I commented last night to one of your posts about computer vs humans that I have just started James Case's book, Competition: The Birth of a New Science, and that, in that, he addresses why machines are poorer at multi-player games, like Go, than two-player, zero-sum games.
I have not got fully into his argumentation yet, but you seemed to be implying that computers do well at Go. Is his data out of date (the book is copyright 2007 so was probably written in 2006), or is he still right that the very best human players of Go still beat machines?
Earlier this year, a computer (AlphaGo) beat the one of the best ranked humans Lee Sodol 4 v 1 after five matches.
Labour going to score 33% at the next general election? Looks like the next polling disaster is already in the making...
33% might just be accurate in the hypothetical event of an election tomorrow with no warning.
In the real world there will be six weeks of Conservative billboards with Corbyn/McDonnell quotes about the IRA/Hamas/printing money, perhaps plus a concerted UKIP campaign in the North.
Opinium had Labour on 32% just over a week ago.
On the whole, Governments tend to lose ground in the formal election campaign period.
I commented last night to one of your posts about computer vs humans that I have just started James Case's book, Competition: The Birth of a New Science, and that, in that, he addresses why machines are poorer at multi-player games, like Go, than two-player, zero-sum games.
I have not got fully into his argumentation yet, but you seemed to be implying that computers do well at Go. Is his data out of date (the book is copyright 2007 so was probably written in 2006), or is he still right that the very best human players of Go still beat machines?
AlphaGo is now beating the leading human players, as of this year.
I would pitch the LibDems lower than that at 9/10% simply because they now have so much competition as the NOTA option.Many who in the past would have supported them on that basis will now opt for the Greens or UKIP.
I think you are running about two years behind the times, Mr Justin.
+1
Why be Green when you can love Corbyn, and why be UKIP when it's 'job done' and both Banks and Farage are about to take their energies and £ elsewhere?
F**k, how is that even possible - is the control fully manual with no safety backups?
From the discussion a few days ago, it appears the speed limit elsewhere is higher, and only the max speed is capped. The tram doesn't "know" it's on a 12 MPH section.
Yes, from a quick skim of the report the speed limit dropped from 80kph to 20kph for the turn, the tram barely slowed and hit the turn at 70kph. Tram had a 'black box' recording driver inputs, but as you say it doesn't appear there's any mechanism to automatically slow the tram as it flies past a speed limit sign. Looks like the driver was either asleep or distracted unfortunately.
There are very few automated systems because trams ('light rail') are treated under different regulations from both heavy rail and busses, AIUI.
From what I've read the trams have a deadmans' handle-type thing, but there is very little signalling. This is because trams traditionally run along roads. The problem appears to be where the tram rums along an old railway line, and therefore can get up to highish speeds, before transitioning to on-street running.
In this, it's a bit like the Spanish high-speed rail crash of a few years ago, which occurred on a transition between high- and low-speed sections.
As an aside, the Cambridge (mis)Guided Bus has suffered a number of accidents over the last couple of years, including busses ending up on their sides. The investigations have appeared short, and the drivers blamed and sacked perfunctorily. This is, to my mind, a really bad thing, as causal factors won't be discovered.
Does anyone agree with me, that the media in this country has never been so inept, so dire and justso embarrassingly awful.
I see Sophy Ridge (SKY) is being given a new Sunday Politics Show. WHY? Do we really need another one, with the same people just dashing from studio to studio?
I've stopped watching it. I was very unimpressed over post-Brexit - and Trump has just confirmed it. The DT has just tweeted an article about a 'source' that 'allegedly part of the transition team' rubbishing Trump.
Seriously? This isn't news - it's gossip of the lowest kind. Fake news is too kind.
Labour going to score 33% at the next general election? Looks like the next polling disaster is already in the making...
33% might just be accurate in the hypothetical event of an election tomorrow with no warning.
In the real world there will be six weeks of Conservative billboards with Corbyn/McDonnell quotes about the IRA/Hamas/printing money, perhaps plus a concerted UKIP campaign in the North.
Opinium had Labour on 32% just over a week ago.
On the whole, Governments tend to lose ground in the formal election campaign period.
We are not in 'on the whole' times here.
You need to be careful with dismissing 'on the whole' arguments completely though. POTUS fundamentally pointed toward a GOP victory off of historic precedents.
Corbyn's Labour definitely looks to be in a piss weak position to me though.
Surely the response is likely to be dependent on the cause! Many would fight to defend their country from invasion who would have no wish to participate in British aggression in Iraq.
The worrying thing about the data is the top two. Totally incompatible with each other and both having nuclear weapons.
Surely it mainly illustrates that what we would call the 'middle class' (and upper) in each of these countries will generally say 'no'? Look at it in that light, and the data seems almost entirely explained. Factor in residual anti-militarism in the WWII aggressors and you are there.
I would offer an alternative explanation. Pakistan and India are in direct conflict at the moment, albeit through the proxy of Kashmir, but people are being killed daily in this direct conflict.
Turkey has recently had a coup and has a war on its doorstep.
China and Russia run militaristic governments and are both in very expansionist mindsets (South China sea for China, Middle East and Eastern Europe for Russia).
The West is at peace and has no real immediate threat of getting sucked into wars on their own patch (terrorism excluded)
My thesis is that it is the proximity of the threat that yields these numbers. If war were to erupt in France with a danger of spilling into the UK, you'd see our numbers go up into the 80s and 90s very, very quickly.
Do you know why there were so many fatalities? It looked from the images like it fell over and slid to a halt - more of a cuts and bruises sort of thing, with a few broken bones for the unlucky.
...We only have a Parliament because we can't have referenda on every matter. It's a practicability issue....
No it isn't. "The People" generally want their cake and eat it too. Lower taxes and better public services. In the single market but no free movement of people. That is the kind of childish behaviour that can be exploited by demagogues backed by foreign owned media.
We need a representative democracy with dedicated MPs to make the difficult trade-offs that "The people" are incapable of. If you don't like the way they do it, you can vote them out next time. That is what our democracy is about.
Germany has forbidden the use of referenda for a reason.
I commented last night to one of your posts about computer vs humans that I have just started James Case's book, Competition: The Birth of a New Science, and that, in that, he addresses why machines are poorer at multi-player games, like Go, than two-player, zero-sum games.
I have not got fully into his argumentation yet, but you seemed to be implying that computers do well at Go. Is his data out of date (the book is copyright 2007 so was probably written in 2006), or is he still right that the very best human players of Go still beat machines?
AlphaGo is now beating the leading human players, as of this year.
Thanks. Is that for the normal version of Go, or a competition version of the rules that limits multi-player aspects and non-zero sum outcomes?
Mr. Barnesian, if Germans had had a vote, perhaps they wouldn't've allowed in over a million totally unvetted foreigners (which had the charming side-effect of pissing off most of their neighbours).
