QTRA - I'm sorry i don't agree with it!

[Albedo]

Nope, five words too long.

 

QTRA is a subjective assessment using numbers.

 

But I would score that as a 2 on the libel scale...

 

Fraudulent = obtained, done by, or involving deception, esp. criminal deception:

 

Nobody would deny that statistics can deceive and by definition fraudulent is not exclusively criminal.

 

So not libelous your honour:001_smile:

[Paul Barton]

Daltontrees: How are you doing this risk assessment of 8000 trees if you don't know what QTRA is?

Edited November 27, 2012 by Paul Barton
Specifying how question is aimed at.

[Amelanchier]

Fraudulent = obtained, done by, or involving deception, esp. criminal deception:

 

Nobody would deny that statistics can deceive and by definition fraudulent is not exclusively criminal.

 

So not libelous your honour:001_smile:

 

It's quite a leap from "stats can be used to decieve" to "QTRA is fradulent" though isn't it? We could just as equally say; "words can be used to decieve, Albedo's arguement contains words, therefore Albedo's arguement is fraudulent"...

 

Affirming the consequent.

[Albedo]

It's quite a leap from "stats can be used to decieve" to "QTRA is fradulent" though isn't it? We could just as equally say; "words can be used to decieve, Albedo's arguement contains words, therefore Albedo's arguement is fraudulent"...

 

Affirming the consequent.

 

I thought twice about the use of that word at the time but was too lazy to think of a less offensive one that meant the same thing.

 

However...assigning a statistical probability to something that you can't and "don't know' is potentially deceitful.

 

My earlier builder scratches chin analogy was more serious than you think.

 

I.e. "that's gonna fail at some point, possibly,...could be tomorrow....could be next year...could be never...I don't know"

 

I stand by the concept that such assigning of a probability, is ........... (insert your own word here...but means...'not true')

[Albedo]

My beloved macbook has a handy feature that I recently discovered by accident.

You just touch on a word with two fingers, and it gives a pop up menu, which includes a direct route to a dictionary.

 

I've never needed it so much:001_smile:

 

I'm currently working out what "affirming the consequent" might mean:001_smile:

[RobArb]

What a thread!

 

Sent from my Galaxy S2

[Albedo]

What a thread!

 

Sent from my Galaxy S2

 

I'd quite like to have Tony Crofts' input here.

 

Come on Tony, you can break that recent promise....This is soooooooooooooooooo your thing:thumbup1:

[HCR]

Wow, you guys are giving this subject a good seeing to!

 

I amn't sure what this QTRA is, is it that sysstem devised by Forbes Laird that you have to pay for to get a license to use?

 

This winter I am surveying around 8,000 trees for a nearby local authority. The contract specifies (words to this effect) that trees must pass the 1:10,000 risk of serious harm test or works must be specified to bring the tree within that probability. Evey single day for every single day I quantify the risk, there is no scope for pontification about the benefits or drawbacks of statistics or the question of whether a tree is Ok or not. It either passes or fails the threshold. And the more I have read your interesting and entertaining debat the more I am reassured that (i) without quantification of risk the whole business is impossible and (ii) most aspects of quantification are not that hard really.

 

Read any survey report, if the author is worth a tosser the report will make it very clear (i.e. not just buried in the small print and disclaimers) that there is no such thing as a safe tree and that the report can only try and predict that the likelihood of harm is or is not above the threshold of acceptibility.

 

The difficulty therefore lies often with the lack of client/public understanding with the limitations (and I don't mean cop-outs) of tree surveys and reports.

 

That's my view anyway. It keeps me sane out on survey day after day after day.... when otherwise I would go ga ga.

 

So that means that there is an 80% chance that one tree you pass as safe (within the scope of the job) will fail anyway

[Amelanchier]

However...assigning a statistical probability to something that you can't and "don't know' is potentially deceitful.

 

My earlier builder scratches chin analogy was more serious than you think.

 

I.e. "that's gonna fail at some point, possibly,...could be tomorrow....could be next year...could be never...I don't know"

 

I stand by the concept that such assigning of a probability, is ........... (insert your own word here...but means...'not true')

 

I disagree (surprise).

 

I don't know whether I will get a flat tyre on my bike tomorrow. As I have at least one tyre that at least raises the probability above zero but I think its unlikely that I will be certain to catch a flat (a probability of 1), so its somewhere in between. A useless prediction you say? But its less likely to happen if I don't use the bike, and more likely to happen if I ride through hawthorn hedge flailings. So we can usefully say that there are situations within which we can comparatively assess risk. Now where's the deception in me estimating that I stand a probability of 0.1 of getting a flat if I don't use the bike compared with 0.8 if I ride through thorns? Who have I misled by using numbers to represent my estimation?

 

It would be deceptive if I presented those figures to you as objective measured probabilites; but QTRA doesn't do that.

 

I guess you might feel decieved if you thought I was referring to objective measured probabilites when I wasn't, but if I've been clear enough in my method then that's kind of your problem.

[Amelanchier]

So that means that there is an 80% chance that one tree you pass as safe (within the scope of the job) will fail anyway

 

Interesting! Or is it that 0.8 trees will fail? Now that's confusing