Top of Site > Clamps as Things > Prices for Clamps > Top of Section :
Here are the key findings, with links to details:
Since 1997, I have tried to understand how dealers set asking prices. I like to do a pricing analysis based on my experiences at Brimfield, on Saturday. It gives me a chance to see lots of clamps, all at one time, and under fairly comparable conditions. However, I have to recognize that some are general dealers selling users, some are selling lookers, some are tool dealers trying to sell to collectors, so there are several "markets" going on all at once.
Of course, some prices are more likely than others. No dealer asks $ 12.95, for example. Often prices (especially higher prices) are multiples of 5 or 10 dollars. Also, these analyses or studies are for "asking" price, which may be considerably different than selling price, after negotiations.
My basic assumption is that the price of a clamp can be estimated from the combined effect of two functions.
Price = F( Size_Function, Condition_Function )This assumption is based on the knowledge that very few dealers recognize that some clamps are scarcer than others. (Some dealers don't even know the difference between factory made and craft made.) Condition here refers only to working or not, clean or not. Married pieces may be less valued by a collector, but most dealers will ignore this aspect of condition.
I usually assume a Linear Model for the Size_Function, that it is a constant plus so much per inch of jaw. This can explain about half of the variation in the data
Sometimes there is enough data to explore an alternative Fitted Model. In a Fitted Model, the assumption is that some sizes are preferred, and priced higher than others.
I constrain the Condition_Function to be monotonic, that is, better condition must not imply lower prices. Usually only a four point scale is used: poor, fair, good, and very good. Then the value of good is set to 1, and the others are adjusted.
Usually, I assume that the two functions can be multiplied. Once I tried the assumption that they can be added.
In earlier studies, I just tried to read values off a plot. More recently, and especially when there are lots of data, I've used Excel Solver to get more precise fits.
In the summary above, "condition irrelevant" means that no consistent pattern could be discovered. This is usually because so many clamps were of similar condition. It is also caused by a few poor clamps having high prices, or some very good clamps having low prices. For any given size, the range in price is often 2 to 1, but it isn't consistently tied to condition! I suspect the variation is an expression of dealer ignorance of (or dis-agreement on) value.
For the statisticians in the crowd: If you know nothing about a clamp, then your best estimate of its price is the average price. If you know the size of a clamp, then a better estimate of its price is the linear model.
You can measure the fit of a model by what results from computing the root of the sum of the squares of the errors made in estimation. ("smaller" or "tighter" is better.) The fit from the linear model is about half the fit from using the average. (It varies from year to year, but half is common.)
No other information changes the fit by more than a few percent.
There are several related pages, that discuss clamps as things:
last revised and validated
Copyright © 1996- Wooden Clamp Journal