I commented last night to one of your posts about computer vs humans that I have just started James Case's book, Competition: The Birth of a New Science, and that, in that, he addresses why machines are poorer at multi-player games, like Go, than two-player, zero-sum games.
I have not got fully into his argumentation yet, but you seemed to be implying that computers do well at Go. Is his data out of date (the book is copyright 2007 so was probably written in 2006), or is he still right that the very best human players of Go still beat machines?
The most recent game series was won by a computer. However that does not mean that your desktop computer has the capacity to run the program to beat a human while complying with the time rules. (There is no time limit IIRC - when you run out of normal time you are given a short period (1 minute) to make each move - Byo Yomi).
There have been accusations in high-level chess games of people going to the toilet to do a quick analysis of their position on a computer.
Go, like chess, is a game where the status is fully analyseable and I assume that tree pruning is still used for simple life-and-death situations of the armies. However there are so many more combinations available that depth-first analysis is pretty hopeless. I would assume that the latest AI machines are able to analyse the whole-board position (influence, territory) as well as target individual areas.
Programmes playing bridge or other card games have incomplete information and they tend to run multiple simulations consistent with the information that they do have and work out the best chance of optimising the result. (Humans cheat in Bridge by exchanging information illegally between them.)
There are very few automated systems because trams ('light rail') are treated under different regulations from both heavy rail and busses, AIUI.
From what I've read the trams have a deadmans' handle-type thing, but there is very little signalling. This is because trams traditionally run along roads. The problem appears to be where the tram rums along an old railway line, and therefore can get up to highish speeds, before transitioning to on-street running.
In this, it's a bit like the Spanish high-speed rail crash of a few years ago, which occurred on a transition between high- and low-speed sections.
As an aside, the Cambridge (mis)Guided Bus has suffered a number of accidents over the last couple of years, including busses ending up on their sides. The investigations have appeared short, and the drivers blamed and sacked perfunctorily. This is, to my mind, a really bad thing, as causal factors won't be discovered.
At least RAIB's involved in Croydon.
The accident was on the corner where the old rail line turns to meet the on-street line, where a significant change of speed is required.
If the RAIB are anything like the AAIB, they'll issue a comprehensive final report and not stop short of understanding all the factors behind the accident. They will make such recommendations as to prevent a single error from causing another similar accident. This will undoubtedly include a means of alerting the driver to an overspeed, and probably recommend research on automated brakeing systems.
Do you know why there were so many fatalities? It looked from the images like it fell over and slid to a halt - more of a cuts and bruises sort of thing, with a few broken bones for the unlucky.
Note: heavy speculation follows:
According to the RAIB document Mr Pit posted, the tram slid on its side for 25 metres. Tram sides are mostly glass, and the ground surface uneven, with tracks, ballast shoulders and other items. The glass shatters, and people get trapped between ground and sliding tram.
I commented last night to one of your posts about computer vs humans that I have just started James Case's book, Competition: The Birth of a New Science, and that, in that, he addresses why machines are poorer at multi-player games, like Go, than two-player, zero-sum games.
I have not got fully into his argumentation yet, but you seemed to be implying that computers do well at Go. Is his data out of date (the book is copyright 2007 so was probably written in 2006), or is he still right that the very best human players of Go still beat machines?
Go is a two player game, it is "too complex" to 'solve'
Computers were terrible at it previously.
Deep Mind thrashed Lee Sedol, the world champ at it - Now. It doesn't 'solve' the game, but uses evolving algorithms as it plays itself.
Surely the response is likely to be dependent on the cause! Many would fight to defend their country from invasion who would have no wish to participate in British aggression in Iraq.
The worrying thing about the data is the top two. Totally incompatible with each other and both having nuclear weapons.
Surely it mainly illustrates that what we would call the 'middle class' (and upper) in each of these countries will generally say 'no'? Look at it in that light, and the data seems almost entirely explained. Factor in residual anti-militarism in the WWII aggressors and you are there.
I would offer an alternative explanation. Pakistan and India are in direct conflict at the moment, albeit through the proxy of Kashmir, but people are being killed daily in this direct conflict.
Turkey has recently had a coup and has a war on its doorstep.
China and Russia run militaristic governments and are both in very expansionist mindsets (South China sea for China, Middle East and Eastern Europe for Russia).
The West is at peace and has no real immediate threat of getting sucked into wars on their own patch (terrorism excluded)
My thesis is that it is the proximity of the threat that yields these numbers. If war were to erupt in France with a danger of spilling into the UK, you'd see our numbers go up into the 80s and 90s very, very quickly.
I'm sure you're right, and anyway it's easy to read far too much into this sort of thing, as Hitler reportedly did after the King and Country Debate.
I would pitch the LibDems lower than that at 9/10% simply because they now have so much competition as the NOTA option.Many who in the past would have supported them on that basis will now opt for the Greens or UKIP.
I think you are running about two years behind the times, Mr Justin.
+1
Why be Green when you can love Corbyn, and why be UKIP when it's 'job done' and both Banks and Farage are about to take their energies and £ elsewhere?
The point is that in the past many people voted LibDem - or Liberal - despite not supporting their policy agendas at all. Many were anti - EU and quite a few had racist views more likely to be associated with the BNP.They voted LibDem simply out of a wish to reject both Tories and Labour.Today they continue to be beneficiaries of such votes - but have much more competition.
I commented last night to one of your posts about computer vs humans that I have just started James Case's book, Competition: The Birth of a New Science, and that, in that, he addresses why machines are poorer at multi-player games, like Go, than two-player, zero-sum games.
I have not got fully into his argumentation yet, but you seemed to be implying that computers do well at Go. Is his data out of date (the book is copyright 2007 so was probably written in 2006), or is he still right that the very best human players of Go still beat machines?
AlphaGo is now beating the leading human players, as of this year.
Thanks. Is that for the normal version of Go, or a competition version of the rules that limits multi-player aspects and non-zero sum outcomes?
Go is a two-player game. Each side places black/white pieces to conquer the board. Rules are straightforward, but it is considered more complex than chess to master.
I commented last night to one of your posts about computer vs humans that I have just started James Case's book, Competition: The Birth of a New Science, and that, in that, he addresses why machines are poorer at multi-player games, like Go, than two-player, zero-sum games.
I have not got fully into his argumentation yet, but you seemed to be implying that computers do well at Go. Is his data out of date (the book is copyright 2007 so was probably written in 2006), or is he still right that the very best human players of Go still beat machines?
The most recent game series was won by a computer. However that does not mean that your desktop computer has the capacity to run the program to beat a human while complying with the time rules. (There is no time limit IIRC - when you run out of normal time you are given a short period (1 minute) to make each move - Byo Yomi).
There have been accusations in high-level chess games of people going to the toilet to do a quick analysis of their position on a computer.
