start time: 2015-08-25 00:37:00 -0500
stop time: 2015-08-25 02:03:00 -0500
duration: 1+ hours (with a 10 minute internet outage)

peak twitch viewers: 16 pre internet outage, 6 post outage
time to reach 10+ viewers: 21 minutes
number of followers at end of stream: 110

FUUUUUUUUUUUUU.

10 minutes of downtime was all it took to completely lose all of the momentum the stream had. That's how long I was offline before internet access came back.

See, here's the problem. Twitch only sends out notifications for when you initially go live. After that, it takes however long to reset before it will send out notifications again, no doubt so that people don't get spammed by "so and so has gone live" messages for someone who keeps stopping and starting their stream.

So my internet cut out, and I lost all of my viewers (save for those I have other means of contacting), and then there was nothing I could do really to notify the others that I was back online. It sucked. I had some new people show up who seemed pretty cool and it was looking to be a good night. But after that initial loss of momentum, my viewer count was low enough that I knew that I was just going to be buried deep in the Outlast directory and was unlikely to go anywhere with the rest of the stream. So I decided to just call it a night and spend that energy streaming at some other time.

What kind of sucks is this was my first night with a new statistics script. Rather than rely on the PS4 viewer counter, or Chatty's imprecise viewer logs, I built my own thing which scrapes my twitch stats approximately every minute. My plan is to eventually graph it to look for patterns (specifically, does there appear to be a good time of night to stream X game). But for now, I get to use it to make it much easier to track things like:

  • peak viewers
  • average viewers
  • time to reach X viewers (like how I'm currently tracking the time to reach 10+ viewers)

This stream was also chock full of 30 minutes of mic testing. I left it in the highlight somewhat because "why not" but also it was an alright demonstration of one random dynamic microphone versus another random condenser microphone and what happens when you have a box fan turned on in the same room.