Credit deflation and the reflation cycle to come (part 2)

[Majorpain]

7 hours ago, sancho panza said:

What matters is that he got his predictions badly,badly wrong.

The Swedish epidemiologist said a couple of months back,words to the tune of,

'never has one non peer reviewed paper had such an effect'.

I think Ferguson should release the full code and the methodology of how he arrived at his conclusions,so that he may be peer reviewed in good order buy the representatives of the taxpayer's who's funding he has benefited from.

Whilst not solely due to Freguson ,people have had cancer treatments suspended,elective operations cancelled and more generally,the nation's health has really suffered due to the excessive focus on Covid.I think we're owed some genuine reflections on what went right and what could have gone better.

 

Completely agree, the media response to the release of the 500,000 figure was off the charts, i don't think any politician would have been brave (or stupid) enough not to have gone into full lockdown.  Any rational analysis about "other deaths" from loss of jobs and lack of healthcare as you highlight was completely swamped by "the virus is going to kill us all" hysteria, he really shouldn't have publicly released something so incendiary without being at least minimally peer reviewed IMO.

[jamtomorrow]

6 hours ago, sancho panza said:

What matters is that he got his predictions badly,badly wrong.

That's certainly the received wisdom, but I'm mystified as to why. This is Table 4 from the March 16 paper, final column arguably corresponds most closely to the lockdown we eventually experienced in UK:

Actual excess deaths in UK has been 62,000 to date, which means final 2 columns are an over-estimate, and CI_HQ_SD column is either an under- or over-estimate depending on your choice of R0 and intervention timing - source: https://www.ft.com/content/3c53ab12-d859-4ceb-b262-f6a0221ca129

Here's the paper: https://www.imperial.ac.uk/media/imperial-college/medicine/sph/ide/gida-fellowships/Imperial-College-COVID19-NPI-modelling-16-03-2020.pdf

None of this is to say I'm happy with the software engineering processes and practices followed by the IC modelling team - I'm not, they were atrocious and it's sheer bloody good fortune if the predictions were anywhere near correct. Completely inappropriate given the importance of the work, and a sacking offence in my view.

[MrXxxx]

8 hours ago, The_Doc said:

There is also a simple reason as to why running the code with the same input files could produce different results. Monte Carlo simulations are used in epidemiology to create forecasts for the spread of a virus. Each Monte Carlo simulation run will produce one result and will use randomisation. As it uses randomisation, running a second time will produce a different result. 

Which is why you use a burn-in period (that you delete) and run sufficient iterations to allow proper heating of your model...inconsistent results are no better than an educated guess!

[MrXxxx]

8 hours ago, Knickerless Turgid said:

Thanks, Harley, I think I am even more confused than ever...!

I empathise!...keep at it though as after a while you get a `Eureka` moment and some bits fall into place :-)

[MrXxxx]

8 hours ago, JMD said:

All excellent points you make Cattleprod. Btw, the 'settled' bit was me being critical of the current zeitgeist, but accept nothing is new under the sun, eg Newtonian science was also settled until Einstein came along. But I did rather stupidly neglect to mention my main point, which was really to state again how blatantly political the scientists were at those Corona briefings. Yes, the government scientists were constrained in what and how they could say things, but I don't mind admiting that i found it a freak show and it horrified me because it all smacked of Orwell's 1984, ie so much propaganda, technocrats and newspeak on parade.

They have been taken to one side and `warned` about their future funding...even senior academics will `prostitute` themselves/their beliefs for money!

[MrXxxx]

7 hours ago, sancho panza said:

took little Miss Panza to Leicester market this morning and we stopped by the pound shop(High St) to get some stuff.Only two tills running for social distancing(ironic really,as it meant there was huge back up of customers in the shop), and the flow through isn't enough to sustain those businesses given how wafer thin their margins are where they're currently sited

I love this (your financial/critical thinking education) education of `The younger Panza`s, it's what my Father did with us as children regarding financial cons, and its probably the most useful thing he gave us...pity more don't do this with their children.

[Harley]

8 hours ago, Knickerless Turgid said:

Thanks, Harley, I think I am even more confused than ever...!

I have a knack!

[MrXxxx]

7 hours ago, sancho panza said:

What matters is that he got his predictions badly,badly wrong.

The Swedish epidemiologist said a couple of months back,words to the tune of,

'never has one non peer reviewed paper had such an effect'.

I think Ferguson should release the full code and the methodology of how he arrived at his conclusions,so that he may be peer reviewed in good order buy the representatives of the taxpayer's who's funding he has benefited from.