Go, like chess, is a game where the status is fully analyseable and I assume that tree pruning is still used for simple life-and-death situations of the armies. However there are so many more combinations available that depth-first analysis is pretty hopeless. I would assume that the latest AI machines are able to analyse the whole-board position (influence, territory) as well as target individual areas.
Programmes playing bridge or other card games have incomplete information and they tend to run multiple simulations consistent with the information that they do have and work out the best chance of optimising the result. (Humans cheat in Bridge by exchanging information illegally between them.)
No human cheats at bridge! People who cheat at bridge are inhuman.
It still strikes me as odd that an accident where essentially a vehicle tilts through 90 degrees should kill seven people. Thee must have been a terrible amount of internal destruction to have inflicted that. I'd be wanting to look at the construction of the whole fleet.
"But if I have taught you anything this year it is that mass delusions – and small group delusions – are totally common and expected in life. On the other hand, the rise of the world’s most powerful super-hypnotist – who can’t win a jury trial – is unlikely..."
I commented last night to one of your posts about computer vs humans that I have just started James Case's book, Competition: The Birth of a New Science, and that, in that, he addresses why machines are poorer at multi-player games, like Go, than two-player, zero-sum games.
I have not got fully into his argumentation yet, but you seemed to be implying that computers do well at Go. Is his data out of date (the book is copyright 2007 so was probably written in 2006), or is he still right that the very best human players of Go still beat machines?
Go is a two player game, it is "too complex" to 'solve'
Computers were terrible at it previously.
Deep Mind thrashed Lee Sedol, the world champ at it - Now. It doesn't 'solve' the game, but uses evolving algorithms as it plays itself.
Do you know why there were so many fatalities? It looked from the images like it fell over and slid to a halt - more of a cuts and bruises sort of thing, with a few broken bones for the unlucky.
Note: heavy speculation follows:
According to the RAIB document Mr Pit posted, the tram slid on its side for 25 metres. Tram sides are mostly glass, and the ground surface uneven, with tracks, ballast shoulders and other items. The glass shatters, and people get trapped between ground and sliding tram.
I would pitch the LibDems lower than that at 9/10% simply because they now have so much competition as the NOTA option.Many who in the past would have supported them on that basis will now opt for the Greens or UKIP.
I think you are running about two years behind the times, Mr Justin.
+1
Why be Green when you can love Corbyn, and why be UKIP when it's 'job done' and both Banks and Farage are about to take their energies and £ elsewhere?
The point is that in the past many people voted LibDem - or Liberal - despite not supporting their policy agendas at all. Many were anti - EU and quite a few had racist views more likely to be associated with the BNP.They voted LibDem simply out of a wish to reject both Tories and Labour.Today they continue to be beneficiaries of such votes - but have much more competition.
I think you will find, however, that all of the parties draw significant support from people who don't support (or are even aware of) their policy agenda. Not least because it is a well established fact that as many people vote 'against' what they don't want as are attracted positively to the party of their choice. There was a recent example in a high profile contest, let me see if I can find a link....
Do you know why there were so many fatalities? It looked from the images like it fell over and slid to a halt - more of a cuts and bruises sort of thing, with a few broken bones for the unlucky.
The interim report didn't make mention of injuries, although the final report will. There were reportedly around 60 passengers on board, knowing if that number would mean lots were standing would be a good starting point. Unfortunately the RAIB believe the CCTV on board to have been unserviceable at the time of the accident. The tram slid for approx 25 metres, during which time it decelerated from 70kph (20m/s) to a standstill - assuming a linear deceleration would have been a 1.2s stop, at just under 2g. That's not a massive g-force, so there must have been serious damage to the vehicle itself. Sad.
AFAIK AlphaGo uses a huge neural network (i.e. pattern recognition) and built up a huge knowledge by playing millions of games against itself, as well as against humans.
There are very few automated systems because trams ('light rail') are treated under different regulations from both heavy rail and busses, AIUI.
From what I've read the trams have a deadmans' handle-type thing, but there is very little signalling. This is because trams traditionally run along roads. The problem appears to be where the tram rums along an old railway line, and therefore can get up to highish speeds, before transitioning to on-street running.
In this, it's a bit like the Spanish high-speed rail crash of a few years ago, which occurred on a transition between high- and low-speed sections.
As an aside, the Cambridge (mis)Guided Bus has suffered a number of accidents over the last couple of years, including busses ending up on their sides. The investigations have appeared short, and the drivers blamed and sacked perfunctorily. This is, to my mind, a really bad thing, as causal factors won't be discovered.
At least RAIB's involved in Croydon.
The accident was on the corner where the old rail line turns to meet the on-street line, where a significant change of speed is required.
If the RAIB are anything like the AAIB, they'll issue a comprehensive final report and not stop short of understanding all the factors behind the accident. They will make such recommendations as to prevent a single error from causing another similar accident. This will undoubtedly include a means of alerting the driver to an overspeed, and probably recommend research on automated brakeing systems.
RAIB are excellent at finding and listing causal factors. If anything, I think their reports are better (and more accessible) than the AAIB.
From what I've read, there are problems with automated systems in trams, not helped by the plethora of different tram systems out there. I'm far from being a tramways expert though.
Do you know why there were so many fatalities? It looked from the images like it fell over and slid to a halt - more of a cuts and bruises sort of thing, with a few broken bones for the unlucky.
The interim report didn't make mention of injuries, although the final report will. There were reportedly around 60 passengers on board, knowing if that number would mean lots were standing would be a good starting point. Unfortunately the RAIB believe the CCTV on board to have been unserviceable at the time of the accident. The tram slid for approx 25 metres, during which time it decelerated from 70kph to a standstill - that's quite a g-force, assuming a linear deceleration would have been 1.2s stop, at just under 2g.
2g is nothing. You get more than that on a roller-coaster*. I think Josias' suggestion of a "cheese-grater" effect is more convincing.
Do you know why there were so many fatalities? It looked from the images like it fell over and slid to a halt - more of a cuts and bruises sort of thing, with a few broken bones for the unlucky.
The interim report didn't make mention of injuries, although the final report will. There were reportedly around 60 passengers on board, knowing if that number would mean lots were standing would be a good starting point. Unfortunately the RAIB believe the CCTV on board to have been unserviceable at the time of the accident. The tram slid for approx 25 metres, during which time it decelerated from 70kph (20m/s) to a standstill - assuming a linear deceleration would have been a 1.2s stop, at just under 2g. That's not a massive g-force, so there must have been serious damage to the vehicle itself. Sad.
I think I read somewhere that the windows against the ground smashed and some people ended up therefore on moving tarmac and getting dragged under the sliding body of the tram.
I commented last night to one of your posts about computer vs humans that I have just started James Case's book, Competition: The Birth of a New Science, and that, in that, he addresses why machines are poorer at multi-player games, like Go, than two-player, zero-sum games.
I have not got fully into his argumentation yet, but you seemed to be implying that computers do well at Go. Is his data out of date (the book is copyright 2007 so was probably written in 2006), or is he still right that the very best human players of Go still beat machines?
