Comcast-Time Warner Merger

[hardblue.livejournal.com]

Yeah, but some of us like Internet porn.

Comments

[lafinjack.livejournal.com]

The graph in that article shows that the significant slowdown (below average) started in January.

[badlydrawnjeff.livejournal.com]

No, the significant slowdown started in December, and probably sometime in November. Graph reading isn't difficult. The net neutrality ruling was on January 14.

[lafinjack.livejournal.com]

Did you look at the graph? Here, let me draw a line on it for you:




The December drop is barely lower than the August recording. You're assuming that because there was a bump from September through November that the bump would continue. That would not be the case if it's just varying from month to month. The January reading is a significant drop, though, and a likely explanation for it is the January ruling.

[badlydrawnjeff.livejournal.com]

Now you're not being serious. November to December was significant. December to January even worse.

[lafinjack.livejournal.com]

Yes! Now we're finally talking data.


November to December was significant

Less than 0.1Mbps. As you can see, the original graph (http://i.imgur.com/tijZBiJ.jpg) is compressed to show only the relevant data points - i.e. the bottom isn't 0. You're assuming a huge jump in speeds when... it wasn't.


The drop from the baseline of August to December's low is... let me get my pixels out:




My claims:

1. September, October, and November were slightly faster than normal (~0.07Mbps per month on average), but an anomaly. Why? I dunno.

2. December isn't much slower than the pre-anomaly normal. Roughly 2.14Mbps to 2.12Mbps. 0.02Mbps. 20Kbps. 2.5KBps. Significant? No.

3. But when you look at the drop from 'normal' of August to January, the drop was ~2.14Mbps to ~1.82Mbps. 0.32Mbps. 320Kbps. 40KBps. Sixteen times the September-November bump. That is a significant aberration, and would line up with the mid-January ruling.

[badlydrawnjeff.livejournal.com]

Again, you're not being serious. There is no January->February data here. You can't compare August->December like that.

You're grasping.

[lafinjack.livejournal.com]

It's your claim that September-November was a significant increase in speeds from August. It was not.

It's your claim that December was significantly slower than normal. It was not, neither from September-November nor from August.

Yes, January is for now a single data point. But it fits what we know.

I've backed up my position, using your data. If anyone's grasping here, it's not me.

[badlydrawnjeff.livejournal.com]

It's your claim that September-November was a significant increase in speeds from August. It was not.

No, it's my claim that November to December was significant. Because it was. Read the graph, read my comments.

It's your claim that December was significantly slower than normal. It was not, neither from September-November nor from August.

Not my claim either. My claim is that December was significantly lower than November. Read the graph, read my comments.

Yes, January is for now a single data point. But it fits what we know.

Yes, it fits that we know that December to January was a significant drop, a drop that began November into December. That's what we know.

I've backed up my position, using your data. If anyone's grasping here, it's not me.

All you've shown so far is that you cannot read a graph or my comments.

[lafinjack.livejournal.com]

...it's my claim that November to December was significant.

Less than 0.1Mbps. A higher limit of 100Kbps. Less than 12.5KBps. Well within the monthly variance shown on your graph.


My claim is that December was significantly lower than November.

Less than 0.1Mbps. A higher limit of 100Kbps. Less than 12.5KBps. Well within the monthly variance shown on your graph.


If this were a climate change discussion you would be handwaving the September-November change as an aberration and calling December a return to near-normal, as it is not significantly lower than normal - look at the red line (http://i.imgur.com/tijZBiJ.jpg), as a statistician would. 2.14Mbps in August to 2.12Mbps in December. 0.02Mbps. 20Kbps. 2.5KBps. Significant? No. December to January? 2.14Mbps to 1.82Mbps. 0.32Mbps. 320Kbps. 40KBps. Significant? Yes.

[badlydrawnjeff.livejournal.com]

Compared to the changes we see otherwise, they're quite significant. That's the point. That's how you compare things.

[lafinjack.livejournal.com]

[badlydrawnjeff.livejournal.com]