Whilst not solely due to Freguson ,people have had cancer treatments suspended,elective operations cancelled and more generally,the nation's health has really suffered due to the excessive focus on Covid.I think we're owed some genuine reflections on what went right and what could have gone better.

 

The problem is either a) he doesn't understand it himself, or b) its badly wrong but he `signed it off`...either is as a result of senior academics farming off their research to postdocs whilst promoting themselves, rather than nurturing them and being more `hands on`....

[MrXxxx]

30 minutes ago, jamtomorrow said:

That's certainly the received wisdom, but I'm mystified as to why. This is Table 4 from the March 16 paper, final column arguably corresponds most closely to the lockdown we eventually experienced in UK:

Actual excess deaths in UK has been 62,000 to date, which means final 2 columns are an over-estimate, and CI_HQ_SD column is either an under- or over-estimate depending on your choice of R0 and intervention timing - source: https://www.ft.com/content/3c53ab12-d859-4ceb-b262-f6a0221ca129

Here's the paper: https://www.imperial.ac.uk/media/imperial-college/medicine/sph/ide/gida-fellowships/Imperial-College-COVID19-NPI-modelling-16-03-2020.pdf

None of this is to say I'm happy with the software engineering processes and practices followed by the IC modelling team - I'm not, they were atrocious and it's sheer bloody good fortune if the predictions were anywhere near correct. Completely inappropriate given the importance of the work, and a sacking offence in my view.

Mmmm but has the actual/real world dataset been `adjusted` to fit/justify the model narrative I.e. you die in hospital of something other than Sars CV2 but as you had a cough on admission you were recorded as asymptomatic CV2.

[Harley]

40 minutes ago, MrXxxx said:

So are you saying that TA as a tool is very much for short/mid term trading whereas FA is for long term investing...or should you be using a combination of both?

I use it everywhere, especially for investing as that's currently my main focus.  TA for trading.  FA for investing with TA to time the purchases.

[Harley]

17 minutes ago, MrXxxx said:

farming

Ah, we haven't discussed the Foot and Mouth experience.  The big picture does not look good.  The key questions for me centre around the clients and other recipients.  Why and how they used the stuff (competence, connections, agendas, etc).

[jamtomorrow]

13 minutes ago, MrXxxx said:

Mmmm but has the actual/real world dataset been `adjusted` to fit/justify the model narrative I.e. you die in hospital of something other than Sars CV2 but as you had a cough on admission you were recorded as asymptomatic CV2.

Those are just raw excess deaths over the period, regardless of cause.

Which of course means that lockdown itself might be a factor/cause of some of those deaths e.g. patients denied access to life-saving diagnostics/treatment, elderly dying of loneliness/isolation.

Strong correlation between excess deaths and lockdown timing would suggest that's a weak effect at best:

Source: https://www.ft.com/content/3c53ab12-d859-4ceb-b262-f6a0221ca129

[The_Doc]

59 minutes ago, Majorpain said:

Completely agree, the media response to the release of the 500,000 figure was off the charts, i don't think any politician would have been brave (or stupid) enough not to have gone into full lockdown.  Any rational analysis about "other deaths" from loss of jobs and lack of healthcare as you highlight was completely swamped by "the virus is going to kill us all" hysteria, he really shouldn't have publicly released something so incendiary without being at least minimally peer reviewed IMO.

Absolutely right. The problem here was that the MSM picked up on the big number (big numbers make big headlines) and ignored the associated probability. The large number of deaths was in the scenario where everyone continued behaving as they had done previously. This was never going to be the case as social distancing had already started even before the government brought in the "2m rule".

I blame the MSM in its interpretation of the Ferguson report.

[sancho panza]

1 hour ago, jamtomorrow said:

That's certainly the received wisdom, but I'm mystified as to why. This is Table 4 from the March 16 paper, final column arguably corresponds most closely to the lockdown we eventually experienced in UK:

Actual excess deaths in UK has been 62,000 to date, which means final 2 columns are an over-estimate, and CI_HQ_SD column is either an under- or over-estimate depending on your choice of R0 and intervention timing - source: https://www.ft.com/content/3c53ab12-d859-4ceb-b262-f6a0221ca129

Here's the paper: https://www.imperial.ac.uk/media/imperial-college/medicine/sph/ide/gida-fellowships/Imperial-College-COVID19-NPI-modelling-16-03-2020.pdf

None of this is to say I'm happy with the software engineering processes and practices followed by the IC modelling team - I'm not, they were atrocious and it's sheer bloody good fortune if the predictions were anywhere near correct. Completely inappropriate given the importance of the work, and a sacking offence in my view.