The most recent game series was won by a computer. However that does not mean that your desktop computer has the capacity to run the program to beat a human while complying with the time rules. (There is no time limit IIRC - when you run out of normal time you are given a short period (1 minute) to make each move - Byo Yomi).
There have been accusations in high-level chess games of people going to the toilet to do a quick analysis of their position on a computer.
Go, like chess, is a game where the status is fully analyseable and I assume that tree pruning is still used for simple life-and-death situations of the armies. However there are so many more combinations available that depth-first analysis is pretty hopeless. I would assume that the latest AI machines are able to analyse the whole-board position (influence, territory) as well as target individual areas.
Programmes playing bridge or other card games have incomplete information and they tend to run multiple simulations consistent with the information that they do have and work out the best chance of optimising the result. (Humans cheat in Bridge by exchanging information illegally between them.)
No human cheats at bridge! People who cheat at bridge are inhuman.
Just looking through some of the data, if Trump is going to end all amnesties and turn back illegal immigration I think that puts Nevada into play for the GOP and takes Texas and Georgia off the endangered list for them. For the former even after all of those reports of record early votes, record Hispanic turnout Trump closed the margin for the GOP. It also puts Colorado into play IMO.
Also, Trump has pulled ahead of Romney in terms of the raw vote. This idea that he pulled in fewer voters than Romney should be consigned to the dustbin because it isn't true. If anything Trump should beat Romney quite handily even with Johnson eating into Trump's base in safe GOP states.
I would pitch the LibDems lower than that at 9/10% simply because they now have so much competition as the NOTA option.Many who in the past would have supported them on that basis will now opt for the Greens or UKIP.
I think you are running about two years behind the times, Mr Justin.
+1
Why be Green when you can love Corbyn, and why be UKIP when it's 'job done' and both Banks and Farage are about to take their energies and £ elsewhere?
The point is that in the past many people voted LibDem - or Liberal - despite not supporting their policy agendas at all. Many were anti - EU and quite a few had racist views more likely to be associated with the BNP.They voted LibDem simply out of a wish to reject both Tories and Labour.Today they continue to be beneficiaries of such votes - but have much more competition.
I think you will find, however, that all of the parties draw significant support from people who don't support (or are even aware of) their policy agenda. Not least because it is a well established fact that as many people vote 'against' what they don't want as are attracted positively to the party of their choice. There was a recent example in a high profile contest, let me see if I can find a link....
That is true but there is evidence that a lot of LibDem support over the years has been 'easy come easy go!'. Little underlying core support.
Do you know why there were so many fatalities? It looked from the images like it fell over and slid to a halt - more of a cuts and bruises sort of thing, with a few broken bones for the unlucky.
The interim report didn't make mention of injuries, although the final report will. There were reportedly around 60 passengers on board, knowing if that number would mean lots were standing would be a good starting point. Unfortunately the RAIB believe the CCTV on board to have been unserviceable at the time of the accident. The tram slid for approx 25 metres, during which time it decelerated from 70kph (20m/s) to a standstill - assuming a linear deceleration would have been a 1.2s stop, at just under 2g. That's not a massive g-force, so there must have been serious damage to the vehicle itself. Sad.
I think I read somewhere that the windows against the ground smashed and some people ended up therefore on moving tarmac and getting dragged under the sliding body of the tram.
That's my impression. See below. Except it's not tarmac. The tram might well have slid along or over rails, meaning there would have been gaps, sometimes momentarily.
That means that Trump's margin there has now dropped below 0.25%. Would that not trigger an automatic recount there?
Michigan: A mandatory recount if 2,000 votes or less separate the top two candidates. Candidates can petition for a recount due to fraud or a counting error.
There are very few automated systems because trams ('light rail') are treated under different regulations from both heavy rail and busses, AIUI.
From what I've read the trams have a deadmans' handle-type thing, but there is very little signalling. This is because trams traditionally run along roads. The problem appears to be where the tram rums along an old railway line, and therefore can get up to highish speeds, before transitioning to on-street running.
In this, it's a bit like the Spanish high-speed rail crash of a few years ago, which occurred on a transition between high- and low-speed sections.
As an aside, the Cambridge (mis)Guided Bus has suffered a number of accidents over the last couple of years, including busses ending up on their sides. The investigations have appeared short, and the drivers blamed and sacked perfunctorily. This is, to my mind, a really bad thing, as causal factors won't be discovered.
At least RAIB's involved in Croydon.
The accident was on the corner where the old rail line turns to meet the on-street line, where a significant change of speed is required.
If the RAIB are anything like the AAIB, they'll issue a comprehensive final report and not stop short of understanding all the factors behind the accident. They will make such recommendations as to prevent a single error from causing another similar accident. This will undoubtedly include a means of alerting the driver to an overspeed, and probably recommend research on automated brakeing systems.
RAIB are excellent at finding and listing causal factors. If anything, I think their reports are better (and more accessible) than the AAIB.
From what I've read, there are problems with automated systems in trams, not helped by the plethora of different tram systems out there. I'm far from being a tramways expert though.
Me neither, but I'm familiar with the work of the AAIB - they are the best in the world in their field.
As a starting point, a tram-activated or GPS-based siren sounding before the corner would make sure a sleepy or distracted driver hit the brake on time. More signage before the corner would also help.
Do you know why there were so many fatalities? It looked from the images like it fell over and slid to a halt - more of a cuts and bruises sort of thing, with a few broken bones for the unlucky.
The interim report didn't make mention of injuries, although the final report will. There were reportedly around 60 passengers on board, knowing if that number would mean lots were standing would be a good starting point. Unfortunately the RAIB believe the CCTV on board to have been unserviceable at the time of the accident. The tram slid for approx 25 metres, during which time it decelerated from 70kph to a standstill - that's quite a g-force, assuming a linear deceleration would have been 1.2s stop, at just under 2g.
2g is nothing. You get more than that on a roller-coaster*. I think Josias' suggestion of a "cheese-grater" effect is more convincing.
Why are the media making such a fuss about the teenager who decided to leave home for a few days? After all, Tony Blair apparently slept rough in Leicester Square when he was 16 and no-one thought that was especially important.
Do you know why there were so many fatalities? It looked from the images like it fell over and slid to a halt - more of a cuts and bruises sort of thing, with a few broken bones for the unlucky.
The interim report didn't make mention of injuries, although the final report will. There were reportedly around 60 passengers on board, knowing if that number would mean lots were standing would be a good starting point. Unfortunately the RAIB believe the CCTV on board to have been unserviceable at the time of the accident. The tram slid for approx 25 metres, during which time it decelerated from 70kph to a standstill - that's quite a g-force, assuming a linear deceleration would have been 1.2s stop, at just under 2g.
2g is nothing. You get more than that on a roller-coaster*. I think Josias' suggestion of a "cheese-grater" effect is more convincing.
