Jump to content
Strawberry Orange Banana Lime Leaf Slate Sky Blueberry Grape Watermelon Chocolate Marble
Strawberry Orange Banana Lime Leaf Slate Sky Blueberry Grape Watermelon Chocolate Marble

Dr Bob

  • Content Count

  • Joined

  • Last visited

  • Days Won


Dr Bob last won the day on March 30

Dr Bob had the most liked content!

Community Reputation

2259 Excellent

1 Follower

Profile Information

  • Gender

Previous Fields

  • Occupation
    Passed it
  • Boat Name

Recent Profile Visitors

4841 profile views
  1. Ours is due in mid May. We will be renewing. Too good a service to loose.
  2. Just looked at the daily data on worldometers. Spain is now showing the 4th day in a row of reducing NEW cases which is following the trend of the Asian countries and certainly looks like they are over the peak of infections (not deaths yet). Germany is showing a 3 rd day of drop of new infections but it is a huge drop. Deaths are also down but that cant be significant. Hopefully we are there as well....difficult to see today as new cases have gone up a little bit but I guess that could just reflect a higher level of testing today - Gov under pressure to get tests up to 10,000 per day.
  3. I've been trying to teach my 10 year old grandson how to play the guitar.....via video link. Slow progress!
  4. Let's hope they work out whether they are ok!😱
  5. If you are asking the question to be able to compare the death rates in say UK, Spain and Germany then its not worth trying. That is what I was trying to say in an earlier post. Because everyone is collecting data in different ways, it becomes almost impossible to compare daily deaths or total number of deaths. The important information is the rate of change of deaths ie how the number of deaths per day changes .....providing the data collection method is the same. If you decide to start including all deaths rather than just hospital deaths then the graph will jump that day and then a new gradient will be established. Peeps who number crunch the data will know the assumptions used to generate the data in each country and those assumptions are bound to be different. They beat Ferrari to it then?
  6. Just to be clear. In each country there will be a number of data points viz new daily cases infected admitted to hospital admitted to ICU deaths ..and maybe some more I am ignoring. In each country each one of those will be effected by local data collection ie lack of testing in the UK, counting deaths wrong in France, making up the numbers in China and therefore it would be dangerous to use that data in that form. If however you look at the rate of change of the data, then it starts to be come more meaningful and you can see peeps looking at graphs in 'log' form which is the way of looking at rate of change ie dX/dT. You then can look at the rate of change of the rate of change ie d(dX/dT)/dT. This gets rid of differences country to country .....but then you need to start looking at changes in reporting within a country, ie the UK start testing more or France counting deaths outside hospitals. These give big changes on the graphs but then the trend settle down and continue as they are. There are 20 countries reporting data and that is good enough to give a good probable output. Trust me the data is useful and the modellers know what they are doing.
  7. that's why you have to be careful with data. The modellers will not be looking at single data points but trends. They will extrapolate or predict from trends which will be common from country to country. Modelling used to be done by taking single data points and creating relationships via Partial least squares, regression analysis etc but now it is more about comparing 'shapes' ie topology, neural networks or sparse data prediction. Imperial College have a good name in the modelling arena and I have had contact with them over the last 20 years for various projects on modelling. They know what they are talking about.
  8. I think you are wrong. The data that is real is admission figures, ICU admissions and deaths. Numbers of new cases will depend on numbers tested where there are more variables. The models they are using will use the real data and the 'inferred' data. The last 25 years of my career was in the chemicals industry which relies totally on modelling for production. I understand models. I understand data. Admission figures will be useful data but has to be used in context...ie it is all about trends not individual data points....and knowing what affects individual data and the trends.
  9. Or perhaps it could be that that it dies out (ie becomes less active in infecting peeps) after 3 or 4 infection cycles because it is mutating into a less harmfull virus. That would explain the asian countries but will it follow here? Time will tell.
  10. Ah, now the Guardian has a better article on the ICL bod. See here. https://www.theguardian.com/world/2020/mar/30/coronavirus-uk-spread-shows-early-signs-of-slowing-key-adviser Sounds like he is using hospital admissons as a key metric to predict from amongst others. Sounds very promising. Nice to hear some positive messages coming out.
  11. I also can't see how they've avoided reinfection but the case is similar in South Korea and Singapore. All those countries are constraining contact but allowing peeps to go back to work etc. Is it because the majority have had the virus without symptoms? That is the only logical solution I can see but we won't know that till the antibody test comes out. The SK numbers are looking like only 50 new internal cases per day which is 20 times less than at the peak. There is something weird about how these countries infections have peaked, reduced and then not taken off again as restrictions have been lifted.
  12. I've got a glass eye. Does that count?
  13. Oh why is the press being so negative? Headlines today using the words 6 months. Mrs Bob read this (she only ever reads the headlines) and pronounced we are in lockdown till October. Yes some peeps will be reading it as this. By the second paragraph it starts saying the ALL the current restrictions won't be lifted for 3 months which written another way means all the current restrictions will be gone in 3 months. Prolly that means a number off the current restrictions will be lifted sooner. Not at all what the headline may seem to infer. In another article in the Guardian, the head bod at UCL...remember the guy who leads the modelling that the government are using .....says that all the data is pointing at us having reached the peak of infections (@jennifer...you may well be right!) although that article is a load of c**p as well with no joined up thinking or data to help us verify the conclusions. If true it would mean the number of new infections per day dropping which is a week ahead of that seen in China etc. That is huge as China is now relaxing a lot of its lockdown a month after its peak with no reoccurrence of the virus with increased social contact. If mimicked here, that would mean significant relaxation of the lockdown by end April. I am criticising the article here, not the ICL guy .... I think he is a hero and the only one talking sense in all of this.
  14. Are they the dangly bits? It hasn't got any of them!
  • Create New...

Important Information

We have placed cookies on your device to help make this website better. You can adjust your cookie settings, otherwise we'll assume you're okay to continue.