Chit Chat
HAVE WE JUST WATCHED THE BIGGEST POLITICAL DISASTER
|
By:
|
|
By:
stewartsrise
Absolutely bang on. The last election was a scandal. The sensible majority were left with a Hobson's choice. Corbyn and his mad, outdated ideas or BJ and his right wing pals. I have always voted Tory but would have gladly voted Labour last time had they not been led by a man whose ideas were mostly mad and unworkable - and it wasn't a media campaign that made me think that. It was a careful read of his manifesto which was a lunatic document with promises for almost everyone but no ideas how to pay for them. Hence, I voted Tory again but with considerable misgivings. In other words, I voted for the 'least worst'. Politics should NEVER have gotten like this. The general public deserved better than they were offered and I just hope that Kier Starmer identifies how many usual Tory voters like me are so unhappy with BJ and his ultra right wing agenda. If Starmer takes the middle ground our politics so badly need we might have a decent opposition. If not, god only knows where we will end up, unless BJ blows Brexit and then the North will flock back to Labour whatever. The next few years do not look promising for the UK. |
|
By:
|
|
By:
very good point
|
|
By:
The Home Office confirms 17 people who are believed to be from Iraq were brought to Dover this morning after being picked up in the Channel by the Border Force. It brings the number who have attempted the crossing to 321 over the Bank Holiday weekend.
It's incredible, they sailed in a dinghy all the way past Saudi Arabia, UAE, Oman and Yemen, through the Suez Canal, past Egypt, Turkey, Greece, Italy, France, Spain and Portugal. No-one saw them till they were 3 miles off Folkestone. How lucky was that? |
|
By:
the knight
|
|
By:
|
|
By:
|
|
By:
|
|
By:
edy
11 May 20 11:38 Joined: 13 Dec 06 | Topic/replies: 227,046 | Blogger: edy's blog May 11, 2020 -- 11:05AM, InsiderTrader wrote: PorcupineorPineapple11 May 20 10:53Joined: 03 Dec 15| Topic/replies: 9,504 | Blogger: PorcupineorPineapple's blogThe models being used at the moment for making massive decisions would never be used for things like the insurance industry or even by gamblers. They are just not good enough.What's your evidence for this assertion?^Look at Ferguson's code. Not withstanding the assumptions he used in his model are wrong.The random number seeding in the code gave a massive variety in results. Sometimes when run the result is 350k rather than 520k then mess up is something else is going on on the CPU thread. It has so many bugs and errors it simply cannot be trusted. If I run the CovidSim with the same input parameters it will come back with vastly different results each time? ^ Correct. |
|
By:
In short the code is riddled with bugs.
Not seen a single positive thing written about it by anyone in the software/modelling sector. Same extracts from reddit... https://www.reddit.com/r/LockdownSkepticism/comments/gesrvr/code_review_of_fergusons_model/ ..... 'I’ll illustrate with a few bugs. In issue 116 a UK “red team” at Edinburgh University reports that they tried to use a mode that stores data tables in a more efficient format for faster loading, and discovered – to their surprise – that the resulting predictions varied by around 80,000 deaths after 80 days: That mode doesn’t change anything about the world being simulated, so this was obviously a bug. The Imperial team’s response is that it doesn’t matter: they are “aware of some small non-determinisms”, but “this has historically been considered acceptable because of the general stochastic nature of the model”. Note the phrasing here: Imperial know their code has such bugs, but act as if it’s some inherent randomness of the universe, rather than a result of amateur coding. Apparently, in epidemiology, a difference of 80,000 deaths is “a small non-determinism”.' Please join me in prayer that the UK government, having not had the stones to avoid lockdowns in the first place, will at least have the brains to see this as an opportunity to back out of their lose-lose policy using Ferguson as a scapegoat. There isn't going to be a better opportunity than this. Software test engineer here. Wow. That is your basic, bare minimum smoke test. Not even a single unit test in sight. Not all that surprised though, given how often testing takes a back seat in software development. Now I'm not familiar with writing tests in Python, but this seems poorly structured and difficult to read (given the way the source code also looks, makes sense). Could they at least have used a test framework to give some semblance of organization here? I say this without hyperbole: this is without a doubt the worst code I have EVER seen in my ENTIRE life. I wanted to vomit when I saw it. I write programs that are in no way critical to health or infrastructure, and I would be IMMEDIATELY FIRED if I wrote code that was even a fraction as bad as this. I wrote better code in fking college getting my computer science degree. Fire these people. Fire all of them. Software Engineer of ~15 years. If this code came across my desk for review, I wouldn't even bother to try to figure out what it does. I would immediately send it back to the dev and recommend that he be fired. This code is unreadable garbage. But this is supremely bad with the level of detail that they seem to be including with no rational structure. I'd fail a student that produced this. Even the braniacs at Microsoft couldn't sanitise this heap of steaming s. They'd been working on cleaning it up for over a month: The latest commit "Squash history for public release" is telling, no one wanted their name associated with this pile of garbage. If someone on my team submitted this for review, it would be kicked back immediately. No test, crazy variable names (e.g. is P a person or place or ?), huge methods, violates many SOLID principles