Rotherham Council mass resignation.....
Discussion
http://www.southyorks.police.uk/news-syp/rotherham...
He took an 11 year old and 13 year old from a park and kept them overnight. He got three years.
How many months ago were the Twitterati and the Great and the Good intent on making Ched Evans public enemy number one? He's a role model?
What message does giving that child abductor 3 years send?
Not my values.
He took an 11 year old and 13 year old from a park and kept them overnight. He got three years.
How many months ago were the Twitterati and the Great and the Good intent on making Ched Evans public enemy number one? He's a role model?
What message does giving that child abductor 3 years send?
Not my values.
La Liga said:
A more extreme version of "very large".
Start from the prior probabilities, and move away when more specific information is forthcoming. Experiment after experiment shows mechanical reasoning out-performs human judgement.
And if the data driving prior probabilities is flawed? Where failure = catastrophic (possibly fatal) then the worst scenario needs to be included to set a boundary for the model. I'll add the model for the design and operation of the space shuttle as another example of where failure modes were discounted as having a probability that was vanishingly small so excluded from the design model. The discounted failure modes should have been included within the model and the risks mitigated. Start from the prior probabilities, and move away when more specific information is forthcoming. Experiment after experiment shows mechanical reasoning out-performs human judgement.
La Liga said:
We're talking about the starting point.
If X number of the population are sex offenders, then Y aren't. We don't work from the assumption 100 X exists in this scenario. We start from X and move away from there when specific information justifies us in doing so.
If we don't know the boundaries for the population of sex offenders then - by definition - "X" could be any value. If you have a good model where X is the only variable, you can run simulations with various values of X to define unfeasible solutions for high values of X. You can do similar where several variables are unknown, but the accuracy decreases.If X number of the population are sex offenders, then Y aren't. We don't work from the assumption 100 X exists in this scenario. We start from X and move away from there when specific information justifies us in doing so.
La Liga said:
Which is nothing to do with modelling. It's was the trust society and the systems in place, placed in doctors who were / are able to issue their own death certificates.
Shipman indicates why basing the concept of the design and operation of a critical system on prior probabilities is flawed.La Liga said:
Because you have quality mapping of your burglary offenders through intelligence and other aforementioned methods. A reduction in burglaries after tacking the offenders provides and indication you are being effective, but its delayed feedback (and subject to normal statistical variance) and doesn't help you resource based on offenders.
Like I say, just because you know how many victims are with burglary, that doesn't mean you can resource it correctly. Few offenders or many? Interesting to mention cross-border as this features quite heavily with burglary (on the more organised front). Data sharing methods and other means mean these people can still be identified and are identified.
Quality mapping costs money (resource), how will you justify this resource if the extent of the criminality is unknown i.e. the scale of the number of victims and the scale of the cost to society? Like I say, just because you know how many victims are with burglary, that doesn't mean you can resource it correctly. Few offenders or many? Interesting to mention cross-border as this features quite heavily with burglary (on the more organised front). Data sharing methods and other means mean these people can still be identified and are identified.
V8 Fettler said:
And if the data driving prior probabilities is flawed?
But it's not. We have lots of data sources including: - Reported crime,
- Recorded crime,
- The crime survey of England and Wales,
To anchor us to where we are most probable to be.
V8 Fettler said:
Where failure = catastrophic (possibly fatal) then the worst scenario needs to be included to set a boundary for the model. I'll add the model for the design and operation of the space shuttle as another example of where failure modes were discounted as having a probability that was vanishingly small so excluded from the design model. The discounted failure modes should have been included within the model and the risks mitigated.
I didn't say disregard the improbable, I said don't start from it. V8 Fettler said:
If we don't know the boundaries for the population of sex offenders then - by definition - "X" could be any value.
You will know the boundaries of the prior probabilities. V8 Fettler said:
Shipman indicates why basing the concept of the design and operation of a critical system on prior probabilities is flawed.