I commented last night to one of your posts about computer vs humans that I have just started James Case's book, Competition: The Birth of a New Science, and that, in that, he addresses why machines are poorer at multi-player games, like Go, than two-player, zero-sum games.
I have not got fully into his argumentation yet, but you seemed to be implying that computers do well at Go. Is his data out of date (the book is copyright 2007 so was probably written in 2006), or is he still right that the very best human players of Go still beat machines?
I commented last night to one of your posts about computer vs humans that I have just started James Case's book, Competition: The Birth of a New Science, and that, in that, he addresses why machines are poorer at multi-player games, like Go, than two-player, zero-sum games.
I have not got fully into his argumentation yet, but you seemed to be implying that computers do well at Go. Is his data out of date (the book is copyright 2007 so was probably written in 2006), or is he still right that the very best human players of Go still beat machines?
Go is a two player game, it is "too complex" to 'solve'
Computers were terrible at it previously.
Deep Mind thrashed Lee Sedol, the world champ at it - Now. It doesn't 'solve' the game, but uses evolving algorithms as it plays itself.
Thought it was just up your street, though you're probably all to aware of it.
Thanks for the link. I had missed that particular article, but know the subject matter well. I wrote the Nature review of the authoritative book on this, Leitenberg and Zilinskas "The Soviet Biological Weapons Program" and I am good friends of long standing of the husband and wife team who did the initial Western epidemiological investigation - Jeanne Guillemin and Matt Meselson:
I think the article you cite somewhat overstates the achievements in stating that "However, the final product was infectious bacteria that could evade immune responses as well as multiple drugs." I think that was the aim, but getting all of the desired features into the bug at the same time was problematic, and preventing the bug from evolving out the new features was even more so. The article does go onto hint at this, but my review of Ray and Milton's book gives you more detail about the problems
I commented last night to one of your posts about computer vs humans that I have just started James Case's book, Competition: The Birth of a New Science, and that, in that, he addresses why machines are poorer at multi-player games, like Go, than two-player, zero-sum games.
I have not got fully into his argumentation yet, but you seemed to be implying that computers do well at Go. Is his data out of date (the book is copyright 2007 so was probably written in 2006), or is he still right that the very best human players of Go still beat machines?
The most recent game series was won by a computer. However that does not mean that your desktop computer has the capacity to run the program to beat a human while complying with the time rules. (There is no time limit IIRC - when you run out of normal time you are given a short period (1 minute) to make each move - Byo Yomi).
There have been accusations in high-level chess games of people going to the toilet to do a quick analysis of their position on a computer.
Go, like chess, is a game where the status is fully analyseable and I assume that tree pruning is still used for simple life-and-death situations of the armies. However there are so many more combinations available that depth-first analysis is pretty hopeless. I would assume that the latest AI machines are able to analyse the whole-board position (influence, territory) as well as target individual areas.
Programmes playing bridge or other card games have incomplete information and they tend to run multiple simulations consistent with the information that they do have and work out the best chance of optimising the result. (Humans cheat in Bridge by exchanging information illegally between them.)
No human cheats at bridge! People who cheat at bridge are inhuman.
Didn't we used to say the same about cricket?
I was simply aiming for the true scotsman... of course people cheat.
I avoid cheating by repeatedly misremembering which convention is which and how to bid them...
The margin in Michigan is the equivalent to a candidate in a typical UK constituency having a lead of just 120 votes.Such a margin would almost certainly result in a full - rather than a bundle - recount.
Me neither, but I'm familiar with the work of the AAIB - they are the best in the world in their field.
As a starting point, a tram-activated or GPS-based siren sounding before the corner would make sure a sleepy or distracted driver hit the brake on time. More signage before the corner would also help.
More signage: definitely. As for this crash, we cannot assume it was the driver's fault. There might have been other factors causing, or contributing, to it.
Most railways prefer to use trackside balises to communicate with the train. Basically: if you pass this point and you're going above x MPH, apply full emergency brake.
Also AIUI, these trams are fitted with electrmagnetic eddy current brakes that are *very* effective in stopping quickly, sometimes to the detriment of passengers within.
I commented last night to one of your posts about computer vs humans that I have just started James Case's book, Competition: The Birth of a New Science, and that, in that, he addresses why machines are poorer at multi-player games, like Go, than two-player, zero-sum games.
I have not got fully into his argumentation yet, but you seemed to be implying that computers do well at Go. Is his data out of date (the book is copyright 2007 so was probably written in 2006), or is he still right that the very best human players of Go still beat machines?
AlphaGo is now beating the leading human players, as of this year.
Thanks. Is that for the normal version of Go, or a competition version of the rules that limits multi-player aspects and non-zero sum outcomes?
Go is a two-player game. Each side places black/white pieces to conquer the board. Rules are straightforward, but it is considered more complex than chess to master.
Just looking through some of the data, if Trump is going to end all amnesties and turn back illegal immigration I think that puts Nevada into play for the GOP and takes Texas and Georgia off the endangered list for them. For the former even after all of those reports of record early votes, record Hispanic turnout Trump closed the margin for the GOP. It also puts Colorado into play IMO.
Also, Trump has pulled ahead of Romney in terms of the raw vote. This idea that he pulled in fewer voters than Romney should be consigned to the dustbin because it isn't true. If anything Trump should beat Romney quite handily even with Johnson eating into Trump's base in safe GOP states.
Just looking through some of the data, if Trump is going to end all amnesties and turn back illegal immigration I think that puts Nevada into play for the GOP and takes Texas and Georgia off the endangered list for them. For the former even after all of those reports of record early votes, record Hispanic turnout Trump closed the margin for the GOP. It also puts Colorado into play IMO.
Also, Trump has pulled ahead of Romney in terms of the raw vote. This idea that he pulled in fewer voters than Romney should be consigned to the dustbin because it isn't true. If anything Trump should beat Romney quite handily even with Johnson eating into Trump's base in safe GOP states.
The margin in Michigan is the equivalent to a candidate in a typical UK constituency having a lead of just 120 votes.Such a margin would almost certainly result in a full - rather than a bundle - recount.
Interesting. The Dems seem to have won a number of R+1 and R+2 seats, indicating they did better in the House than an even contest, so more in line with the popular vote.
The margin in Michigan is the equivalent to a candidate in a typical UK constituency having a lead of just 120 votes.Such a margin would almost certainly result in a full - rather than a bundle - recount.
Neither particularly likely to be overturned, Michigan less so than a UK constituency.
I commented last night to one of your posts about computer vs humans that I have just started James Case's book, Competition: The Birth of a New Science, and that, in that, he addresses why machines are poorer at multi-player games, like Go, than two-player, zero-sum games.
I have not got fully into his argumentation yet, but you seemed to be implying that computers do well at Go. Is his data out of date (the book is copyright 2007 so was probably written in 2006), or is he still right that the very best human players of Go still beat machines?
