streaks vs statistical streaks

Hacker News et al are obsessed with streaks, but I think they have some problems:

1. A single regression resets to zero.

2. There’s not an easy way to gradually ramp up your streak-commitment over time.

I prefer a different approach: statistical streaks.

Suppose I made a commitment to do something differently on 2016-08-26, and did it for the next 5 days; then my 30-day statistical streak avg = 0.166, but my 5-day statistical streak avg = 1.0.

If I then have one off-day (5 consecutive days then one failure day) the 30-day statistical streak avg is still 0.166, but the 5-day statistical streak avg drops to 0.80.

If I pick it back up the next day, it goes to 0.2 and 0.80 for the 30- and 5-day statistical streak averages (respectively), while if I fail again, it drops to 0.1667 and 0.6 for the 30- and 5- variants, respectively.

Here’s a tiny bit of python to demonstrate these scenarios:

    assumed_history = 0
    explicit_recent_history = [1,1,1,1,1,0,0]

    for window_days in [30, 5]:
        # lazy...
        history = [assumed_history]*(window_days)+explicit_recent_history   
        recent_history = history[-window_days:]
        success_days = sum(recent_history)

        print window_days, float(success_days) / window_days , recent_history