of OOP, etc. tl;dr; it is junk - Original author of the code review is spot on. In a stochastic simulation, let's say you want to test what changing the proportion of susceptible individuals does. So you change it, and rerun the model, but it gives you a higher peak infection rate than the with a lower number of immune individuals. We know this is wrong, it could be something wrong in the model, or it could be that the particular random seed that was chosen, gave you all the 'worst' random numbers for the 2nd run compared to the first one. You need to isolate the random stochastic noise from the changes to the model. So what you do is you run it with the same pseudo-random seed number, then you know that the randomness is not causing differences between runs and can isolate what individual variables are doing based on that single run. Once you've done this and unit tested all the paramaterisations, you could then run the model with say, a million different random seeds, and take the 95th percentile or the median or whatever you want from those million random runs. That is the basis of all stochastic modelling. What you don't have, is a model that gives a different output with the SAME random seed, because that is just nonsense and indicates a massive bug, not only that, how can you isolate and test individual parameters to perform sensitivity analysis in that environment? You can't. Not seen one positive comment anywhere on the web about the code and model that Boris used as his frame of reference last night. |
|
By:
|
|
By:
Because he read it on the internet
 |
|
By:
Won't he have analysed it himself with his vast knowledge in computer science and interest in the subject?
|
|
By:
I'm not doubting it's bad code btw.
I do think however that IT is concentrating too much on the model over some potential inaccuracies. Other countries have done their own, independent, models. He is also charging too much at Ferguson when it was the Tories that went far and beyond the recommendations of his model. |
|
By:
edy
11 May 20 12:08 Joined: 13 Dec 06 | Topic/replies: 227,054 | Blogger: edy's blog May 11, 2020 -- 11:56AM, InsiderTrader wrote: edy11 May 20 11:38Joined: 13 Dec 06| Topic/replies: 227,046 | Blogger: edy's blog May 11, 2020 -- 11:05AM, InsiderTrader wrote: PorcupineorPineapple11 May 20 10:53Joined: 03 Dec 15| Topic/replies: 9,504 | Blogger: PorcupineorPineapple's blogThe models being used at the moment for making massive decisions would never be used for things like the insurance industry or even by gamblers. They are just not good enough.What's your evidence for this assertion?^Look at Ferguson's code. Not withstanding the assumptions he used in his model are wrong.The random number seeding in the code gave a massive variety in results. Sometimes when run the result is 350k rather than 520k then mess up is something else is going on on the CPU thread. It has so many bugs and errors it simply cannot be trusted.If I run the CovidSim with the same input parameters it will come back with vastly different results each time?^Correct. How do you know? ^ Because the 'model' is based on multiple random number seeds. Anyone found a positive review of his model/code? |
|
By:
edy
11 May 20 12:14 Joined: 13 Dec 06 | Topic/replies: 227,055 | Blogger: edy's blog I'm not doubting it's bad code btw. I do think however that IT is concentrating too much on the model over some potential inaccuracies. Other countries have done their own, independent, models. He is also charging too much at Ferguson when it was the Tories that went far and beyond the recommendations of his model. ^ Any links to other influential models/code? |
|
By:
https://www.rki.de/DE/Content/InfAZ/N/Neuartiges_Coronavirus/Modellierung_Deutschland.pdf?__blob=publicationFile
. |
|
By:
Ferguson's headline figure of 510k is massively misrepresented.
It was an estimate of deaths if nothing was done what so ever. No case isolation, no voluntary home quarantine, no social distancing, no closure of schools and universities. Is that a realistic scenario? UK society would have continued as if nothing had happened, you would read about thousands of new cases, hundreds of deaths. Yet society would continue, stiff upper lip, business as usual. We have evidence that society wouldn't have carried on in this manner, actions were already being taken before the government intervened. Sporting events were being cancelled despite the governments advice that there was no requirement for this to happen. As such, you need to look at the other estimates contained within the report. In scenario 1, an estimate of deaths was provided in a range of 85-98k. This scenario was dependant on the introduction of three key measures i.e. case isolation, voluntary home quarantine, and social distancing In scenario 2, an estimate of deaths was provided in a range of 12-53k. This scenario was dependant on the introduction of three key measures i.e. closure of schools and universities, case isolation, and social distancing In scenario 3, an estimate of deaths was provided in a range of 9-39k. This scenario was dependant on the introduction of three key measures i.e. closure of schools and universities, case isolation, voluntary home quarantine, and social distancing Viewing the report through this lens provides a significantly more balanced approach than you arrive at if you consider the 510k figure as being the inevitable result once the virus had arrived in the UK. Now, does this mean that I am a raging fan of the report produced by Ferguson and his team? No, it does not. It does, however, demonstrate that I am prepared to go behind the headlines and spend some time to actually read the report that created the headline. |
|
By:
a very quick look from anyone at the study they keep talking about would've revealed that the authors themselves described the worst case number as highly unrealistic.
