[delinka]

That's not a prediction of the outcome. That's a statistic. A prediction of the outcome would be "you won't roll a 6."

[Jarwain]

The article doesn't claim that Cambridge Analytica predicted the outcome. The only factual claim about the outcome made by the company, according to the article, is that statistic.

The article, however, took that statistic and implicated it as a prediction when all it really was was a statistic.

[dagw]

If you asked me to predict whether you will roll a six or not, then I will predict that you won't. Is that a bad prediction?

[delinka]

It's certainly an incorrect prediction if I roll a six. You didn't simply state the odds and leave it at that, you actually made a prediction. Is it a "bad" prediction? Depends on how you're defining "good" vs. "bad" predictions. If the quality of the prediction is not based on the outcome (i.e. whether the prediction is correct), but is instead based on the odds themselves, then that's a conversation that leads to "bad but correct" predictions.