Search This Blog

Saturday, August 17, 2024

8/17/24 Report - Supermoon Coming Soon. Importance of Testing. Two Important Types of Metal Detecting Errors.

 


ti

CAPE CANAVERAL, Fla. (AP) — The first of four supermoons this year rises next week, providing tantalizing views of Earth’s constant companion... 

Here is the link for more about this years supermoons, the first of which is next week.

How to watch August's supermoon, which kicks off four months of lunar spectacles | AP News

Watch for high tides.

----


Gold Rush; Mine Rescue was on yesterday and I noticed something.  What is the first thing they do?  Test.  And what do they do next?  More testing.  Before they make any suggestions or changes, they test the system and the current procedure.  They pan what is coming out of the sluice or whatever to see how well the equipment is catching the gold, and then they do a larger test, running for hours and checking the results.  

Then what do they do after they make the recommended changes?  They test again to see how much improvement (if any) there was.

One thing I often recommend is testing.  Test your metal detector at home on various targets.  And test it at the beach, in salt sand or wherever you hunt.  And then test yourself every once in a while, in the field.

I've said before that you never know what you miss, and that is one reason it is so important to do tests.   You can continue doing something unproductive for a long time and never find out unless you do some tests.

Maybe I'm wrong, but I think most detectorists do pretty much the same thing most of the time.  They might use the same detector, use the same settings, and even visit the same beaches and work the same part of those beaches over and over.  

One of the things I like to test is the junkiness of the site or area I am detecting.  There is no reason to discriminate when there is no, or almost no, junk. But how do you know when there is no junk?  You might detect for a little while, digging up everything, or almost everything, simply to find out what kinds of things are there.  

If your test shows that there is very little junk, there is little reason to discriminate or even bother much with target ID.  

Of course, if you are in wet sand, it generally takes longer to recover targets, so the cost of digging junk is higher.  That can be taken into account. 

Referring to the illustration above, there is no reason to discriminate at either end (all junk or all good targets).  The area between one and three would probably most justify discrimination.  You would be trying to pick some good targets out of a lot of junk.

Remember that there will always be some mistakes.  Some junk targets will be dug and some good targets will be missed.  That should be taken into account along with the expected value of the targets.  If you expect an especially high value target and want to make sure to not miss anything, your no-dig criteria would be adjusted accordingly.

One thing you should try to determine is your error rate.  How often do you misidentify good and junk targets.  That requires some testing.  Again, you quickly find out when you dig junk but may never find out what you miss.  As a result, there is a natural tendency to over adjust to miss junk, and the natural result of that is missing more good targets.

Unfortunately, the things that are most clearly signaled by most metal detectors are common items such as coins.  They will give a good coin tone or conductivity number and you can identify common coins pretty well, which is good if you are targeting common coins.  But how often do you want to target common coins.  The problem is the best targets are not the most common targets and are often difficult to identify by tone or ID number.  As a result many are missed if you are going by easily identifiable signals or easy to identify readouts.  Take a Rolex watch, for example.  What detector will tell you the target is a Rolex watch?  Most are missed by most detectorists who discriminate or try to avoid junk.

As you know, signals and readouts will vary when you go over a target depending upon several factors that I've mentioned before.  You won't always get a 22, or whatever the number is for a penny on your detector.  Sometimes it might be 21 or 20 when you swing back over the target, or sometimes you might even get a 30 or something as the result of electrical or radio interference of some other source of noise in the area.  On average you might get a 22, but the number will sometimes be a little different.  The distribution of the average number will vary some.  The distribution signal or readout and any noise or variance can be represented as a curve with the average being the highest point in the center.

Below is an illustration representing two signals, one (A) being a junk item, and the other (B) representing a good target.


The vertical axis would be the probability of the specific readout for that item and the horizontal axis, the specific conductivity number.

The area under the blue curve represents the distribution of the signal or readout around the average for the junk target - let's say it is 17.  The area under the red curve, the distribution for the good target - let's say 22.  The curves overlap, meaning that sometimes the signal or readout could be mistaken.  For items that give readouts that are more different, the curves would be farther apart and there would be less overlap, and you would less frequently mistake the junk item for good target (false positive) or vice versa (false negative).

The shape of the distribution will vary depending upon the precision of your detector and how you operate it.  Cutting down on noise will narrow the curves and decrease the overlap.   How you operate the detector can also increase your precision or accuracy and make the readouts more consistent, narrowing the curves and decreasing both types of errors.

I find it useful to consider the amount of junk relative to good targets, the cost and values, and the effect of varying the dig no-dig criteria on both finds as well as both types of error.

I'm always trying to simplify this and represent this more clearly, but there are so many variables that it always seems to get complex and confusing and I have to quit before feeling that I've done the job. That is the case now.

---

Sourcea; surfguru.com.

According to SurfGuru we're supposed to have three to four feet today.  Not bad.  

Also, the second high tide today is supposed to be pretty good.  

Maybe worth checking some of the more vulnerable areas.

Good hunting,

TreasureGuide@comcast