Shipman is an extreme example. Small samples have much higher probability of an extreme outcome. Larger samples, like sex offending in a constabulary, have greater statistical reliability. V8 Fettler said:
Quality mapping costs money (resource), how will you justify this resource if the extent of the criminality is unknown i.e. the scale of the number of victims and the scale of the cost to society?
Because it's fundamental intelligence work. Offenders commit crimes, therefore it's fundamental to know who they are. You will also now the extent of the criminality in other ways i.e. recorded crime, CSEW etc. La Liga said:
But it's not. We have lots of data sources including:
- Reported crime,
- Recorded crime,
- The crime survey of England and Wales,
To anchor us to where we are most probable to be.
You're on very shaky ground if you believe that any data set is perfect and complete.- Reported crime,
- Recorded crime,
- The crime survey of England and Wales,
To anchor us to where we are most probable to be.
La Liga] said:
I didn't say disregard the improbable, I said don't start from it.
If the primary aim is to avoid catastrophic failure then the design of processes to avoid that catastrophic failure needs to be the key building block and starting point. If you want to take the value engineered approach based on "it's not been a problem previously" then rely on prior probabilities and take the risk of another Rotherham/Chernobyl/FukushimaLa Liga said:
You will know the boundaries of the prior probabilities.
See above re: risk La Liga said:
Shipman is an extreme example. Small samples have much higher probability of an extreme outcome. Larger samples, like sex offending in a constabulary, have greater statistical reliability.
Shipman is an important example, it demonstrates that relying on "It's never been a problem before" is a dangerous approach where the results may be catastrophically fatal, Shipman could have been stopped at an early stage if someone had measured [average time since last visited by doctor before death]. I thought we had agreed that no-one knows the scale of sex offending in any geographical area?La Liga said:
Because it's fundamental intelligence work. Offenders commit crimes, therefore it's fundamental to know who they are. You will also now the extent of the criminality in other ways i.e. recorded crime, CSEW etc.
It may well be fundamental, but someone still has to pay for it and someone needs to justify the expenditure against an identified and quantified need, primarly because it's my tax money.
The Shipman case is used in police training isn't it?
Police training also mentions the Stephen Lawrence Inquiry, Report of an Inquiry by Sir William Macpherson of Cluny. Hasn't that report and the police pushing it in their training been largely discredited by the relationship between Norris's father and a dodgy Met police officer and the fact that it's now known that the police spied on the Lawrence family and Duwayne Brooks.
http://www.stephenlawrence.org.uk/news/theresa-may...
I'm not sure how off thread that is given the relationship between South Yorkshire Police and Councillors that failed in their attempts to cover up the wholesale sexual exploitation of minors in Rotherham.
Police training also mentions the Stephen Lawrence Inquiry, Report of an Inquiry by Sir William Macpherson of Cluny. Hasn't that report and the police pushing it in their training been largely discredited by the relationship between Norris's father and a dodgy Met police officer and the fact that it's now known that the police spied on the Lawrence family and Duwayne Brooks.
http://www.stephenlawrence.org.uk/news/theresa-may...
I'm not sure how off thread that is given the relationship between South Yorkshire Police and Councillors that failed in their attempts to cover up the wholesale sexual exploitation of minors in Rotherham.
Edited by carinaman on Saturday 28th February 07:27
V8 Fettler said:
You're on very shaky ground if you believe that any data set is perfect and complete.
Where did I say it was perfect? Little data for complex matters of this nature is ever going to be perfect, yet we need not just makeup a starting point because of its absence. V8 Fettler said:
If the primary aim is to avoid catastrophic failure then the design of processes to avoid that catastrophic failure needs to be the key building block and starting point. If you want to take the value engineered approach based on "it's not been a problem previously" then rely on prior probabilities and take the risk of another Rotherham/Chernobyl/Fukushima
The aim is to resource correctly and take a starting point in which to do so. We're not talking about improbable / high impact scenarios that cause a power plant to blow up. We're talking about the starting point for allocating police officers to begin to tackle a crime-type in an area. Allocating based on there being an improbably high amount of offenders / potential offenders, well beyond the national data is a reckless use of resources that increases risk in other areas without justification.
