Written by the TreasureGuide for the exclusive use of the Treasure Beaches Report.
I'm going to start what might become a lengthy discussion of discrimination today. I'll start with an illustration that I originally made and posted over ten years ago. If I made the same chart today, I'd do it a little differently today, but it is still useful and handy as a starting place.
The illustration was made when I was targeting mostly gold rings. Ignoring for the moment some of the exceptions, gold rings are generally more valuable when they are big and contain more gold. The weight of the gold is one of the primary determinants of a rings value. Of course there are other things that will affect value, such as valuable stones or important brand names. The illustration simply shows that large rings are in general more valuable along with the generally smaller ladies rings with valuable stones.
Since value is correlated with size and, since conductivity and discrimination is also related to size, if you discriminated out the smaller items, you could also be discriminating out many smaller valuable rings. That is something that actually happened to me when I was using the Fisher 1280 Aquanaut with linear discrimination. As I said before, there was a time when I thought men lost many more rings than women, but that was because I was discriminating out things that included smaller gold rings. After turning down my discrimination setting, I discovered that there were many lost ladies rings to be found. That teaches a variety of important lessons. One of those is that you can't confuse what is out there with what you are finding if you use discrimination. To put it another way, you never know what you miss.
Back then the discrimination control on most metal detectors was a simple linear control. You would have a knob, and you could turn the discrimination setting up or down. The conductivity numbers were not displayed and you had no digital graphic display to help with target ID. With few exceptions, it was as simple as a single knob. When you increased discrimination to eliminate pull tabs, you would also lose a lot of smaller rings. That could be a big mistake because the smaller rings were often valuable too.
The two red lines on the illustration represent two different discrimination settings. Everything to the left of the respective line is discriminated out. The left most line, would not discriminate out the bulk of women;s rigs. While the line representing a higher discrimination setting would. The illustration also shows that rings with less gold than large men's rings such as women's engagement rings can be valuable, but they, being smaller would often be discriminate out.
As modern metal detectors clearly show us, the conductivity numbers of small gold items can be the same as other items such as pull tabs. You therefore have to make a strategic decision: Is it really worth discriminating that kind of junk if you also lose some of the more valuable items. For making the optimal decision, several factors should be considered.
There are personal factors. What limitations are you dealing with? Do you have plenty of time? What is your physical condition? What are your goals? Can you tolerate digging junk? Are you patient? Will you have multiple shots at the same area?
Then there are factors that have to do with the detecting conditions. Do you have a good idea of what is in the ground? Do you know the history of the area? Have you done some sampling that would tell you the amount and type of junk and targets that probably remain? Will the tides, weather and other things be changing or is it a stable environment?
And there are other factors having to do with your detector and skill level. And your recovery skill and speed. There is less cost involved in digging a piece of junk if you can pinpoint and recover it quickly and easily, for example.
These days many buyers, such as pawn shops will not include gems as contributing much if anything to what they are willing to pay for gold jewelry. That has to do with several things, one of which is the recent availability and popularity of lab made diamonds and other developments in the gemstone market. So, if I made that same illustration today, the bump in value around women's rings would probably not be as high. Valuable rings in that discrimination range are therefore less common and many of the rings in that range are now less valuable. It that is the way you value rings, the value of rings is now better (but not perfectly) correlated with the amount of gold in a ring. That should be considered in your discrimination setting if you are targeting gold rings.
Discriminating, as I said the other day is tricky business. Your decisions have consequences and to optimize results, there are a lot of things to be considered.
As I've said before, it is not always about the economic value of finds. Sometimes you'll consider things other than the economic value of targets. You might value some things as valuable even if they have little economic value. You might be more interested in the historical value, for example. You have to decide what is valuable to you, but it will help you optimize your decision making if you can quantify the value of various items. It can be subjective or include weighting for subjective factors.
Having some idea of what is likely in the ground is important. If you have not idea of what is in the ground start with no discrimination until you get a small sample. You might find that the predominant junk items are pull tabs or bottle caps, for example, but if you find the ground is pretty clean, there is no reason to discriminate at all. It could be that you would be discriminating more good items than junk. Or if there are very few or new pull tabs, you don't need to discriminate out that range of items, which might also eliminate some gold rings or other good items. You can make better decisions when you have a better estimate of what is in the ground. Sampling is one of the best ways to improve your understanding of that.
On a recent hunt, I was using an Equinox metal detector in an area that has been hunted a good bit before but still has a lot of junk of various kinds. One of the most common items was pull tabs that were usually showing a 15, or was it 13, conductivity number on the Equinox. I knew I've found gold rings with that same number before, so I had to consider the trade-off. Was it worth possibly losing gold rings or maybe nickels or things like that to avoid the many pull tabs? Several factors were taken into account. The probability of a pull tab was high while the probability of a gold ring in that conductivity range in that environment was pretty low, but there was some chance. What were my goals for the hunt and what were the constraints? In that case, I was severely limited on time and the digging was tough. Recovery was difficult and time consuming. Without mentioning all the factors that went into the decision, I decided to skip those pull tabs on this hunt. It was a fairly narrow range (15 rather reliably on the Equinox) so with the high probability of it being pull tab and low probability of it being gold, in this case I decided to skip most of those readings.
One thing about aluminum pull tabs is that in many environments, including the hunt I was just talking about, they are typically on or near the surface, meaning that the cost of checking them is minimal. I could often swipe off a layer of leaves with my foot and 80 or 90% of the time see the pull tab without digging. They seldom go deep, so the depth indicator can be another useful piece of information when you might for some reason be wondering. On items like that, occasionally verifying your interpretation can help your confidence level and decision making.
