And that also became, de facto, the top score in the gifted validation sample, using the standard norms. That is to say, we have no idea whether that 151 was really 151, or should have been 161 or 181. The idea of the exercise with extended norms is to more finely distinguish all the examinees who bunch up at between about 140 and 151 (if you factor in the confidence interval). If you look at the increased spread from standard to extended norms (e.g., FSIQ top score goes from 151 to 159, VCI from 155 to 188), clearly not all 150s are created equal. Also, extended norms are a way of rescoring the raw data into scaled scores above 19. It overlays the standard norms, rather than purely replacing them (you can think of the 151 on the standard norms as 151+). The rule about needing two max scaled scores is supposed to restrict use of extended norms to those children whose confidence intervals include the maximum Index score under the standard norms. But I grant you that the extended norms have more clinical utility than strict psychometric robustness. I'm sure there was a fair amount of fun with curve smoothing involved.
Pearson also really had to release some technical guidance, because people were starting to make their own rules for this. Plus, more cynically, the Stanford-Binet LM had been clinging to the gifted eval market for a long time, and part of its appeal was undoubtedly the possibility of scores in the 200s. The extreme agedness of the LM created a hole in the market, which had not been definitively captured.
I hope what I just wrote makes sense, because it is definitely past my bedtime!