If the EDL come to do a protest in an area, do we put the whole national police service on standby, because X could theoretically be 10s of millions of people? Or can we make assessments based on the prior probabilities of similar protests and resource accordingly based on what is most probable for the initial resourcing / planning?
V8 Fettler said:
Shipman is an important example, it demonstrates that relying on "It's never been a problem before" is a dangerous approach where the results may be catastrophically fatal, Shipman could have been stopped at an early stage if someone had measured [average time since last visited by doctor before death]. I thought we had agreed that no-one knows the scale of sex offending in any geographical area?
No one is saying to rely on it, we're talking about a starting point that is probable rather than one that is improbable.If specific information (once developed) moves us to more extreme areas, then we can react accordingly. A rational approach.
You're talking about a lack of consideration, as oppose to a starting point. Different things.
V8 Fettler said:
It may well be fundamental, but someone still has to pay for it and someone needs to justify the expenditure against an identified and quantified need, primarly because it's my tax money.
It results in people being prosecuted for offences, which is one of the main functions of the police (to gather evidence in order to lead to prosecutions). La Liga said:
Where did I say it was perfect? Little data for complex matters of this nature is ever going to be perfect, yet we need not just makeup a starting point because of its absence.
I stated ">And if the data driving prior probabilities is flawed?"< You replied "> But it's not<" Meaning that it isn't flawed, therefore it must be perfect.La Liga said:
The aim is to resource correctly and take a starting point in which to do so. We're not talking about improbable / high impact scenarios that cause a power plant to blow up.
We're talking about the starting point for allocating police officers to begin to tackle a crime-type in an area. Allocating based on there being an improbably high amount of offenders / potential offenders, well beyond the national data is a reckless use of resources that increases risk in other areas without justification.
If the EDL come to do a protest in an area, do we put the whole national police service on standby, because X could theoretically be 10s of millions of people? Or can we make assessments based on the prior probabilities of similar protests and resource accordingly based on what is most probable for the initial resourcing / planning?
We are talking about flawed design and management processes leading to catastrophic failure. The objective should be to correctly allocate the resource to deal with the issue from day one. If the scale of the extent of the problem is not accurately identified then the level of resource will either be too low (ineffective) or too high (a waste of my tax money). For your EDL example, I'm sure that most of the victims of EDL criminality will report the crimes to the police, thus enabling the authorities to measure the scale of the problem over time. Many of the victims of CSE in Rotherham did not report the crimes to the authorities, so the authorities had (and have) only a vague view of the level of crime.We're talking about the starting point for allocating police officers to begin to tackle a crime-type in an area. Allocating based on there being an improbably high amount of offenders / potential offenders, well beyond the national data is a reckless use of resources that increases risk in other areas without justification.
If the EDL come to do a protest in an area, do we put the whole national police service on standby, because X could theoretically be 10s of millions of people? Or can we make assessments based on the prior probabilities of similar protests and resource accordingly based on what is most probable for the initial resourcing / planning?
La Liga said:
No one is saying to rely on it, we're talking about a starting point that is probable rather than one that is improbable.
If specific information (once developed) moves us to more extreme areas, then we can react accordingly. A rational approach.
You're talking about a lack of consideration, as oppose to a starting point. Different things.
The best starting point is the right starting point, this ensures effective use of my tax money. The early detection of Shipman could have occurred if it had been understood by the authorities that nobody in a position of absolute power should be absolutely trusted, thus enabling processes to be put into place to "flag up" doctors murdering patients. Additionally, the police should have picked up on the key factor of [average time since last visited by doctor before death] at an early stage in their investigations. If specific information (once developed) moves us to more extreme areas, then we can react accordingly. A rational approach.