By confidence level I mean more than emotional reaction. I'm thinking more of the experimental scientific meaning having to do with the probability of error.
I used the old and somewhat limited illustration above to introduce a few important points.
Detectorists often judge results in terms of number of good finds or the objective or subjective evaluation of finds. They almost never take into account what was missed that could have been found with different strategies, actions or decisions. I've always hated the idea of missing good things even though that is seldom known. I've often dug junk in order to minimize misses, but digging junk has multiple benefits, one of which I already mentioned. That is one way to assess what is in the ground - the relative likelihood of junk compared to good items. Digging junk also eliminates target masking. It prepares the ground for future hunts. As long as junk remains you might can over the same junk every time you visit the site. You can encounter the same junk items multiple times but only have to remove them it once. Some people even dig junk items, throw them back on the ground and dig them again on another hunt. There are times when I'd just rather remove the junk, but there may be times when you decide to leave them. It should be a considered decision that considers various factors.
Here is a reality that can have a big effect on decision making.
Valuable finds are rare. And very valuable finds are very rare and infrequently encountered. It is an exponential curve.
The above illustration shows the general relationship between find values and rarity. It is obvious enough that pennies, dimes and quarters are found pretty frequently while gold rings are found less often than common change, and gold coins are found even less often. Rarity is one important determinant of values. Common items, almost by definition, are less valuable. There are many types of common finds and the values increase exponentially with increasing rarity.
Although valuable finds are rare, their value is many times that of common finds. Would you rather find one gold ring or a couple hundred pennies? Even thousands of common coins or one gold ring or gold coin? But decisions that minimize the probability of a gold ring find are often made in favor of finding common and less valuable items.
Unfortunately, metal detectors generally do a very good job of identifying common coins but not so much rarer and more valuable finds. Detectors often have often had displays clearly marked penny, dime and quarter, for example, but how often have you seen an indicator on a metal detector display that says something like diamond ring, Rolex watch or one escudo.
The flat surface round shape of coins makes it easier for a metal detector to identify. Detectors are less reliable on items have irregular surfaces. I've demonstrated in the past how difficult it is to identify a buried watch, for example. Watches vary widely on composition and can lay on the ground in different ways and all that affects the variability in the detector signal and readout. All readouts have some amount of variability. There is a confidence level and an error rate. Some items, such as coins, produce a tight range around a central or average number. Other items have a larger range of variability.
On the hunt I was talking about above, the pull tabs produced a 15 with a high degree of reliability, but occasionally a pull-tab would produce maybe a number or two higher or lower. The range of numbers was pretty tight even though there were exceptions.. The error rate for an item is an important factor. The problem is, as I mentioned above, it is very often more difficult to positively identify valuable items than common items.
This is a good time to bring up the bell curve.
As you've probably notice conductivity numbers for specific items will not always produce the same number. There will be a central number and a range of numbers higher and lower than the what the detector centers on. On some types of items the range of numbers will be smaller than others. For example a coin lying flat will produce a fairly consistent conductivity number and a fairly small range of numbers a few points higher or lower as you continue to sweep over it. Irregular shaped or complex items, as I've already discussed, will show a wider range of numbers depending upon a variety of things, including the part of the object the coil passes over on various sweeps.
Variations in the item will also produce on occasion different center numbers. For example, the pull tabs that usually produced a 15 on a few occasions produced slightly different numbers. Perhaps it was because of the pull tab being positioned on the ground at an agle rather than lying flat or the pull tab being bent. Below are some bell curves. Some are tighter than others.
 |
| Variety of Sample Bell Curves. |
One of the better examples of the bell curve, although it is not related directly to metal detecting is IQ. If you take a test once, your IQ might be slight different than when you took the IQ test maybe some years ago. According to theory, IQ should not change much with age, however there is such a thing as test or measurement error. In many, if not most, practical situations, there is some degree of measurement error. Sometimes the test measures real things it is not supposed to be influenced by. For example the test taker might be very tired and drowsy when he takes the test on a particular day. Or the scorer makes an error.
If you give a large representative sample of the population, the average IQ of the population should be 100, but there is variance. Some people will score very high while others will score very low. Very high or very low scores will be less common than scores around the average IQ of a hundred. The farther the score is away from the average of 100, the less common or frequent it is. For example, few people have an IQ of 150, but more people will have an IQ of 120, which will be less frequent than an IQ of 100. Extreme deviations from the norm tend to be less common. It is unusual to have people that are eight over seven feet tall or under four feet tall. That is the kind of thing the bell curve describes. The largest deviations from the normal are also the less common. That is kind of like valuables, except that distribution is skewed in one direction. With the increase of value is a decrease in frequency, but the chart doesn't display negative values. The left side of the graph starts at zero.
In metal detecting target ID there can be a range of readings, some of which are due to variables, such as coin position or perhaps some other overlapping item. But the important thing is to realize that the numbers won't be the same every time for a specific item no matter what the reason and there will be a range of numbers. Some items, as I've described will have a small variance, while others produce more widely varying numbers.
The blue curve above represents the numbers of an item that produces a night tight distribution of conductivity numbers. That would be like coins, which are fairly easy to identify. The red curve represents an item that produces a wider range of numbers, such as a watch might do.
The important thing to recognize is that there is always some degree variation or error. Sometimes the variation is significant and consequential and other times less so.
I'm going to stop there today and continue with this discussion some other time.
---
 |
| Fort Pierce South Jetty Sunday Afternoon. |
The surf is still pretty flat, but a higher surf is expect Tuesday.
 |
Surf Chart for the Fort Pierce Jetty Area From Surfguru.com. |
Good hunting,
TreasureGuide@comcast.net