|
|
By:
Regarding the Imperial College study
|
|
By:
Suddenly, the estimates produced don't appear to be so out of whack after all, do they.
|
|
By:
I pointed it out in another thread the other day, but for some weird reason it was ignored and some people continued to pretend that the Imperial College predicted it was likely that 500k would've died without a lockdown.
|
|
By:
I wonder why that might do that
 |
|
By:
surely IT hasn't been lying again
|
|
By:
Just blindly believed what his masters told him imo. No bad faith behind it.
|
|
By:
Without him being willing to do his own research or consider the points of others though.
|
|
By:
 |
|
By:
you expected that from hard leftists on this forum but a proud traditional conservative, you expect better
|
|
By:
I posted last night ,those who didnt want a lockdown use the stats achieved through the lockdown as their arghument for not having it,not the stats we may have got without it.
to quote the well trodden Brexit phrase having their cake and eating it,  |
|
By:
|
|
By:
|
|
By:
aaronh
11 May 20 12:32 Joined: 18 Sep 09 | Topic/replies: 272,474 | Blogger: aaronh's blog surely IT hasn't been lying again ^ Where? |
|
By:
Angoose
11 May 20 12:23 Joined: 18 Jul 02 | Topic/replies: 16,275 | Blogger: Angoose's blog Ferguson's headline figure of 510k is massively misrepresented. It was an estimate of deaths if nothing was done what so ever. No case isolation, no voluntary home quarantine, no social distancing, no closure of schools and universities. Is that a realistic scenario? UK society would have continued as if nothing had happened, you would read about thousands of new cases, hundreds of deaths. Yet society would continue, stiff upper lip, business as usual. We have evidence that society wouldn't have carried on in this manner, actions were already being taken before the government intervened. Sporting events were being cancelled despite the governments advice that there was no requirement for this to happen. As such, you need to look at the other estimates contained within the report. In scenario 1, an estimate of deaths was provided in a range of 85-98k. This scenario was dependant on the introduction of three key measures i.e. case isolation, voluntary home quarantine, and social distancing In scenario 2, an estimate of deaths was provided in a range of 12-53k. This scenario was dependant on the introduction of three key measures i.e. closure of schools and universities, case isolation, and social distancing In scenario 3, an estimate of deaths was provided in a range of 9-39k. This scenario was dependant on the introduction of three key measures i.e. closure of schools and universities, case isolation, voluntary home quarantine, and social distancing Viewing the report through this lens provides a significantly more balanced approach than you arrive at if you consider the 510k figure as being the inevitable result once the virus had arrived in the UK. Now, does this mean that I am a raging fan of the report produced by Ferguson and his team? No, it does not. It does, however, demonstrate that I am prepared to go behind the headlines and spend some time to actually read the report that created the headline. ^ What do you dislike about the Ferguson report? |
|
By:
(I don't think Angoose is the real author of that piece of text)
|
|
By:
edolf and aaron giving it the high and mighty with a guy IT who's trying to have a genuine discussion as evidenced by his posts
Yet this is the sort of thing YOU TWO post edy • May 9, 2020 9:26 PM BST Funny you would say that. I did dream of your soft, but masculine, hands rubbing oil, lavender scented, all over my naked body before you gently went on to spoil me further by massaging my penis with your very skilled tongue and mouth. aaronh • May 9, 2020 9:38 PM BST LoveLoveLoveLove Are you two paid to post on here? Your reactions lack judgement and common sense it's as if you're being paid to silence opposing views whilst embarrassing yourselves as bait. |
|
By:
Denzil, you're the one who has just spent his time being a raging reactionary and posting conspiracy nonsense since you came back
|
|
By:
Denzil - is there something you're trying to get off your chest. All friends here mate. No judgments.
|
|
By:
Denzil posting an unrelated hissy fit in a thread in his homophobic obsession towards an anonymous heterosexual male on an internet forum.
Imagine my shock! |
|
By:
wonder if Denzil has those 2 posts copy pasted in a document or just keeps going back to that thread to find it
|