Market Breadth EMAs V2

Second version of Market Breadth EMAs for $SPY. Getting a little more complicated than V1 but removed noise.
Key:

• Green line = % of stocks above their 20-period moving average, the "twitch line"
  
• Red line = % of stocks above their 200-period moving average, the "long term trend"
  
• White line = weighted average of the % of stocks above the 20/50/100/200 averages, the "general trend." Captures bursts that the 200 misses, and is more trustworthy than the 20.
  
• Background colors = limits of the red/green/white where reversals have happened historically. The darker the color, the stronger the signal.
  
• Histogram = the change in the white line over time, for different time periods: 1/4/10/20, the "trend strength/confidence." i.e. If the white line "General Trend" has been drifting lower for a month but started increasing the past 2 days, you might have 3 red histograms and 1 green one.
  

Techniques:

• If the green, red, or white line is above 50%, then more than half the stocks are above that average. So, if they're in the top half, bullish market. Bottom half, bearish market.
  
• If the green line is above the red, market has rising/bullish momentum. If red is above green, market has falling/bearish momentum.
  
• If the white line is rising, bullish momentum. If it's falling, bearish momentum.
  
• If the histograms are all green, there is strong momentum in that direction. The % of stocks above their important averages has been increasing each day for both the short term and long term.
  
• If the histograms go from all green to a mix of green and red, be on the lookout for a reversal from one of the background levels. Usually initiates from the 20 (green line) first.
  
• If price dips without the histogram changing, HODL.
  

版本注释:

Change: The green, red, and white lines were behind the histogram. Re-arranged them so they'd be in front.

版本注释:

Histogram bars overlapped when you zoom in and out. Switched them to columns and it's perfect now.
style=plot.style_histogram -> style=plot.style_columns

版本注释:

Changed picture accidentally, slight changes to code/labeling. Mostly just fixing my picture mistake