You're talking about a lack of consideration, as oppose to a starting point. Different things.
You've mentioned "probable" and "improbable", but you've neglected "possible".
La Liga said:
It results in people being prosecuted for offences, which is one of the main functions of the police (to gather evidence in order to lead to prosecutions).
Which requires resource, how are you going to justify that resource to resolve that problem unless you know the scale of the problem? Increase in measured crime = increase in resource.
V8 Fettler said:
I stated ">And if the data driving prior probabilities is flawed?"< You replied "> But it's not<" Meaning that it isn't flawed, therefore it must be perfect.
In what world does "not flawed" = "perfect"? There are obviously degrees between. V8 Fettler said:
We are talking about flawed design and management processes leading to catastrophic failure.
No, we're talking whether or not you can resource without knowing the children unknown to services. We're not talking about power-plants, space ships, or extreme murdering doctors, we're talking about your assertion there were "no means" to plan to resource, when clearly there are means. V8 Fettler said:
But - as Jay stated - there is no data available concerning CSE where the victim is not known to any agency. Without this data there is no means to accurately identify trends (are we winning?), or to plan resource (do we have enough to continue to win?).
V8 Fettler said:
For your EDL example, I'm sure that most of the victims of EDL criminality will report the crimes to the police, thus enabling the authorities to measure the scale of the problem over time. Many of the victims of CSE in Rotherham did not report the crimes to the authorities, so the authorities had (and have) only a vague view of the level of crime.
It has little to do with criminality. There's relatively little criminality at such protests relative to the numbers and risk. The intelligence is often ambiguous and / or deliberately misleading and there are quite a few unknowns, yet the police manage to resource the demonstrations effectively. If we apply the same standards as assuming the number of offenders / potential offenders will be "very large", as oppose to inline with probabilistic data, then we'd end up grossly over-estimating the number of protesters and grossly over-resourcing every time when faced with unknowns.
V8 Fettler said:
The best starting point is the right starting point, this ensures effective use of my tax money.
The right starting point is circumstantial. When facing a complex task with more general information, the right starting point is one inline with the base rates. When more specific information is available to move us away from this probable default, then that becomes the right starting point. Things require actioning with incomplete information. V8 Fettler said:
You've mentioned "probable" and "improbable", but you've neglected "possible".
Improbable encompasses 'possible' by its very definition. V8 Fettler said:
Which requires resource, how are you going to justify that resource to resolve that problem unless you know the scale of the problem?
You're defining "problem" at the bottom of the sequence i.e. victims / recorded crime. The problem is also "offenders", their scale and specific information. If you know who they are, you can target them. Prevention is better than the cure. No offenders = no crime. La Liga said:
In what world does "not flawed" = "perfect"? There are obviously degrees between.
Not flawed = flawless = perfectLa Liga said:
No, we're talking whether or not you can resource without knowing the children unknown to services. We're not talking about power-plants, space ships, or extreme murdering doctors, we're talking about your assertion there were "no means" to plan to resource, when clearly there are means.
The process for ensuring that children in Rotherham were safe was clearly flawed by design and by management. Part of the problem was the lack of accurately directed adequate resource. This is a common thread where catastrophic failure occurs in other scenarios.La Liga said:
It has little to do with criminality. There's relatively little criminality at such protests relative to the numbers and risk. The intelligence is often ambiguous and / or deliberately misleading and there are quite a few unknowns, yet the police manage to resource the demonstrations effectively.
If we apply the same standards as assuming the number of offenders / potential offenders will be "very large", as oppose to inline with probabilistic data, then we'd end up grossly over-estimating the number of protesters and grossly over-resourcing every time when faced with unknowns.
You've highlighted the key issue at Rotherham: the lack of knowledge regarding the scale of the criminality, hence your EDL analogy is flawed.If we apply the same standards as assuming the number of offenders / potential offenders will be "very large", as oppose to inline with probabilistic data, then we'd end up grossly over-estimating the number of protesters and grossly over-resourcing every time when faced with unknowns.