Go is a two player game, it is "too complex" to 'solve'
Computers were terrible at it previously.
Deep Mind thrashed Lee Sedol, the world champ at it - Now. It doesn't 'solve' the game, but uses evolving algorithms as it plays itself.
Thought it was just up your street, though you're probably all to aware of it.
Thanks for the link. I had missed that particular article, but know the subject matter well. I wrote the Nature review of the authoritative book on this, Leitenberg and Zilinskas "The Soviet Biological Weapons Program" and I am good friends of long standing of the husband and wife team who did the initial Western epidemiological investigation - Jeanne Guillemin and Matt Meselson:
I think the article you cite somewhat overstates the achievements in stating that "However, the final product was infectious bacteria that could evade immune responses as well as multiple drugs." I think that was the aim, but getting all of the desired features into the bug at the same time was problematic, and preventing the bug from evolving out the new features was even more so. The article does go onto hint at this, but my review of Ray and Milton's book gives you more detail about the problems
Seems to be clutching at straws. If Trump doesn't want to quit he can overrule the declaration of him being incapacitated, as per the same amendment.
Unless he is actually incapacitated. Which given he didn't release his medical records and we now all believe the crazyiest wacked out conspiracy shit means he only has 3 months left to live.
After watching PMQs today, I have to say this as a Tory supporter, I was disappointed with May. She seems oddly uncomfortable at the dispatch box, at times overtly nervous (similar to Gordon Brown), and even somebody as ridiculous as Jeremy Corbyn has appeared vaguely competent in recent weeks. May and the Tories are lucky that Labour are not anywhere near a government in waiting, with a charismatic and popular leader, otherwise they could be in real trouble. Imagine a young Tony Blair at the dispatch box instead of Corbyn.
Incredible, also, that not a single MP had the courage to congratulate Trump's presidency. They may find him deeply objectionable, however, there are greater things at stake than tribal instincts and personal feelings - like trade deals and economic growth.
Just looking through some of the data, if Trump is going to end all amnesties and turn back illegal immigration I think that puts Nevada into play for the GOP and takes Texas and Georgia off the endangered list for them. For the former even after all of those reports of record early votes, record Hispanic turnout Trump closed the margin for the GOP. It also puts Colorado into play IMO.
Also, Trump has pulled ahead of Romney in terms of the raw vote. This idea that he pulled in fewer voters than Romney should be consigned to the dustbin because it isn't true. If anything Trump should beat Romney quite handily even with Johnson eating into Trump's base in safe GOP states.
Just looking through some of the data, if Trump is going to end all amnesties and turn back illegal immigration I think that puts Nevada into play for the GOP and takes Texas and Georgia off the endangered list for them. For the former even after all of those reports of record early votes, record Hispanic turnout Trump closed the margin for the GOP. It also puts Colorado into play IMO.
Also, Trump has pulled ahead of Romney in terms of the raw vote. This idea that he pulled in fewer voters than Romney should be consigned to the dustbin because it isn't true. If anything Trump should beat Romney quite handily even with Johnson eating into Trump's base in safe GOP states.
But his share is less than Romney.
Due to traditional voters in safe Dem/GOP seats going for Johnson, IMO. Trump will win the popular vote quite easily in 2020 if he runs.
Just looking through some of the data, if Trump is going to end all amnesties and turn back illegal immigration I think that puts Nevada into play for the GOP and takes Texas and Georgia off the endangered list for them. For the former even after all of those reports of record early votes, record Hispanic turnout Trump closed the margin for the GOP. It also puts Colorado into play IMO.
Also, Trump has pulled ahead of Romney in terms of the raw vote. This idea that he pulled in fewer voters than Romney should be consigned to the dustbin because it isn't true. If anything Trump should beat Romney quite handily even with Johnson eating into Trump's base in safe GOP states.
Eh. I don't know.
Me bringing it up officially makes it a cliche but demos in those states are still going to favor the Dems. There are going to be more Hispanic voters four years from now.
The margin in Michigan is the equivalent to a candidate in a typical UK constituency having a lead of just 120 votes.Such a margin would almost certainly result in a full - rather than a bundle - recount.
You can't scale up directly as the probability of overturning a result for a given percentage lead diminishes with the size of the vote, unless there are specific structural issues as to why the count might bias one way or another.
With a turnout of 100 in a town council election, a 1% lead is a single vote. On a turnout of 10k, it amounts to 100 votes. On a turnout of 1m, it amounts to 10k votes.
Finding a single miscounted vote might not be too difficult, even in a relatively small sample. Finding a hundred, all miscounted the same way, even on a sample a hundred times the size is a much bigger ask. If the miscounting is random then you'd expect errors to largely cancel out. However, there is the risk that in UK elections a bundle has been misallocated and so 100 votes shifts all at once. The chances of 10,000 votes *all* being miscounted the same way (or a net gain of 10k being made through errors), even on a huge sample is trivially small. Those with more recent advanced maths can crunch the numbers; I imagine there's a normal distribution that could be used to calculate the likelihood for given sample sizes.
Me neither, but I'm familiar with the work of the AAIB - they are the best in the world in their field.
As a starting point, a tram-activated or GPS-based siren sounding before the corner would make sure a sleepy or distracted driver hit the brake on time. More signage before the corner would also help.
More signage: definitely. As for this crash, we cannot assume it was the driver's fault. There might have been other factors causing, or contributing, to it.
Most railways prefer to use trackside balises to communicate with the train. Basically: if you pass this point and you're going above x MPH, apply full emergency brake.
Also AIUI, these trams are fitted with electrmagnetic eddy current brakes that are *very* effective in stopping quickly, sometimes to the detriment of passengers within.
Interested in more about eddy current brakes, these are what rollercoasters use for 2g or 3g stops, which would be damn uncomfortable in a tram but way better than a crash. If the Croydon trams have them, why are they not used as emergency brakes before these tight corners?
Obama deported 2.5 million people, so Trump deporting 3 million is a big deal, why?
The rhetoric will mean more than 3m will leave either by deportation or fear of deportation. There are 11m illegal immigrants in the US, over half could end up leaving within the next 4 years. It will change the nature of politics in border states if it does happen.
Seems to be clutching at straws. If Trump doesn't want to quit he can overrule the declaration of him being incapacitated, as per the same amendment.
Unless he is actually incapacitated. Which given he didn't release his medical records and we now all believe the crazyiest wacked out conspiracy shit means he only has 3 months left to live.
If we are all believing the craziest whacked out conspiracies, doesn't that mean we believe the records he did release during the campaign? If so, we have nothing to worry about.
After watching PMQs today, I have to say this as a Tory supporter, I was disappointed with May. She seems oddly uncomfortable at the dispatch box, at times overtly nervous (similar to Gordon Brown), and even somebody as ridiculous as Jeremy Corbyn has appeared vaguely competent in recent weeks.
