It is difficult to make sense of the hit points in the new Monster Manual.
Looking at the average of the official hit points at each CR, the rate of improvement sometimes accelerates, sometimes decelerates, at certain points reverses so the next CR actually has less hit points, and finally at epic CRs (20+) wildly accelerates deviating from any pattern.
The simplest formula to explain the numbers is: hit points = 15(CR+1).
In other words, CR 1 starts at 30 hit points, and after that, each higher CR improves by 15 hit points.
This simple formula closely matches the actual average hit points from CR 1 thru CR 10. Overall, the formula stays true even up to CR 21 where an expected 330 hit points matches the actual 335.8 hit points on average. In the CRs in between, the formula also helps shore up the anemic CRs. For example CR 14 averaging about 200.5 hit points, should instead expect about 225 hit points, which is at least a bit better. CR 17 which averages 254 hit points, should instead be about 270 hit points.
However the simple formula fails to explain the epic CRs. From CR 22 thru 24, the improvement to hit points lurches upward derailing from any formula, improving about 70 hit points per CR on average. Because the epic CRs tend to have fewer monsters, and many of them are unique individuals, it is fair to say the small sampling pool is nonrepresentative of any expected CR. On the other hand, these wildly higher hit points at the upper CRs, might instead suggest, the hit points should have been improving at higher rate all along at every CR.
The Hit Points table below compares the formula, hit points = 15(CR+1), to the actual hit points in the 2024 Monster Manual. Actual hit point averages that are less than the formula by 10% are in red. Actual averages that are more than the formula by 10% are in blue. The formula is solid and helpful, except at the very highest CRs.