La Liga said:
The right starting point is circumstantial. When facing a complex task with more general information, the right starting point is one inline with the base rates. When more specific information is available to move us away from this probable default, then that becomes the right starting point. Things require actioning with incomplete information.
So what is the "right" starting point for CSE when the base rates are unknown because the scale of CSE is unknown.?La Liga said:
Improbable encompasses 'possible' by its very definition.
To a reasonable person, possible is more likely than improbable, as in the expression: "It's very possible that .... " meaning "It's very likely that... ". A numerical scale provides a better quantitative description.La Liga said:
You're defining "problem" at the bottom of the sequence i.e. victims / recorded crime. The problem is also "offenders", their scale and specific information. If you know who they are, you can target them. Prevention is better than the cure. No offenders = no crime.
V8 Fettler said:
Not flawed = flawless = perfect
Not fatally flawed, if I need to be explicit and you genuinely think I believe data like this will be perfect. V8 Fettler said:
The process for ensuring that children in Rotherham were safe was clearly flawed by design and by management. Part of the problem was the lack of accurately directed adequate resource. This is a common thread where catastrophic failure occurs in other scenarios.
We're talking about what is needed and not needed to resource. We're not talking about the flawed approach which created the issues over Jay's time period. V8 Fettler said:
You've highlighted the key issue at Rotherham: the lack of knowledge regarding the scale of the criminality, hence your EDL analogy is flawed.
The key issue was not acting upon the information they had and going through established processes and procedures that are the foundation for us having full prisons. The EDL analogy is sound, you can resource properly with unknowns.
V8 Fettler said:
So what is the "right" starting point for CSE when the base rates are unknown because the scale of CSE is unknown?
There are various data sets for sex offenders to work from. Assuming an area has a sex / potential sex offender population grossly larger than what we see nationally, over a large sample, isn't the right starting place. Considering extremes to be possible when starting out, I agree with. I'm not suggesting we don't consider the atypical or improbable, and have plans to manage that. I'm talking about making a practical decision of where to start from when addressing a problem.
V8 Fettler said:
To a reasonable person, possible is more likely than improbable, as in the expression: "It's very possible that .... " meaning "It's very likely that... ". A numerical scale provides a better quantitative description.
There are degrees of probable and improbable, which cover all ranges of "possible". V8 Fettler said:
Odd that you place the victims at the bottom, I would place them at the top of the list of priorities and at the start of the sequence. How do you know there are no offenders unless you are closing the information loop by measuring the scale of the number of victims?
I've made no comment as to their priority. I've said offender-focused data gathering and targeting allows sound resourcing in itself. Ideally every aspect will be taken care of, but that's not the context, it's about what information and data we need to allocated some of our finite resources. I'm not talking about closing the data loop, I'm talking about resourcing based on knowing your offenders.
RobinOakapple said:
Is anyone apart from the two participants reading this stuff?
I do have a look now and again to see if anyone else has posted but always just see a mirage of quotes and unquotes, it almost gives me a headache, so quickly click 'back'. Yours is quite refreshing to see.
jogon said:
RobinOakapple said:
Is anyone apart from the two participants reading this stuff?
I do have a look now and again to see if anyone else has posted but always just see a mirage of quotes and unquotes, it almost gives me a headache, so quickly click 'back'. Yours is quite refreshing to see.
TLDNR for me every time, I pop back here to see if there is any meaningful update, even a cross reference to the arrests Manchester way today, but no....
I often Wonder if they are both Involved in some way in this case and posting relentless quotes on the semantics is just a form of obfuscation so the rest of us don't read or contribute to the thread.......
t bingo.
La Liga said:
Not fatally flawed, if I need to be explicit and you genuinely think I believe data like this will be perfect.
You now agree that the data is flawed, but you state that it might not be fatally flawed. Goalposts, for the moving of.La Liga said:
We're talking about what is needed and not needed to resource. We're not talking about the flawed approach which created the issues over Jay's time period.
