osheaf01.bsky.social

If anyone is in Providence tomorrow and has a bit of spare time, I am doing a lunchtime book talk at Brown. Please come along.

He's talking about correlation.

He's implying or inferring causation, because her presence is correlated with all the "suspicious" events.

That is his meaning, in plain English.

"That wasn't the prosecutor's statement."

BBC report on opening day of trial. ⬇️

Care to read the bit under the image for me? Slowly now, so as not to tax your levels of English comprehension...

But immunoassays are definitive, right, and anyone claiming otherwise is a "conspiracy theorist"?

I'll repeat the question: is everyone who dares to question your worldview a Conspiracy Theorist? 🤭

So the babies died, then?

You didn't address the issue that further testing wasn't done (of course!)

It's stating that correlation - Letby's presence at "suspicious" events - equals, infers or implies causation.
Correlation is statistical, is it not?
It's saying her presence cannot be by chance - the prosecutor referred to her as the "common denominator."

Well, have a question.

Why did they not avail of the services of the Warwick Professor of statistics?

Well, I asked AI to fact check your assertion that immunoassays were definitive, not indicative, as to exogenous insulin administration...the results didn't surprise me.

I strongly suggest you acquaint youself...

May well have been...ya think?

Yawn.

No, they spent 10 months working out explanations, with expert in same, Dr Evans.

Really?
But they were called in by the doctors to investigate Letby, based on Brearey and Jayaram's suspicions, as voiced in subsequent interviews. Hence, the internal fight to get her removed, which is what Thirlwall is all about.

No individual mentioned?

Oh dear, ad hominem.

Lost the argument, have we?

I'll go with the Prof of Stats in UCL and my own judgement on that one, thanks.

Guess I must be a "conspiracy theorist" for daring to disagree with you? 🤭

So, her presence circumstantially proves...what?

She was around when "suspicious" deaths occurred, so she must have been doing something?

This is identical to the Lucia de Berk case, isn't it?

But most murders are unquestionably murders. Therefore, one's presence at the scene, forensically proven, is evidence. Correlation is causation.

Whereas, these cases are not unquestionably murders, and Letby's presence is proof of nothing, as she was where she was contracted to work.

And I'm a computer scientist with a maths and engineering background, who can see at a glance that that chart, shown to the jury, is a complete crock of shit.
If that's the standard of evidence in an 18 month trial costing millions, you have to wonder what else passes muster.

Which is where I came in: that chart is a complete crock.
If that passes evidentiary standards, you'd have to wonder how subterranean they are?

Actually, most crimes do have witnesses, or clear forensic evidence and, most of all, it's clear a crime actually occurred.

What this case has in common with Horizon IT is that that's the actual issue.

Really?

"Sounds like my kind of case."

What is your expertise in statistics?

You say they weren't used, he says he does.

Is *everybody* who disagrees with your view a "conspiracy theorist"? 🤭

All these suspicious cases, and she's always there (but no-one actually saw anything), is a statistical argument, claiming correlation is causation.

Have to say I'm amused that this started with you slagging off antivaxers for their own versions of reality 🤭

Not suspicious, of course, because we already have our Prime Suspect, from Day 1 of the investigation.

It's only a matter of "working out an explanation" for how she nefariously did it, thereafter.

And who better to do that that Dr Evans...

John O'Quigley, Professor of Stats at UCL, publishing his crank theories in that conspiracy theory journal, the Daily Telegraph

Alibi?

She was at work when she was contracted to be at work.

Common denominator is an explicitly statistical statement.

No, I don't think so, and, more to the point, the CPS ordered the police not to involve a Warwick statistician, which really ought to have been done.

Just as the immunoassays really ought to have been sent for further testing.

It's like getting heads 27 times in a row with a coin. The p value for that is infinitesimally small.
But is the coin fair, or rigged?

So you've really asked the wrong question, haven't you?

And as for it having no evidentiary value?
The jury were shown the chart.
Prosecutor Johnson kept calling her the "common denominator", the "*constant* malevolent presence".
This is a statistical argument, and to claim otherwise, as you do, is intellectually dishonest.