The K factor giveth - but not necessarily taketh away

The January 2014 FIDE Rating list contained the usual familiar facts (Carlsen number 1, the gap not really closing, juniors improving), but also at least 1 surprise. The spectacular jump of Parviz Gasimov, an Azerbaijan junior, from 1949 to 2517 in the space of three months, has resulted in some excitable commentary. While a number of people have questioned how this happened, the simple explanation is the new FIDE K factor of 40 for players who have played less than 30 games, or are under 18 and are rated below 2300. Gasimov seems to have simply timed his run well, reaching 2295 on the December 2014 rating list, before picking up another 220 points for the January list.
While there seems to be a bit of 'something should be done' in the air, I personally do not think this is a defect in the system. One of the results that Gasimov achieved was a 12.5/15 in a tournament that was not particularly strong (a lot of 1900's and 2000's). But as he managed to beat almost all the players with that rating it is valid to say that he must be significantly stronger than they are.
A similar outcome occurred a number of years back under the old Australian rating system (ELO with bonus points). Pat Halpin played a series of events in Queensland, and did very well (he scored 75%-80% IIRC). The fields were mainly around his own rating (2100) but as performed consistently well, his rating jumped to 2350, which mad him the 5th highest rated player in Australia. Despite fielding calls from concerned chess players (as I was the ACF Ratings Officer at the time), the system was not changed, and Halpin's rating soon dropped back to a level that other people found acceptable (and reflected his subsequent performances).
I'm pretty sure the same thing will happen with Gasimov, although he may benefit from a few extra tournament invitations in the meantime.

(BTW If you want to catch up with the new FIDE Rating Regulations they are here. Anyone planning to organise or play in a FIDE rated event should have a read)