Question about Kurtosis

[luso]

I have some questions about Kurtosis and leptokurtic / platykurtic
I found different books said it differently.

for Wikipedia and Schweser's FRM study guide and some other books
Leptokurtic = positive (or >3) and it is peaked at the mean and with fat tail
Platykurtic = negative (or <3) and it is less peaked at the mean and with thin tail
Kurtosis - Wikipedia, the free encyclopedia


But for some other books i.e. The Mathematics of Options Trading
The Mathematics of Options Trading - Google Book Search
says differently about the tails:
Leptokurtic => long and skinny/thin tail
Platykurtic => short and fat tail

I feel the second one is more convincing, do u guys have any idea?

 

[joe_bradley]

I presume it depends where you look in the tail. You have to look outside a few standard deviations. If you look at a leptokurtic distribution and compare it with a platykurtic distribution, the tail does indeed look skinnier - close to the mean. But if you go further away, the tail doesn't decay nearly as quickly as it does for a standard curve. You have this tiny little bit that keeps going. That tiny little bit is what causes those "once in 10,000 year" events that happen every 5 years on wall street.