Computational Investing Part I (Coursera) - Week 7

Week 7

Module 1

1.1

• Information feeds -- for event studies
• Thomson Reuters Machine Readable News
  • Machine readable news feeds
• StarMine
• InsiderInsights
• What to look for
• Historical data you can back test
• Survivor bias free
• Ease of integration with your sys
• Low latency
  • Get to you quickly so you can act on it.

1.2

• TA uses historical price and volume data only to compute "indicators"
• Indicators are calculated from recent price and volume data to predict future price movements
• Indicators are "heuristics"
• Discusses the controversial side of technical analysis
• An information source?
• Depends on info in historical price and volume
• "Market physics"
• Limit to how quickly market can move
• Tucker's view on Technical Analysis is interesting

1.3

• How to compute Bollinger Bands
• Read in historical closing prices
• mid = "rolling" mean over look back period (average over 5 week period, for example)
• std = "rolling" STDEV over look back
• upper = mid + std
• lower = mid - std
• Current "value" of Bollinger
• if current price at Upper band = 1.0
• if current price at Lower band = -1.0
• Val = (price - mid / std)