There seems to be a lot of misunderstanding surrounding this. Hopefully this clarifies the matter :
A traditional high-low test involves moving one wheel at a time away from the area where we have found a suspected gate.
When the wheel containing the suspected gate is moved away to a higher or lower test number, the suspected gate/low spot will have been moved out from under the fence, and therefore the contact area will
increase - because the low spot is no longer under the fence.
For example, lets say we have found a suspected gate at 47, we then dial the following test combinations & get these results :
20 47 47 9.0 (this test positions wheel one elsewhere)
47 20 47 9.0 (this test positions wheel two elsewhere)
47 47 20 9.2 (this test positions wheel three elsewhere)
We can see that when wheel three is positioned elsewhere(at 20)the contact area increases. This tells us that moving wheel three took the low spot away from being underneath the fence. So the low spot/suspected gate is on wheel three.
Put another way, when wheel three is not set to 47, our good indication disappears, and when wheel 3 is at 47, our good indication is present. Therefore the low spot/gate we have found is on wheel three.
These tests must be performed taking rotational direction into account. If a low spot was found with left rotation, that number must be dialled with left rotation in all test combinations. 47L 47R 20L is not the same as 47L 47L 20R.
Some people choose to perform a slightly different test, where two wheels are thrown off from the suspected gate. Again, with a suspected gate at 47, using a test number of 20, the test combinations would be as follows :
47 20 20
20 47 20
20 20 47
If testing in this manner, the test combination with the smallest contact area will identify the wheel that the suspected gate belongs to.
The traditional high-low test, that throws off only one wheel at a time from the area of interest is more likely to be accurate than this alternative method. That said, neither method is fail safe.
On another note :
webpirate wrote:The smaller the gap the more likely you have a correct number. But only if the range is 1-2 numbers on either side. If the gap gets smaller for 4 or more numbers you can ignore that and move on.
is wrong.
It is true that narrower contact areas correlate with gates, but definitely NOT only when the indication is 1-2 numbers wide! I have never seen a gate signature only 1 increment wide & doubt I ever will.
As Mark Bates, a master of manipulation, describes here
http://www.safeventures.com/news.php?news_id=16, a common gate signature that is often missed is when the contact area narrows, but does not widen again, so do not ignore indications that are more than 4 numbers wide!
...Mark