Dealing with Data

Question 1

tick

Answer 1

finest resolution of data

represents a successful buy/sell match

Question 2

OHLCV

Answer 2

open - price at start of time period
high - highest price in period
low - lowest price in period
close - price at end of period
volume - total volume

Question 3

stock splits

Answer 3

generally because price is too high
appears as a price drop

solution is adjusted close
Divide closing price by stock split amount
eg
adjclose = actual close after split
adjclose = actual close / 2 for time before a 2:1 split
adjclose = actual close / 4 for time before 4:1 split etc

Question 4

dividends

Answer 4

cash payout from the company

adjust stock price by proportion of dividend
eg
adjclose = actual close after dividend
adjclose = actual close adjusted down by dividend amount

Question 5

survivor bias

Answer 5

only considering live stocks from the benchmark universe (eg current s+p)

this inherently favors all stocks that did well in the past, and ignores those that shutdown.

Need to use bias free data