is 'officially' boat-tail, it's such a measly little thing I barely consider it as such and regard it as an aid to bullet seating rather than an external ballistics advantage. Likewise when it comes to stability, it'll have less effect on the bullet's needs than the 'proper' long angled tail sections you find on long-range 6mm designs.
If my hypothesis is correct, it also applies to many other lighter small calibre match models such as the 52/53gn Sierra and Hornady .224s and similar. Whenever I've run them through Miller, they're barely stabilised in commonly used slower twist barrels. At 3,200 fps in standard ballistic conditions, the Sg for the 224 52gn Sierra MK is only 1.05 in a 14-inch twist barrel, yet they're use very effectively in .222 Rem rifles with this twist rate.