Interesting aspect about the models... As you know, BAM (Beta and Advection Model) has three species

1 BAMS; 850mb to 700mb steering depth near the edges where vortex opens to environmental flow. Essentially better for predicting track guidance with lower tropospheric events such as depressions.
2 BAMM; 850mb to 400mb steering depth is essentially better for predicting track guidance for TS'
3 BAMD; 850mb to 200mb steering depth is essentially better for predicting track guidance for hurricanes..

One must wonder if the establishment of TS intensity will somehow negate the inclusion of BAMS in lieu of the BAMM; and so on to BAMD if hurricane status is achieved.

The interesting aspect leaps out that the LBAR does not include the Beta Effect, nor does the statistical based climatology and persistence - "cliper", and those are the ones that in the 06z, 12z and 18z runs do not want to take Beta westward. It is almost suggestive that to remove beta effects allows the in situ albeit weak tendency to be the predominant steering mechanic, which based on the observations on those three runs of the global based models indeed is a tendency for NNW/N drift. In fact, speaking of the BAM runs...in that order above (1, 2 and 3), particularly in the 12z run we note that the lowest altitude advection parameters have the deepest latitude track, and the tracks bend slightly north albeit slightly, as the model that is run considers a deeper steering field (pressure means above)

In a way, that tells me that the west track could be dubious because it is based primarily on the absence of critical values deemed negligible, which is risky out in time. The CMC for example is also in the LBAR/CLIPER camp, and I'm fairly certain it doesn't use tricky considerations having to do with the physics of tropical disturbances (like the product of the beta parameter and meridional wind speed..etc)

Forgive me though because it's been awhile since tropical meteorology and these models were new when I was there. But, sufficed to say, at 18z the GFN - or interpolated GFS? - is also showing "somewhat" more north at 18z.

If nothing else...it was interesting to see the BAMS be ~1.5 lat S of the BAMM, which was ~1.5 lat S of the BAMD in the 3 runs. Underscoring this is that these have ever deeper mean steering depth respecitively, AND, the deepest (BAMD) is at least in the direction of none Beta Effected model runs.

It's probably not too cogent to use these observations in a meaningful correction because lets face it...deep layer steering field is weak so Beta Effect becomes large when scaling the influences. However, I still believe that because Beta's wind radii is relatively small, as she intensifies she will be less susceptible to Beta Effect; we may just see lest W tug in the models in the 00z run - if not, probably afterwards...

It is all hairy, because if I am right about this basic premise and supposition thereafter, it is entirely possible that Beta will drift more N and never make the W turn; then, we have an important full latitude trough timed for beyond 100hours, for the most part unanimously agreed. That ups the stakes farther N...
Actually, it should be noted that the LBAR is trending W in time. It wouldn't be for Beta Effects however; there would have to be something else imparting the trend..