You need to separate out the deaths that were caused by Covid-not insubstantial I admit- and those deaths caused by the response-also substantial(you'd need a higher level clinical opinion than mine for an assessment of this).That's hard to do because of the way that deaths have been certified over the period(as I understand,but happy to be educated by someone that knows.)

Inherently this discussion is backward looking but the damage the lock down has done more broadly to the health of the nation going forwards,also needs to be considered.

The key figure-ie do nothing- and we get 500,000 deaths was blatantly wrong if we look at Sweden.

Without Sweden,there'd be no natural comparator and many sceptics wouldn't have an argument.

[sancho panza]

53 minutes ago, MrXxxx said:

The problem is either a) he doesn't understand it himself, or b) its badly wrong but he `signed it off`...either is as a result of senior academics farming off their research to postdocs whilst promoting themselves, rather than nurturing them and being more `hands on`....

A big thing in the clinciual circles I swim in(Paramedic- but I think this is across the board) is reflecting after jobs.What went right,what went wrong,what could we have done differently??

A key part of effective reflection is to be open and honest with both yourself and your colleagues,so that all sides can input constructively for the improvement of all.Imperial has been coy about it's code/methodology and the more I'm learning from the opinion of pro's in industry,the worse it's starting to look.You can't have one rule for people like me on the bottom of the heap and then another rule for the chief medical officer/Neil Ferguson.

[sancho panza]

25 minutes ago, The_Doc said:

Absolutely right. The problem here was that the MSM picked up on the big number (big numbers make big headlines) and ignored the associated probability. The large number of deaths was in the scenario where everyone continued behaving as they had done previously. This was never going to be the case as social distancing had already started even before the government brought in the "2m rule".

I blame the MSM in its interpretation of the Ferguson report.

Absolutely.The problem we have is that our political class was cowed by them.I always remember watching TV news and Beth Rigby was hounding poor old Bozza about when he was going to lock down..........unbelieveable to see in the Western world given what lock down was going to entail.

We'll never know who was pushing the agenda-whether it was Ferguson or the govt that was looking for a reason to do what most other govts were doing- but we do knwo that it's had a devastating effect on society more broadly, not jsut in terms of health but individual liberties,kids educations and health.

I hope the political class in London reflect meaningfully on it but expereince teaches us they won't.

[sancho panza]

High Grade Luis

https://twitter.com/crofin67/status/1273901808147214336/photo/1

 

[DurhamBorn]

For anyone holding Telefonica,here are the details on the tax situation of the divis.

Do nothing and take the rights no tax,

Elect to take the cash Spanish Withholding tax of 19%

Sell the rights,no withholding tax,but telephone dealing charges,with Hargreaves these are 1% with a minium of £20/max £40.

I would guess Repsol will be the same.

So option two is a no no of course,so its let the rights become shares,or sell the rights.Repsol do this most years,hopefully Telefonica will as well going forward.

 

What is happening?

Telefonica SA recently announced details of a Stock Dividend to Shareholders as at close of business on 16 June 2020 with a cash option, giving you the right to receive additional Shares or a cash payment. As a holder of Telefonica SA Shares in your Hargreaves Lansdown Account you have been issued with one Right for each Share held at the qualifying time.

What are my choices?

You now have 3 options to consider relating to the Rights and these are explained below in more detail.

Option 1 - Do nothing - DEFAULT. If you do not return an election by the deadline below and take no action you will receive 1 New Telefonica SA Share for every 24 Rights held. The issue of New Ordinary Shares will not be subject to Spanish withholding tax. Any Rights that are not converted into New Shares i.e. not multiples of 24, will lapse and a cash payment will be made in their place. The New Shares are expected to be issued from 17 July 2020.

Option 2 - You can receive a cash payment. If you elect for this option by our noon Tuesday 30 June 2020 deadline you will receive a cash payment of EUR0.193 per Right. The proceeds will be subject to a 19% Spanish withholding tax, i.e. you will receive EUR0.15633 per Right after tax. Cash proceeds are expected from 3 July 2020 and will be converted to Sterling upon receipt subject to the prevailing exchange rate at that time our standard currency conversion fees.

