@parados,
No surprises here. Now I have to explain statistics ? You will have to rely on more formal methods to improve your education. When you go to High School and begin to learn statistics, you will find the first point is to list all possibilities and give them an equal chance of occurring. Then begins the analysis. Why is that so hard to understand ?
Quote:No, I never said they were all wrong in one direction. Random errors are just that, random.
Random errors don't normally affect a trend line in a large series of data.
This is not random error. Every single reading is inaccurate. It is the only data available and Global Warming Thuggees who like to cite that references are written by superman needed something to justify their stance. The first question they were asked was what are you basing Global Warming on. Totally devoid of accurate data from the past, they seized on current weather data. Weather fluctuates far more widely than climate. Climate by its very nature is a long term trend. Needing to show a trend, and totally lacking any observable data for the earth's 3.2 billion years it has had a climate, they use the only data available. 30 years of satellite measurements and say they have enough for a trend. Prior to that, they have weather stations with increasing inaccuracy as you go back in time.
The data collection made by climatologists from weather stations is inaccurate. You seem to be repeatedly missing this point so I will explain it for a 8 year old. You wake up in the morning and it is very cold at 2 deg C. It stays that way till 14:00 when the temp climbs quickly to 10 deg C because the sun came out . By night time, the sun has gone down and it is very cold again. Most of the day, from midnight when you were asleep till the next midnight when you were asleep again, it was very cold. The average temperature was 2.7 deg C. The mid point between the max and min temps is 6 deg C. In the town nearby, there was less cloud and rain so the temp range was 2 deg C to 15 deg C. The mid point is 8.5 deg C but the average temp was 10 deg C. The first town has an error of + 3.3 deg C. The second town has an error of - 1.5 deg C. The third and fourth towns have no weather stations because they are of no interest to meteorologists and have no airfield.
With a random number generator you can also "create" numbers. Why measure anything ?
This means that to assume the mid point is always the average is wrong every single time. Why dont we do this everywhere else in science ? Because it is nonsense. If every single piece of data is inaccurate, apart from why are you using it, any trend you find is random. Are you unaware that some numbers come up more often then others in Lotto ?
Quote:No, I never said they were all wrong in one direction.
If you find a warming trend and you have said that, and they are all wrong then yes, they are all wrong in one direction.
Quote:Simple problem - Graph the numbers 1 - 10. You will see a trend line.
Now let's introduce errors.
Subtract 2 from each number and graph it again.
You are graphing the numbers -1 to 8. The trend line is exactly the same as the first graph just slightly lower on the graph.
Are you assuming the errors are all in the same direction ?
Quote:Now let's introduce random errors at a 50/50 rate.
subtract 2 from the even numbers and add 2 to the odd numbers.
Graph that and tell us what the trend is.
The only way you can have a 50/50 rate is if the original assumption is correct and that is : the mid way point is also the average and errors will be evenly distributed about the average. This also assumes all temperatures, everywhere, always follow a bell curve. Whilst this is good enough for weather prediction, as it quickly becomes obvious if it was accurate or not and then they move on to the next prediction, it is not good enough to demand world wide change.
It is usual in science to measure any error and include it in the data and say why it has been discounted. This is not done because no one knows the original error. By accurately plotting the original data and then forming a trend line, one can visually demonstrate a trend and error. It is unheard of to take data that is known to be erroneous and make an assumption it must be correct because the final decision has already been made.