
timeweighted return
treat each time period individually, then find the geometric mean

bank discount yield
(Face value  Current value)/Face value * 360/time to maturity

money market yield
HPR*360/time to maturity

dollar(money)weighted return
Exact same thing as IRR, just make NPV = 0

sampling error
difference between sample statistic and population statistic

Trading index (TRIN) or Arms Index
 measures trading volumes of advancing vs declining stocks
 less than 1 indicates more volume in advancing stocks


priori probability
a simple probability based on the population size. If a stock can go up, down, or stay the same, then using priori each would have a 1/3 chance.

cross sectional vs time series data
 cross sectional: sample taken at a given point in time
 time series data: sample taken over a period of time

parametric vs nonparametric tests
 parametric: many assumptions about the population distribution
 nonparametric: may be most appropriate when heavily skewed
 both: often used in conjunction with each other

trading volume, head and shoulder trend
head and shoulder trends are more validated when volume decreases along them, showing that supply and demand forces that were acting before are weakening

Kurtosis
> 3 => fat tails
leptokurtic = more kurtosis
platykurtic = little kurtosis

roy's safety first criteria
choose portfolio with highest (Return  threshold)/standard deviation

%K and %D, MACD and signal line
%K is (current low  lowest low)/(highest  lowest). %D is moving average of %K. K above D implies buy, K below D implies sell
signal line is moving average of MACD line. MACD < signal implies sell, MACD > signal implies buy

technical analysis
market moves (supply and demand) are based on rational and irrational behavior

