papers/resources about the math behind tagtime

[cdep.i...@gmail.com]

Hi,

I'm wondering if there are any good resources (papers, books, etc) about the math that tagtime is based on.

Briefly looking through the code, I have a couple questions and I think they could be answered by understanding some of the math behind tagtime:

1) What do you get by having every user of tagtime on the same ping schedule?  Does this workout nice mathematically somehow?

2) In cntpings.pl and when syncing to beeminder, it looks like every ping is given 45 minutes worth of time.  Is this correct?

Is this how sites like tagtime-trends are implemented as well? https://alexschell.shinyapps.io/tagtime-trends/

It seems like some days would have much less than 24 hours of activity logged.  If you were generating graphs like the site above, wouldn't this throw them off?  Is there any solution to this other than smoothing out the graph?

3) I'd like to learn more about the math below which calculates confidence intervals for time per tag.  Can you point me to any books/papers/resources?  Or even just general concepts?

https://groups.google.com/d/msg/tagtime/MqI44eIyFkE/ccCbOM4FpJkJ

Thanks!

Dennis

[Andrew Quinn]

Response to 1) and 2)

I think the best place to look for a quick-and-dirty explanation of the math is here: http://messymatters.com/tagtime

But as I understand it, it uses the Poisson distribution, and the 45 minute mark is just a nice convenient number. You can easily edit it to ping at, instead, 15 minutes or 90 minutes, but the more often it pings the more likely you are to get annoyed and stop using it.

The Poisson distribution is essentially a really nice lil' thing that could theoretically take edging-towards-infinity number of minutes for you to get a signal, but in practice it averages out to sending a signal every 45 minutes. I imagine someone who took a calculus-based statistics or probability theory course could tell you a lot more about it than I can!