Type Set Weighting Idea - CEO Thread Variant

itsnotjustme

I had posted this on TypeTone's thread on CEO's weighting idea, but it drew no replies. That hurt my ego so I had to start a new thread...

I was thinking this would work well for type sets using the variation of all eligible coins in higher grades...

...but (sorry) then I pulled out my $20 Saint and realized some coins have such huge populations slabbed, they will drive the scores.

Example:
1976-S Silver Kennedy
Total graded in type 439
No Coin, score=439
MS67 pop 167, ~$50, score=36
MS68 pop 36, ~$450, score=0

1924 $20 Saint
Total Graded in type 364,030
No coin, score=364,030
MS63 pop 128,614, ~$375, score=92,934
MS64 pop 71,445, ~$475, score=21,489
MS65 pop 18,031, ~$850, score=3,458
MS66 pop 3,279, ~$1,600, score=179
MS67 pop 177, ~$6,000, score=2
MS68 pop 2, ~$XXX,000, score=0

So, if you don't have a MS66 Saint, none of the lower total pop coins in your set matter. If you change the scoring to the percetage of eligible coins that are higher, thing are brought back into balance. All no-coins would get you 100 points. Any top pop coins earns 0 points. That MS67 76-s Kennedy would bring you (36/439*100) 8.2 points. An MS63 Saint would earn you 25.5 points. An MS65 Saint would earn you just .94 points. I'm sure there are weaknesses in this scheme as well, but it appears to overcome the downfall of some coins in a set that are far more frequently graded.

keithdagen

Brian,

Interesting idea. That could work out pretty well. Except that you would need to combine the MS and PR together for their populations, since the set is mixed at this point.

Keith

CEO

Brian,

I think it's a great idea. The same issue exists for date/mintmark sets as well, but I think the problem would arise far more often in type sets. After all, if you have a really great Morgan dollar set, would you really have an MS-63 1881-S?

- Corey.

clevegreg

I think it would just be easier and just as effective to weight the slot. For example, weight Barber Halves at a 7 and the Lincoln Cent ( current dates) at a 1. You can easily go through the entire set applying relative rarity on the slot and then not have to worry about the coin population itself. Since it is a Type set, it shouldn't really matter which date/mintmark coin the collector has selected.

Greg

clevegreg

Here is an example of what I had in mind:

Sorry the formatting didn't carry over so great, but you get the general idea.
Item Weight

Indian Cent (1900-1909) 4
Lincoln Cent (1909 VDB on reverse) 3
Lincoln Cent (1909-1917 no VDB) 3
Lincoln Cent (1918-1958 VDB on obverse) 2
Lincoln Cent (1943 Steel) 2
Lincoln Cent (1944-1946 Shell case) 2
Lincoln Cent (1959-1982 Bronze Memorial) 1
Lincoln Cent (1983-present Zinc Memorial) 1
Liberty Head Five Cent (1900-1912) 3
Buffalo Five Cent - Type 1: Buffalo on Mound (1913) 3
Buffalo Five Cent - Type 2: Buffalo on Base (1913-1938) 2
Jefferson Five Cent - Type 1: Copper- Nickel (1938-1942, 1946-present) 1
Jefferson Five Cent - Type 2: Copper- Silver (1942-1945) 2
Barber Dime (1900 -1916) 4
Mercury Dime (1916-1945) 3
Roosevelt Dime (1946-1964 Silver) 2
Roosevelt Dime (1965-present Copper Clad) 1
Roosevelt Dime (1992-present Silver) 1
Barber Quarter (1900-1916) 5
Standing Liberty Quarter - Type 1 (1916-1917) 6
Standing Liberty Quarter - Type 2 (1917-1924) 4
Standing Liberty Quarter - Type 3 (1925-1930) 4
Washington Quarter (1932-1964 Silver) 2
Washington Quarter (1965-1998 Copper Clad) 1
Washington Quarter (1976 Bicentennial Silver Clad) 1
Washington Quarter (1976 Bicentennial Copper Clad) 1
Washington Quarter (1992-1998 Silver) 1
Washington Statehood Quarter (1999-present) 1
Barber Half Dollar (1900-1915) 5
Walking Liberty Half Dollar (1916-1947) 2
Franklin Half Dollar (1948-1963) 2
Kennedy Half Dollar (1964 Silver) 1
Kennedy Half Dollar (1965-1970 Silver Clad) 1
Kennedy Half Dollar (1971-present Copper Clad) 1
Kennedy Half Dollar (1976 Bicentennial Silver Clad) 1
Kennedy Half Dollar (1976 Bicentennial Copper Clad) 1
Kennedy Half Dollar (1992- Present Silver) 1
Morgan Dollar (1900 -1921) 2
Peace Dollar (1921-1922 High Relief) 3
Peace Dollar (1922-1935) 2
Eisenhower Dollar (1971-1978 Copper Clad) 1
Eisenhower Dollar (1971-1974 Silver Clad) 1
Eisenhower Dollar (1976 Bicentennial Silver) 1
Eisenhower Dollar Ty 1 (1976 Bicentennial Copper Clad) 1
Eisenhower Dollar Ty 2 (1976 Bicentennial Copper Clad) 1
Susan B. Anthony Dollar (1979-1999) 1
Sacagawea Dollar (2000-present) 1
Liberty Head Quarter Eagle ($2 1/2) (1900 -1907) 5
Indian Head Quarter Eagle ($2 1/2) (1908-1929) 5
Liberty Head Half Eagle ($5) - Type 2: With Motto (1900 -1907) 5
Indian Head Half Eagle ($5) (1908-1929) 5
Liberty Head Eagle ($10) - Type 3: With Motto (1900 -1907) 5
Indian Head Eagle ($10) - Type 3: No Motto (1907-1908) 7
Indian Head Eagle ($10) - Type 3: With Motto (1908-1932) 5

