WBAUnofficial
FAQ Statistical biffs (non football) - Printable Version

+- WBAUnofficial (https://wbaunofficial.org.uk)
+-- Forum: WBAUnofficial (https://wbaunofficial.org.uk/forumdisplay.php?fid=3)
+--- Forum: WBAUnofficial (https://wbaunofficial.org.uk/forumdisplay.php?fid=4)
+--- Thread: FAQ Statistical biffs (non football) (/showthread.php?tid=27236)

Pages: 1 2


FAQ Statistical biffs (non football) - SophLad - 10-19-2022

What does Delta mean?  Used in the context of two (different but similar trend forecasts)  I've googled but, tbh, I can't understand the answers and have 8 minutes to work it out! Ta.

Plain English if possible please.


RE: FAQ Statistical biffs (non football) - Birdman1811 - 10-19-2022

In statistics, the delta method is a result concerning the approximate probability distribution for a function of an asymptotically normal statistical estimator from knowledge of the limiting variance of that estimator.

Is the Wikipedia definition.

I realise now it's been a long time since I studied mm aths at that level and I'm struggling to break it down into an easy explanation.


RE: FAQ Statistical biffs (non football) - SophLad - 10-19-2022

(10-19-2022, 03:14 PM)Birdman1811 Wrote: In statistics, the delta method is a result concerning the approximate probability distribution for a function of an asymptotically normal statistical estimator from knowledge of the limiting variance of that estimator.

Is the Wikipedia definition.

I realise now it's been a long time since I studied mm aths at that level and I'm struggling to break it down into an easy explanation.

Kind of a blend of standard cause and special cause deviation?


RE: FAQ Statistical biffs (non football) - Birdman1811 - 10-19-2022

It's basically a function for deriving a trend in data, apparantly it only works in a pretty normal data spread since it heavily downplays outliers leading to a negative bias when they're present.


RE: FAQ Statistical biffs (non football) - SophLad - 10-19-2022

(10-19-2022, 03:16 PM)Birdman1811 Wrote: It's basically a function for deriving a trend in data, apparantly it only works in a pretty normal data spread since it heavily downplays outliers leading to a negative bias when they're present.

That makes sense in the context.  I'll chuck in the 'negative bias' line and look like I know what I'm doing.  I suspect most on the call won't know what it is either!


RE: FAQ Statistical biffs (non football) - Borin' Baggie - 10-19-2022

Change/variance in.


RE: FAQ Statistical biffs (non football) - SophLad - 10-19-2022

(10-19-2022, 03:18 PM)Borin' Baggie Wrote: Change/variance in.
Huh?


RE: FAQ Statistical biffs (non football) - Borin' Baggie - 10-19-2022

(10-19-2022, 03:19 PM)SophLad Wrote:
(10-19-2022, 03:18 PM)Borin' Baggie Wrote: Change/variance in.
Huh?

That's what it means, the delta is the change/variance in a function or value.

That could be either through a derivative function - such as change in speed relating to distance - or variance from the mean to outline error. The latter seems more pertinent in your example with the context specifically in regards to either the difference between the two curves or the variance from the mean of both curves.


RE: FAQ Statistical biffs (non football) - SophLad - 10-19-2022

(10-19-2022, 03:26 PM)Borin' Baggie Wrote:
(10-19-2022, 03:19 PM)SophLad Wrote:
(10-19-2022, 03:18 PM)Borin' Baggie Wrote: Change/variance in.
Huh?

That's what it means, the delta is the change/variance in a function or value.

That could be either through a derivative function - such as change in speed relating to distance - or variance from the mean to outline error. The latter seems more pertinent in your example with the context specifically in regards to either the difference between the two curves or the variance from the mean of both curves.

Perfect, cheers


RE: FAQ Statistical biffs (non football) - Birdman1811 - 10-19-2022

Knew someone would know better than me.