Due to traditional voters in safe Dem/GOP seats going for Johnson, IMO. Trump will win the popular vote quite easily in 2020 if he runs.
Especially if the mainstream in 2018 has come around to his way of thinking on issues like infrastructure, immigration and Russia many of the Democratic candidates might think they're better off sitting it out than facing the full barrage of an election campaign against Trump.
After watching PMQs today, I have to say this as a Tory supporter, I was disappointed with May. She seems oddly uncomfortable at the dispatch box, at times overtly nervous (similar to Gordon Brown), and even somebody as ridiculous as Jeremy Corbyn has appeared vaguely competent in recent weeks.
How is it good for Mr Corbyn, when he spent his questions to get the answers he was expecting anyway, and for which the PM has popular support - before leaving the PM with an open goal at the end? Will he be asking the same question six times next week too, knowing he'll be getting the same answer six times?
Just looking through some of the data, if Trump is going to end all amnesties and turn back illegal immigration I think that puts Nevada into play for the GOP and takes Texas and Georgia off the endangered list for them. For the former even after all of those reports of record early votes, record Hispanic turnout Trump closed the margin for the GOP. It also puts Colorado into play IMO.
Also, Trump has pulled ahead of Romney in terms of the raw vote. This idea that he pulled in fewer voters than Romney should be consigned to the dustbin because it isn't true. If anything Trump should beat Romney quite handily even with Johnson eating into Trump's base in safe GOP states.
Eh. I don't know.
Me bringing it up officially makes it a cliche but demos in those states are still going to favor the Dems. There are going to be more Hispanic voters four years from now.
I'm not sure that there will be, the increasing demographics advantage relied on Clinton's 11m illegal immigrant amnesty in the border states. Without the amnesty, the process will be a lot slower. I remember a few liberal commentators saying that Texas would be ultra marginal because of it, how will the Dems make the gains without that massive increase in Hispanic voters.
Interested in more about eddy current brakes, these are what rollercoasters use for 2g or 3g stops, which would be damn uncomfortable in a tram but way better than a crash. If the Croydon trams have them, why are they not used as emergency brakes before these tight corners?
I've just done some 'research# (i.e. Googled) and I was wrong: they're magnetic brakes, but not eddy brakes:
The third braking system is the magnetic 'track brakes' system for emergencies, basically a 'shoe', each side of each bogie, each apply 6.66 tons of pressure onto the track and stops the tram very fast indeed, the shoes are assisted by applying sand onto the track to aid grip. This is sometimes known as the 'Hazard Brake' and all uses have to be reported to Controllers and logged.
As for why they're not automatically used: this is the first fatal crash on a UK tram system for many decades (as opposed to collisions with trams) - albeit there were virtually no trams for much of that time. They were probably thought 'safe' without them.
I would pitch the LibDems lower than that at 9/10% simply because they now have so much competition as the NOTA option.Many who in the past would have supported them on that basis will now opt for the Greens or UKIP.
I think you are running about two years behind the times, Mr Justin.
+1
Why be Green when you can love Corbyn, and why be UKIP when it's 'job done' and both Banks and Farage are about to take their energies and £ elsewhere?
The point is that in the past many people voted LibDem - or Liberal - despite not supporting their policy agendas at all. Many were anti - EU and quite a few had racist views more likely to be associated with the BNP.They voted LibDem simply out of a wish to reject both Tories and Labour.Today they continue to be beneficiaries of such votes - but have much more competition.
An interesting point, Mr Justin. I suppose that, in the same spirit, you would argue that very many people voted Conservative at the last election because they wanted the Coalition Government to continue. And many more did so because they liked that charming Mr Cameron and the utterly astounding Mr Osborne. (The fall in the Tory majority at Witney from 25.000 to only 5000 might support this point.)
Due to traditional voters in safe Dem/GOP seats going for Johnson, IMO. Trump will win the popular vote quite easily in 2020 if he runs.
Especially if the mainstream in 2018 has come around to his way of thinking on issues like infrastructure, immigration and Russia many of the Democratic candidates might think they're better off sitting it out than facing the full barrage of an election campaign against Trump.
There's bound to be one republican primary opponent who he will crush utterly too.
Due to traditional voters in safe Dem/GOP seats going for Johnson, IMO. Trump will win the popular vote quite easily in 2020 if he runs.
Especially if the mainstream in 2018 has come around to his way of thinking on issues like infrastructure, immigration and Russia many of the Democratic candidates might think they're better off sitting it out than facing the full barrage of an election campaign against Trump.
There's bound to be one republican primary opponent who he will crush utterly too.
It really depends, if he does a decent job the GOP might let him run unopposed.
Obama deported 2.5 million people, so Trump deporting 3 million is a big deal, why?
How screwed up is politics in the US, where the concept of deporting illegals is even seen as an issue?
If one overstays their welcome or misbehaves in any country they can be expected to be put on a plane back home, no?
Isn't it clouded a bit in the US by the fact that anyone born on US soil is automatically a US citizen? It means there are a lot of kids who are legal and a lot of parents who are illegal. If you deport the parents you deprive US citizens of a natural right and increase your own cost-base.
Comments
AlphaGo is now beating the leading human players, as of this year.
Why be Green when you can love Corbyn, and why be UKIP when it's 'job done' and both Banks and Farage are about to take their energies and £ elsewhere?
From what I've read the trams have a deadmans' handle-type thing, but there is very little signalling. This is because trams traditionally run along roads. The problem appears to be where the tram rums along an old railway line, and therefore can get up to highish speeds, before transitioning to on-street running.
In this, it's a bit like the Spanish high-speed rail crash of a few years ago, which occurred on a transition between high- and low-speed sections.
As an aside, the Cambridge (mis)Guided Bus has suffered a number of accidents over the last couple of years, including busses ending up on their sides. The investigations have appeared short, and the drivers blamed and sacked perfunctorily. This is, to my mind, a really bad thing, as causal factors won't be discovered.
At least RAIB's involved in Croydon.
Seriously? This isn't news - it's gossip of the lowest kind. Fake news is too kind.
Corbyn's Labour definitely looks to be in a piss weak position to me though.
Turkey has recently had a coup and has a war on its doorstep.
China and Russia run militaristic governments and are both in very expansionist mindsets (South China sea for China, Middle East and Eastern Europe for Russia).
The West is at peace and has no real immediate threat of getting sucked into wars on their own patch (terrorism excluded)
My thesis is that it is the proximity of the threat that yields these numbers. If war were to erupt in France with a danger of spilling into the UK, you'd see our numbers go up into the 80s and 90s very, very quickly.
Miss Plato, if you ask nicely, Lily Allen may apologise to you on behalf of Sky.
We need a representative democracy with dedicated MPs to make the difficult trade-offs that "The people" are incapable of. If you don't like the way they do it, you can vote them out next time. That is what our democracy is about.