keithdagen

Greg,

I have a feeling that a variant of your idea will be adopted. I would tweak the weights somewhat, but the general feel is definitely there. Very few coins after 1964 deserve a weight greater than 1, Barbers and gold weighing heavily, with the $5 Indian and the No Motto $10 Indian getting the highest weights.

Keith

CEO

Ah, but then we're back to the same old problem that we were trying to address. For example, a Barber half in MS-67 is easier to come by than a 1921 Peace dollar in MS-67, but the situation is just the opposite at lower grades.

Weights by themselves won't do the job unless they are weights per grade. But, weights per grade would be an accounting nightmare, spawning a brand new crop of Excel spreadsheets necessary to calculate one's ranking!

keithdagen

CEO,

I already offered my services as an accountant to anyone who is confused on how the weighting is affecting their sets. Fees payable in pre-1900 slabbed gold.

Keith

Typetone

The beauty of the CEO system, is we don't need an expert to weight or pontificate on anything. Sorry Keith, I hope I'm not costing you a job. I do however like the % variant proposed by Brian. I realized I make most of my decisions on grade to acquire by targeting a % of finest, rather than a number away from finest. I wasn't sure the CEO system would work for type, but I think the % variant would work very well. For series sets, I don't think it matters, although % seems cleaner mathimatically.

Has anyone shot the idea to DH yet?

Greg

cosmicdebris

Greg one fix:

Eisenhower Dollar Ty 1 (1976 Bicentennial Copper Clad) 2
Eisenhower Dollar Ty 2 (1976 Bicentennial Copper Clad) 2

keithdagen

Typetone,

The only problem is that if we keep hammering DH with new ideas on how to weight (no matter how good they are) we will never get that Basic Type Set added.

With Brian's take of CEO's system, there are still percentages and heavy math involved. I figure I can squeeze something out of it, so it wouldn't cut my business too badly.

Keith

clevegreg

Bill,

You are probably correct on the Ikes. They are a one year type and relatively scarce as such. I just hope to keep this simple yet meaningful for the type sets.

Greg

keithdagen

OK,

It was a slow afternoon, so I gave weighting a scientific method under the current method of doing things.

Assumptions:

1. MS or Proof coins allowed.
2. Price is based on second highest graded coin (this was how they weighted PSA cards)
3. Prices are out of PCGS guide, regardless of whether a higher coin with a known price regularly exists (1943 Lincoln MS-68 anyone?) Also regardless of whether a higher priced coin really exists (Sac in PR-70 anyone?)
4. Grades were dumbed down for Barber Half, Indian $2.5, Indian $5, Indian $10 No Motto, and Indian $10 Motto). If you didn't do this, mathematically, the Indian $5 would get 10 points and all other coins would get one.
5. Formatting ain't good on this board and you can't post Excel spreadsheets direct.
6. This was just goofing off, so don't take the fact that your favorite series didn't get squat for points seriously.

