Mrs Trellis

Well-known member

• Aug 12, 2020

• #1

.......here's today's episode:-

https://www.bbc.co.uk/programmes/m000llw2

 

Mrs Trellis

Well-known member

• Aug 19, 2020

• #2

...and here's today's (the algorithm explained, covid testing, poker , and being hit by a bus):- https://www.bbc.co.uk/programmes/m000ls8k

 

Mrs Trellis

Well-known member

• Aug 26, 2020

• #3

..and here's today's 26/08/2020. Covid plasma treatment claims, early breast cancer screening and harlots (numbers thereof).

https://www.bbc.co.uk/programmes/m000lzgy

 

Mrs Trellis

Well-known member

• Sep 2, 2020

• #4

...and today's 02/09/2020. Back to school & uni., Covid test and trace, GDP, and why maths?

https://www.bbc.co.uk/programmes/m000m5j9

 

The Old Ruminator

Well-known member

• Sep 2, 2020

• #5

I actually like this program. Sometimes it shows how a single flawed assumption way back in the data can lead to a spurious result later on. Today's issue was using an assumption of R Rate of 3 in students returning to university. Ergo quite a nasty number of COVID cases further down the line. Whilst it may be near the truth it might well be miles adrift. There are statistics and damn statistics. Similarly, algorithms are only of use if the initial data is correct as you are using something to predict an unknown quantity in the future.

 

PeteHall

Moderator

• Sep 2, 2020

• #6

Before publishing any modelled scenario, it's always important to "sensitivity test" it. Basically make small incremental changes to the variables and see if the result is stable.

If the result is wildly different following a small change to one particular variable, you either need to be very confident in that variable, or you need to re-work your model until it is more stable.

I assume that all the published models have been checked in this way, but given how little we know about Coronavirus and the urgency to get these models published, I suspect that none of them are particularly stable or reliable...

 

Mrs Trellis

Well-known member

• Sep 9, 2020

• #7

...and today's 09/09/20 episode - Covid cases rising, a guide to life?s risks, and racing jelly-fish (and a shamefaced confession lol.)

https://www.bbc.co.uk/programmes/m000mb06

 

Mrs Trellis

Well-known member

• Sep 16, 2020

• #8

...and today's 16/09/20 - Covid testing capacity,  asylum refugee numbers, Covid in areas of deprivation, and mascara claims.

I noticed on huge error in today's discussion though - in the areas of deprivation piece Sale was described as having a very good rugby team. He obviously didn't watch last Saturday's game. 

 

Mrs Trellis

Well-known member

• Sep 16, 2020

• #9

oops - senior moment - here's the link:-  https://www.bbc.co.uk/programmes/m000mksw

 

Mrs Trellis

Well-known member

• Sep 23, 2020

• #10

..and today's 23/09/20 Covid curve queried, false positives, caffeine & pregnancy, and the Queen?s head.

https://www.bbc.co.uk/programmes/m000mr42

 

Mrs Trellis

Well-known member

• Oct 2, 2020

• #11

30/09/20 'Record' Covid cases, Trump on the death count, and ant pheromones - how topical is that? - the ant pheromones of course
https://www.bbc.co.uk/programmes/m000mywc

 

Mrs Trellis

Well-known member

• Oct 7, 2020

• #12

Last one in the current series (I can hear the sighs of relief) :- Spreadsheet snafu, ?Long Covid? quantified, and the birth of probability.

https://www.bbc.co.uk/programmes/m000n4vq