The flawed approach at Rotherham occurred because there were/are not enough checking/hold/referral points within the existing system, primarily because of the insular nature of the two "authorities" involved. You have stated that similar (but to a lesser degree) has occurred elsewhere, this rings alarm bells.La Liga said:
The key issue was not acting upon the information they had and going through established processes and procedures that are the foundation for us having full prisons.
The EDL analogy is sound, you can resource properly with unknowns.
The EDL analogy is sound, you can resource properly with unknowns.
The key issue is that the existing insular management systems permitted the two failing organisations concerned to continue serving their own interests rather than prioritise the people they are supposed to protect. That insular incompetence needs to be eliminated, particularly as it involves my tax money.
Full prisons is a measure of the effectiveness of deterrence, i.e. current deterrence is not particularly effective, if it was then the prisons would not be full because the offenders wouldn't offend due to the effectiveness of the deterrence.
Re: unknowns, are you referring to unknown criminals offending against unknown victims? If so, how do you know the scale of the criminality? You can only guess, I would prefer that my tax money wasn't used for guesswork.
La Liga said:
There are various data sets for sex offenders to work from. Assuming an area has a sex / potential sex offender population grossly larger than what we see nationally, over a large sample, isn't the right starting place.
Considering extremes to be possible when starting out, I agree with. I'm not suggesting we don't consider the atypical or improbable, and have plans to manage that. I'm talking about making a practical decision of where to start from when addressing a problem.
Considering extremes to be possible when starting out, I agree with. I'm not suggesting we don't consider the atypical or improbable, and have plans to manage that. I'm talking about making a practical decision of where to start from when addressing a problem.
So what is the scale of CSE offending then? If you have that information then please forward it to Jay. It's currently an unknown. If you are risk adverse then you start at the worst case scenario (see UK designs for AGR nuclear power plants), if you want to take a chance with starting with something based solely on previous experience then be prepared for Chernobyl/Fukushima/Rotherham.
La Liga said:
There are degrees of probable and improbable, which cover all ranges of "possible".
And in the real world, "improbable" typically means so unlikely that it can be discounted. But as I said, a numerical means to measure probability is preferred, to eliminate misunderstanding leading to error. La Liga said:
I've made no comment as to their priority. I've said offender-focused data gathering and targeting allows sound resourcing in itself. Ideally every aspect will be taken care of, but that's not the context, it's about what information and data we need to allocated some of our finite resources.
There's an issue here with a mindset that places victims at the bottom in any context. I'll ask the same question again: How do you know there are no offenders unless you are closing the information loop by measuring the scale of the number of victims?La Liga said:
I'm not talking about closing the data loop, I'm talking about resourcing based on knowing your offenders.
So how do you identify the scale of the number of offenders you don't know? More guesswork perhaps?RobinOakapple said:
Is anyone apart from the two participants reading this stuff?
Yes. Me.Nobody cares about the kids getting screwed do they, because it's not their kids getting screwed.
Isn't the need to keep the police brand pristine part of the reason they covered up Hillsborough and lied about the victims at that football stadium?
The victims are to blame in Rotherham. Just as they were at Hillsborough.
V8 Fettler said:
@ RobinOakapple, jogon, HarryW, spadriver and smegmore. Are you not concerned about the waste of your tax money in Rotherham?
Very, thats why its possibly best to leave the main three posters to it and start another thread-perhaps on Child abuse in Oxford.This one used to be interesting.
spadriver said:
V8 Fettler said:
@ RobinOakapple, jogon, HarryW, spadriver and smegmore. Are you not concerned about the waste of your tax money in Rotherham?
Very, thats why its possibly best to leave the main three posters to it and start another thread-perhaps on Child abuse in Oxford.This one used to be interesting.
Gassing Station | News, Politics & Economics | Top of Page | What's New | My Stuff