Hacker News Reader

Comment by rduchnik | Hacker News Reader

There is a comment on my blog from a developer from MLB.com for anyone who is interested:

"Hi -

Engineer from mlb.com here. It appears this goes a bit deeper down the rabbit hole than meets the eye. Apparently there’s some code laying around in our tests run by a CI setup that randomly generates a tracking code to mock third party scripts (Google analytics, ad tracking, etc) instead of using our actual IDs as to not mess with our marketing guys’ numbers (we run a LOT of tests on CI).

The strange thing is that your IDs aren’t being pulled from your site, but have randomly been generated the same way many, many times and then been shipped out to our production server by mistake.

We can’t figure out why this is happening, but are looking into the build system and how it caches data. Luckily I read HN or we might have never caught this!"

View on HN

cough bullshit cough, The statically probability that you would generate an actual AND proper tracking code that also just happens to be used by a site that shares all you code is less than can be described. Your team got caught outsourcing jobs to code copy/paste firms.

"Wow, what a coincidence... we just happened to randomly generate a google analytics ID that was the exact same as the one used on a scratchpad plugin that we copied the code from..."

If only calculating the odds of this were as fun as 2048 we'd have a bunch of different solutions on the frontpage of HN for the next 2 weeks!

And wouldn't it make more sense to create a test GA account and just use the ID from there for everything, instead of generating randomly each time?

I don't see any reason for the guy to lie. He could just have not responded, or quietly changed the code, but he came to the blog and posted a publicly viewable comment about it ... maybe his diagnosis is wrong and something different is happening, but I'm very willing to believe it was an honest mistake given the circumstances.

This is exactly the reason they responded. By changing the code quietly, they would implicate themselves. An official PR statement like this is obligatory.

What is more probable, that they a lying, or that the random generator generated not a single code that matched, which is very(!) improbable by itself, but two codes that belong to the same account, generated at the same iteration.

This is not an official PR statement, it's an engineer responding at 11 PM (MLB is based on the east coast) with the information he/she had available. I don't think there's an intent to decieve here, just a mistaken guess of the probabilities involved. Hanlon's razor applies.

He has no idea that his team is a bunch of lazy outsourcers. He knows, he is just confused.

I don't think he's lying, but I do think he's mistaken. They probably do generate random codes, but this was not one of them. It looks/smells like a copy/paste.

That aside, though, why not just create a different Google Analytics account and use that one consistently for testing?

Yeah, it could be their CI doesn't generate random codes voor javascript code or something...

The thing that most inclines me to believe it is that, by my reading of the article (which admittedly might be reading to much into it), only a subset of MLB pages are using the ID. Hard to imagine why that would be the case if someone was simply copying code wholesale from the "victim" site.

Everything about that is dumb. If you need to test without affecting your numbers just have an alternate ID for tests.

Hi MLB.com engineer, I thought I should remind you that you're dealing people who think, and perhaps programmers too. We see through it.

Reply on news.ycombinator.com