Applied Nate Silver – misleading charts

It’s interesting how trendlines with low correlation coefficients are, in fact, almost useless. Here is yesterday’s short-term trend:

weekly23-190s1

Downward slope (loss) but at the low rate of 1.1lbs/week, but notice the low r^2. Now we add today’s data point and drop the oldest one and get:

weekly23-190s1-1

This looks like big improvement, slope is much greater (more loss) and r^2 is much higher, probably high enough to consider the trendline as real.

But what actually happened? I add a new point which is a small gain and then rolled-off the older point which was a significant deviation (and lower). So any “improvement” is an illusion. If anything the point I rolled off and the point I added are the same, within statistical significance, but this creates the illusion of a steeper drop by removing an earlier “good” point.

This is very frustrating. While having an absolute increase (rather than just stalling) while sticking to extreme diet and extreme exercise was worse (when that happened four weeks ago and freaked me) this resistance to any further loss now is worse. I’ve actually increased the activities that should make a difference and the weight stubbornly persists as unchanged. Without any positive feedback, from rather severe diet and exercise, it’s hard to keep it up without any reward.

And bogus statistical analysis isn’t helping.

 

Advertisements

About dmill96

old fat (but now getting trim and fit) guy, who used to create software in Silicon Valley (almost before it was called that), who used to go backpacking and bicycling and cross-country skiing and now geodashes, drives AWD in Wyoming, takes pictures, and writes long blog posts and does xizquvjyk.
This entry was posted in comment, musing, rant and tagged , , . Bookmark the permalink.

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s