Germany has forbidden the use of referenda for a reason.
https://docs.google.com/spreadsheets/d/133Eb4qQmOxNvtesw2hdVns073R68EZx4SfCnP4IGQf8/edit#gid=19
There have been accusations in high-level chess games of people going to the toilet to do a quick analysis of their position on a computer.
Go, like chess, is a game where the status is fully analyseable and I assume that tree pruning is still used for simple life-and-death situations of the armies. However there are so many more combinations available that depth-first analysis is pretty hopeless. I would assume that the latest AI machines are able to analyse the whole-board position (influence, territory) as well as target individual areas.
Programmes playing bridge or other card games have incomplete information and they tend to run multiple simulations consistent with the information that they do have and work out the best chance of optimising the result. (Humans cheat in Bridge by exchanging information illegally between them.)
If the RAIB are anything like the AAIB, they'll issue a comprehensive final report and not stop short of understanding all the factors behind the accident. They will make such recommendations as to prevent a single error from causing another similar accident. This will undoubtedly include a means of alerting the driver to an overspeed, and probably recommend research on automated brakeing systems.
According to the RAIB document Mr Pit posted, the tram slid on its side for 25 metres. Tram sides are mostly glass, and the ground surface uneven, with tracks, ballast shoulders and other items. The glass shatters, and people get trapped between ground and sliding tram.
Go is a two-player game. Each side places black/white pieces to conquer the board. Rules are straightforward, but it is considered more complex than chess to master.
I am not aware of any multi-player aspects.
http://blog.dilbert.com/post/153259019196/the-hypnosis-lawyer
"But if I have taught you anything this year it is that mass delusions – and small group delusions – are totally common and expected in life. On the other hand, the rise of the world’s most powerful super-hypnotist – who can’t win a jury trial – is unlikely..."
http://arstechnica.co.uk/science/2016/11/decades-after-deadly-lab-accident-a-secret-russian-bioweapon-decoded/
Thought it was just up your street, though you're probably all to aware of it.
AFAIK AlphaGo uses a huge neural network (i.e. pattern recognition) and built up a huge knowledge by playing millions of games against itself, as well as against humans.
From what I've read, there are problems with automated systems in trams, not helped by the plethora of different tram systems out there. I'm far from being a tramways expert though.
* Sometimes a lot more: http://rollercoaster.wikia.com/wiki/Highest_G-Force_on_a_Roller_Coaster
Also, Trump has pulled ahead of Romney in terms of the raw vote. This idea that he pulled in fewer voters than Romney should be consigned to the dustbin because it isn't true. If anything Trump should beat Romney quite handily even with Johnson eating into Trump's base in safe GOP states.
No.
As a starting point, a tram-activated or GPS-based siren sounding before the corner would make sure a sleepy or distracted driver hit the brake on time. More signage before the corner would also help.
State website gives:
http://miboecfr.nictusa.com/election/results/2016GEN_CENR.html
2,277,914 Trump
2,264,807 Clinton
http://news.sky.com/story/bored-schoolboy-arthur-heeler-frood-slept-rough-in-london-10659432
http://www.politicsandthelifesciences.org/Biosecurity_course_folder/readings/guillemin.pdf
I think the article you cite somewhat overstates the achievements in stating that "However, the final product was infectious bacteria that could evade immune responses as well as multiple drugs." I think that was the aim, but getting all of the desired features into the bug at the same time was problematic, and preventing the bug from evolving out the new features was even more so. The article does go onto hint at this, but my review of Ray and Milton's book gives you more detail about the problems
http://www.nature.com/nature/journal/v489/n7416/full/489364a.html?message-global=remove
I avoid cheating by repeatedly misremembering which convention is which and how to bid them...
AIUI heavy rail's had considerable problems with the reliability of GPS-based systems, though there is promise:
http://www.railway-technology.com/features/feature121215/
Most railways prefer to use trackside balises to communicate with the train. Basically: if you pass this point and you're going above x MPH, apply full emergency brake.
Also AIUI, these trams are fitted with electrmagnetic eddy current brakes that are *very* effective in stopping quickly, sometimes to the detriment of passengers within.
https://en.wikipedia.org/wiki/Balise
If not, who is likely win the popular vote there, the Dems or the GOP?
https://twitter.com/TheSunShowbiz/status/798645504288571392
https://twitter.com/redstate/status/798737906970087424
Dave Wasserman – Verified account @Redistrict
Clinton's lead in safe states just surpassed 2 million (+2.4%), but she's behind in swing states by 880k (-1.9%). https://docs.google.com/spreadsheets/d/133Eb4qQmOxNvtesw2hdVns073R68EZx4SfCnP4IGQf8/edit#gid=19 …
5:34 pm - 15 Nov 2016
twitter.com/doctor_oxford/status/798523007371276288
(Probably reflects the views of most of my family, incidentally).
Incredible, also, that not a single MP had the courage to congratulate Trump's presidency. They may find him deeply objectionable, however, there are greater things at stake than tribal instincts and personal feelings - like trade deals and economic growth.
Me bringing it up officially makes it a cliche but demos in those states are still going to favor the Dems. There are going to be more Hispanic voters four years from now.
With a turnout of 100 in a town council election, a 1% lead is a single vote. On a turnout of 10k, it amounts to 100 votes. On a turnout of 1m, it amounts to 10k votes.
Finding a single miscounted vote might not be too difficult, even in a relatively small sample. Finding a hundred, all miscounted the same way, even on a sample a hundred times the size is a much bigger ask. If the miscounting is random then you'd expect errors to largely cancel out. However, there is the risk that in UK elections a bundle has been misallocated and so 100 votes shifts all at once. The chances of 10,000 votes *all* being miscounted the same way (or a net gain of 10k being made through errors), even on a huge sample is trivially small. Those with more recent advanced maths can crunch the numbers; I imagine there's a normal distribution that could be used to calculate the likelihood for given sample sizes.
Interested in more about eddy current brakes, these are what rollercoasters use for 2g or 3g stops, which would be damn uncomfortable in a tram but way better than a crash. If the Croydon trams have them, why are they not used as emergency brakes before these tight corners?
If one overstays their welcome or misbehaves in any country they can be expected to be put on a plane back home, no?
The third braking system is the magnetic 'track brakes' system for emergencies, basically a 'shoe', each side of each bogie, each apply 6.66 tons of pressure onto the track and stops the tram very fast indeed, the shoes are assisted by applying sand onto the track to aid grip. This is sometimes known as the 'Hazard Brake' and all uses have to be reported to Controllers and logged.
https://croydon-tramlink.co.uk/info/trams/index.shtml
As for why they're not automatically used: this is the first fatal crash on a UK tram system for many decades (as opposed to collisions with trams) - albeit there were virtually no trams for much of that time. They were probably thought 'safe' without them.
It seems once they know who the president they don't particularly give a monkeys anymore.
And she doesn't have public support for her handling of Brexit.