Using an Algorithm

I am in the process of creating an algorithm that will help me determine true value. By true value I mean adjusted value. The criteria will include factors like composition (melt value), mintage, population, survival rates, historical value comparisons, etc. There will be several other factors I will not disclose so no one copies me exactly. What do you people think?

 

Log in or Sign up to hide this ad.

its possible..

 

It would be interesting to see how you weight each factor in your algorithm. Depending on the series, many of the silver coins were melted in the early 1900's under the Pittman Act. A second melt occurred during the 80's when the Hunt brothers essentially cornered the silver market. How to account for this in survival rates, when no one logged the dates and mint marks of the coins being melted. How will you account for toning? Toning is truly in the eye of the beholder and purely subjective, yet carries large premiums in many cases.

After you run your program, I'd be interested to see how those prices compare to the auction prices at Heritage and the other great auction houses. IMO, pricing can only be determined by transactional data.

We had a discussion about this on another coin site yesterday. There was a 1870 Half Dime MS66 Gold Bean coin that just sold for 1.5x the "printed" price. IMO, there were many factors that went into this premium price: condition rarity (15 graded MS66 and only 9 in higher grade), toning, gold bean, availability to purchase a coin in this grade. If a condition rare coin only comes up for sale or auction once every 3 years, how do you properly "price" it? BTW, there was an MS67 also sold in the same auction but for 1/3 the price of the MS66. The MS67 also had toning, but it did not add to the eye appeal of the coin.

In summary, I applaud your effort to use a scientific method to price coins, but you cannot properly calculate the "human factor" into the equation, so the output will likely be flawed.

 

I kinda disagree with pricing - in reality, all you need is two buyers to battle each other to raise prices.

At times, some coins may appear to be scarcer depending on the market condition and this could change over time. I also believe that there are too many variables - it's not just a matter of supply and demand. While we may be able to track historical pricing to factor in demand, supply can be a rather interesting area to tackle. For instance recently Bank of Canada has announced to release some of the hoards of gold coins that was stashed in their bank vault. Of course, this only tells you a small portion of a story - while you have total mintage of what the mint has struck, it doesn't reflect on how much coins were released to the public, how much were melted down, how high the current demand are and how much coins are stashed up by the bank, eventually impacting on price of coin.

On top of that, coin appeal just makes it too complicated. No one coin are exactly the same especially circulated coins.

Best of luck

 

Like it...don't forget about demand. TC

 

I don't know your background but you've just set yourself one heck of a task.
At a minimum acquiring valid data inputs will be a HUGE chore.

 

Value is an individual opinion and not a formula. Didn't they try this in Good Will Hunting?

 

And by adjusted value you mean...true value? What is your starting point? Face value? Melt value?

 

At one time I tried writing an algorithm that would interpolate the values of intermediate grades.
Basically taking CDN or CW values for a specific coin and calculating the values of grade not included in their charts such as MS-61 or MS-62.
Since the chart of the values from VG(-8) to MS-65 is not linear, strict averaging is not correct, i.e., the XF-45 value is not halfway between the XF-40 and AU-50 values.

Each coin would have its own equation.
In theory I would put in the data points (value at each grade), hit ENTER and out would pop the coefficients for the equation (in my mind some sort of polynomial).
I also wanted to force the equation to go through all the data points.
Then I could select an "in between" grade and interpolate its value.

I never did get it right.
I think was trying to use a math model and method that wasn't correct for the way I wanted it to operate.

Okay, did I glaze over everyone's eyes?
Sorry about that.

 

I like the idea (being a math major) but think you need to have some sort of adjustment factor based upon recent auction prices.

There's another side of me that says what you come up with with have limited "value" (pun intended) because many times the "value" of a specific coin has more to do with the buyer's mindset than with anything having to do with the type of coin.

Having said that, I think you make have a practical tool for those of us that are new to the collecting game or that are looking at coins outside of our main collection effort.

If you'd like any input when you get a "beta version," I'll be more than happy to assist.

Ken

 

I understood you perfectly!!!

 

If coin collectors were logical they would not "buy the highest grade they could afford;"they would buy the grade that gave them the most for the money. Just as a business logically seeks the most utilization for their fixed costs. But we're not logical, we're whimsical. And that's on a good day. On a bad day we're almost insane-AI55.

 

It's actually very easy to calculate a coin's value. Just use this simple equation...

 

I'd give this multiple likes if I could!!!!!!!!!

 

I got -1 for an answer. TC

 

See. Your coins are less than worthless. Whose fault is that?

 

42

 

You guys do know that 48.7% of all statistics are fabricated, right?

 

Oh no, not a paradox! Hello there, you have found one of three of "thecoinlover's" hidden messages! Find two more and win a cookie!

 

Nope - just a single dox