Option 3 - You can sell your Rights. If you elect for this option by our 4:30pm Tuesday 30 June 2020 deadline you will receive the prevailing market price of the Rights at the time of the sale subject to our standard telephone dealing fees. The proceeds of the sale will not be subject to Spanish withholding tax. You can only elect for this option by telephone on 0117 980 9800. N.B. we are unable to accept instructions to sell your Rights that are sent via the Corporate Actions section below or by email and such instructions will not be executed. No guarantee can be given that a market will continue to exist for these Rights until the 4:30pm Tuesday 30 June 2020 deadline. If we are unable to sell your Rights you will receive Shares under the default option (Option 1) instead.

Full details of this event will be sent to Qualifying Hargreaves Lansdown clients by either post or secure message. Please note that any instruction you give must be based on the full details provided in the letter and not on the summary information outlined above.

You currently hold 4,383 shares. 

 

[PaulParanoia]

11 hours ago, The_Doc said:

The code from Ferguson has been independently verified and has produced the same results. See https://zenodo.org/record/3865491

There is also a simple reason as to why running the code with the same input files could produce different results. Monte Carlo simulations are used in epidemiology to create forecasts for the spread of a virus. Each Monte Carlo simulation run will produce one result and will use randomisation. As it uses randomisation, running a second time will produce a different result. 

The whole point of doing this is to obtain a large set of possible outcomes and critically to obtain confidence intervals related to those outcomes. So it it perfectly possible that in the Imperial College report, a large numbr of deaths was one possible outcome. It would have come with a likelihood based on the number of Monte Carlo simulations that produced that figure or greater. This is standard practice.

 

https://lockdownsceptics.org/how-replicable-is-the-imperial-college-model/

The headline is “Codecheck confirms reproducibility of COVID-19 model results”, and the article highlights this quote:

I was able to reproduce the results… from Report 9.

This is an unambiguous statement. However, the press release quotes the report as saying: “Small variations (mostly under 5%) in the numbers were observed between Report 9 and our runs.”

This is an odd definition of “replicate” for the output of a computer program, but it doesn’t really matter because what ICL doesn’t mention is this: the very next sentence of Eglen’s report says:

I observed 3 significant differences:

1. Table A1: R0=2.2, trigger = 3000, PC_CI_HQ_SDOL70, peak beds (in thousands): 40 vs 30, a 25% decrease.
2. Table 5: on trigger = 300, off trigger = 0.75, PC_CI_HQ_SD, total deaths: 39,000 vs 43,000, a 10% increase.
3. Table 5: on trigger = 400, off trigger = 0.75, CI_HQ_SD, total deaths: 100,000 vs 110,000, a 10% increase.

In other words, he wasn’t able to replicate Report 9. There were multiple “significant differences” between what he got and what the British Government based its decisions on.

[PaulParanoia]

Shared some incorrect data so have deleted it.

[jamtomorrow]

1 hour ago, sancho panza said:

The key figure-ie do nothing- and we get 500,000 deaths was blatantly wrong if we look at Sweden.

Without Sweden,there'd be no natural comparator and many sceptics wouldn't have an argument.

I'd want to see data on behavioural change in Sweden (or lack of) before chalking it up as a control experiment. There's good evidence that Government-proclaimed lockdowns elsewhere in Europe were a lagging indicator - a chunk of the population made up their own minds (rightly or wrongly) and changed their behaviour pre-emptively.

Here's Taleb's tweet on the topic this brought to mind: 

 

[jamtomorrow]

8 minutes ago, PaulParanoia said:

That graph was retracted and replaced with this one ...

Got a link? Asking because the date on mine is more recent

[PaulParanoia]

7 minutes ago, jamtomorrow said:

Got a link? Asking because the date on mine is more recent

You are correct, I was confusing this with a previous chart they published which showed total excess deaths along the horizontal axis.  I'll remove the content from my previous post.

[geordie_lurch]

1 hour ago, DurhamBorn said:

You currently hold 4,383 shares. 

Might want to remove that bit @DurhamBorn 

I will just let my meagre amount of shares turn into new ones 

[DoINeedOne]

1 hour ago, DurhamBorn said:

For anyone holding Telefonica,here are the details on the tax situation of the divis.

Do nothing and take the rights no tax,

Elect to take the cash Spanish Withholding tax of 19%

Sell the rights,no withholding tax,but telephone dealing charges,with Hargreaves these are 1% with a minium of £20/max £40.

I would guess Repsol will be the same.

So option two is a no no of course,so its let the rights become shares,or sell the rights.Repsol do this most years,hopefully Telefonica will as well going forward.

 

Just seen the message in my HL account option 1 for me