Keith

Slot Type Price Weight Notes
1 Indian Cent (1900-1909) $850 3 MS-66
2 Lincoln Cent VDB (1909) $195 1 MS-66
3 Lincoln Cent (1909-1917) $155 1 MS-66
4 Lincoln Cent (1918-1958) $20 1 MS-67
5 Lincoln Cent (1943) $30 1 MS-66
6 Lincoln Cent (1944-1946) $17 1 MS-66
7 Lincoln Cent (1959-1982) $33 1 PR-69DC
8 Lincoln Cent (1983-Curr) $14 1 PR-69DC
9 Liberty Nickel (1900-1912) $800 2 PR-66
10 Buffalo Nickel (1913) $235 1 MS-66
11 Buffalo Nickel (1913-1938) $57 1 MS-66
12 Jefferson Nickel (1938-Curr) $15 1 PR-69DC
13 Jefferson Nickel (1942-1945) $200 1 MS-67FS
14 Barber Dime (1900-1916) $1,175 4 MS-66
15 Mercury Dime (1916-1945) $52 1 MS-66FB
16 Roosevelt Dime (1946-1964) $108 1 PR-68DC
17 Roosevelt Dime (1965-Curr) $12 1 PR-69DC
18 Roosevelt Dime (1992-Curr) $33 1 PR-69DC
19 Barber Quarter (1900-1916) $2,000 6 MS-66
20 Standing Liberty Quarter (1916-1917) $2,000 6 MS-66FH
21 Standing Liberty Quarter (1917-1924) $2,350 7 MS-66FH
22 Standing Liberty Quarter (1925-1930) $1,650 5 MS-66FH
23 Washington Quarter (1932-1964) $132 1 PR-68DC
24 Washington Quarter (1965-1998) $27 1 PR-69DC
25 Washington Quarter (1976 Silver) $33 1 PR-68DC
26 Washington Quarter (1976 Clad) $33 1 PR-68DC
27 Washington Quarter (1992-1998) $35 1 PR-69DC
28 State Quarter (1999-Curr) $20 1 PR-69DC
29 Barber Half (1900-1915) $2,850 9 MS-65**
30 Walking Liberty Half (1916-1947) $215 1 MS-66
31 Franklin Half (1948-1963) $750 2 PR-68DC
32 Kennedy Half (1964) $595 2 PR-68DC
33 Kennedy Half (1965-1970) $375 1 PR-69DC
34 Kennedy Half (1971-Curr) $33 1 PR-69DC
35 Kennedy Half (1976 Silver) $25 1 PR-68DC
36 Kennedy Half (1976 Clad) $40 1 PR-68DC
37 Kennedy Half (1992-Curr) $39 1 PR-69DC
38 Morgan Dollar (1900-1921) $360 1 MS-66
39 Peace Dollar (1921) $2,500 7 MS-65
40 Peace Dollar (1922-1935) $625 2 MS-66
41 Ike Dollar (1971-1978) $20 1 PR-68DC
42 Ike Dollar (1971-1974) $80 1 PR-69DC
43 Ike Dollar (1976 Silver) $50 1 PR-68DC
44 Ike Dollar (1976 Clad T1) $45 1 PR-68DC
45 Ike Dollar (1976 Clad T2) $40 1 PR-68DC
46 SBA Dollar (1979-1999) $40 1 PR-69DC
47 Sacagawea Dollar (2000-Curr) $49 1 PR-69DC
48 Liberty $2.50 (1900-1907) $2,200 7 MS-66
49 Indian $2.50 (1908-1929) $3,350 10 MS-65**
50 Liberty $5 (1900-1908) $2,850 9 MS-65
51 Indian $5 (1908-1929) $2,750 8 MS-64**
52 Liberty $10 (1900-1907) $3,100 9 MS-65
53 Indian $10 (1907-1908) $3,000 9 MS-64**
54 Indian $10 (1908-1933) $3,150 9 MS-65**
55 Liberty $20 (1900-1907) $2,600 8 MS-65
56 Saint $20 (1907-1908) $2,150 6 MS-66
57 Saint $20 (1908-1933) $2,400 7 MS-66

tradedollarnut

My conversations with DH lead me to believe that they are dead set against a weighting system that varies the weight by grade or value. They intend to place the weights at an appropriate level for mid range B.U. coins and leave it there - even tho this creates situations like I illustrated on my thread ($90,000 MS68 1875-S trade dollar adds 0.78 and $1500 AU55 1873CC trade dollar adds 6.67 to the set rating). PCGS prefers the simplicity and expects it to work for the majority of the sets in the registry.

If you think about it, it actually "dumbs down" the registry so that the middle range keys really drive the set ratings. This allows the 20th (or 40th, etc) place set to appear much closer to the 1st place set. Instead of having a huge disparity, you can get close just by concentrating on the key coins and ignore that beautiful supergem common date because it really doesn't help you much. In the long run, this will drive the prices of the keys upward and the condition rarities downward (that is if the set registry can actually influence the values of coins).

CEO

I could understand DH's decision if the system we're proposing was more difficult than present weight-based scheme, but it's actually simpler and more elegant. With their present system, experts have to work to develop the weights, and must continue to do so after new coins are added every year. With the population system I was proposing with Brian's modification for type sets, they already have the data and the data would continually evolve as more coins are certified.

I think that having weights is better than having none at all, but it's really only a small step toward a more accurate ranking. Why would they not desire to use their own populationd data? They already display it for all the sets!

- Corey.

tradedollarnut

My opinion is migrating toward the extreme: that an inadequate weighting system is worse than no weighting at all. I can live with the imperfect concept that a rare MS68 common date is exactly equal in weight to a rare MS68 key date more readily than the current weighting system. The idea of a $90,000 superb gem coin adding 0.78 to the set rating when a $1500 circulated key date coin adds 6.67 just sticks in my craw!