Thanks

, my understanding is that the higher the ballistic coefficient is the accurate the RV is.
This is largely correct, but it partially misses some important caveats.
Broadly speaking, the higher the ballistic coefficient (beta), the lower the amount of drag there is on the RV. As a consequence, as the beta increases, the amount of deceleration due to atmospheric drag decreases, which means that higher beta RV designs will spend less time in the atmosphere, will reach the ground faster, and will have higher velocities when they reach the ground (or the target detonation altitude).
This does not directly translate to improved accuracy. Instead, the improvement comes from how the higher velocity leads to a reduced time of transit through the atmosphere. Thanks to those two factors, the effects of high-altitude winds on accuracy are greatly lessened, as the RV has a much shorter window of exposure to them, and is more resistant to being pushed off course by said winds. While they are still a significant source of error, the amount of error they can introduce is significantly lessened relative to older designs with lower beta RVs.
However beta alone is not the be all and end all of accuracy. Improved RV designs with higher betas are usually co-packaged with other accuracy-improving technology, like improved nose tips, improved fuzes, improved radars, etc.
Nose tips are particularly critical, as it is immensely important that the nose tip can withstand the exceptionally high stress environment of high-speed reentry while wearing down as close to perfectly evenly as possible. This is far from a trivial task, and there are extreme challenges that they must be able to handle. Something as simple as raindrops turn into a lethal threat when you're entering the atmosphere at Mach 25, and while the RV will decelerate significantly, raindrops are still incredibly damaging at lower (typically still double digit) Mach numbers. They will quite literally eat holes into the nosetip as if it had bullets fired at it. The greatest threat is asymmetric ablation, which can send a RV far off course.
Improvements in nose tip design alone can eke out dramatic improvements in accuracy even without changing anything else about the design of the RV.
Similarly, fuze and radar design is an area where there was tons of room for further improvement. Much of the accuracy improvement in higher-beta RVs was actually due to improved fuzes and radars. The Mk21 RV uses radars of a totally different design from the previous Mk12 and Mk12A warheads. The Mk5 RB introduced the radar updated path length (RUPL) fuze, more commonly known as the "super fuze". Both the Mk21 and Mk5 introduced a wide variety of major improvements to the design of their fuzes relative to previous generations of RVs/RBs.
In the end, while it's true that all else being equal a RV with a higher beta will outperform a RV with a lower beta on accuracy due to atmospheric reentry errors, this is not the full picture, and it's rare for all else to be equal.
Even more importantly, you can backport improvements in nose tip and fuze technology to older lower beta RV designs to get dramatic improvements in accuracy without increasing the beta.
Just look at the W76, where:
- The Mk5's RUPL fuze ("super fuze") has already been backported into the W76-1 via the Mk4A RB in order to provide a substantial accuracy improvement.
- An improved nose tip design (the Shape Stable Nose Tip (SSNT) that was originally developed for the Mk5) is scheduled to be backported to the W76 RBs in a upcoming mod that will create the Mk4B RB (which is basically just a Mk4A RB that has had its nose tip replaced with the new SSNT). The new nose tip will significantly reduce reentry errors, and should therefore result in another substantial increase in accuracy.
These changes are introducing technology that was originally developed for ultra-high-beta RVs via the ABRES program, which resulted in the ABRV, which in turn was developed into the Mk21 RV (beta = 3000) and Mk5 RB (beta = 2500). And yet now this technology is being backported into the Mk4 RB (beta = 1800).
Sure, the Mk4 will never gain the improved resistance to high altitude winds that the Mk5 and Mk21 have, but it will gain enough improvement in accuracy from backporting the fuze and nose tip designs that the difference in beta won't really matter as much anymore. The introduction of the RUPL alone was enough to give the Mk4 a credible hard target counterforce capability despite its much lower yield. The introduction of the SSNT will only further enhance the lethality of the Mk4.
So yes, higher beta designs do tend to have better accuracy, but it's not just from the higher beta. It's from the whole package. If you threw a Mk4's fuze and nose tip into a Mk5, it'd have pretty atrocious accuracy despite the far higher beta. Heck, the Mk21 lacks a RUPL, so even though it has even higher beta than the Mk5, it's at a disadvantage from a fuzing perspective, and therefore it is significantly less accurate than it could be. The Mk21A mod will integrate a new fuze with RUPL capabilities into the Mk21, which promises to substantially increase the accuracy of Mk21 RVs without any increase in beta.