The typical HL Limited Edition piece is on the market for 759 days when it is sold out to dealers (retired).

Here are the pieces that have exceeded that average; sorted by the number of days they've been available to dealers since released by HL:

1437 Selkirk (4 years on 10-1-99)
1406 Brant Point (4 years on 11-1-99)
1314 Cape Meares
1285 Mukilteo
1254 Charlotte-Genesee
1254 Point Pinos
1254 Saugerties
1254 Matinicus
1254 Ida Lewis Rock
1163 Sandy Point Shoals
1163 Peggy's Cove
1163 Toledo Harbor
1132 Concord
828 Navesink
798 Long Ships
798 Hook Head

Next on the list, but inside the averages is:
736 Presque Isle

Rgds,
__
/im

[This message has been edited by JTimothyA (edited 09-06-99).]

Tim, is this 759 days a simple arithmetic mean, (sum of x)/n ?

If so, then what is the significance? Funny thing about any group of numbers...

(1) They will all have an average.

(2) About half of the individuals in the group will be above average, about half below average (depending upon the definition of 'average' and the shape of the underlying distribution).

Even in the group you used to calculate the average from, probably about half were above average, yet these are now retired, too, I am guessing.

I guess that being above average here only matters if they still have not retired? If so, then anxiety should eventually set in for about half of the models released, for one should expect about half of the models to be 'above average'.

A more meaningful number: average +3 standard deviations. Those pieces exceeding that number differ significantly from 'the pack'.

From Lake Woebegon, where the women are all pretty, the men are all strong and all the children are above average,

-Walter Shewhart

The 759 number is based on totaling the days on the market for the LE models, Christmas, Event pieces, etc. issued from 1991 through 1998 then dividing by the number of these pieces to get the average days on the market. (For each piece, its Retirement date is subtracted from its Available date yielding the number of days the piece was on the market.) I don't believe any pieces issued in 1999 are retired except for Hillsboro and the Canal set - so a slight adjustment could be made for these.

Out of that group of models, 53 are not yet retired. 16 of those have not retired after 759 days, about 30%.

>>what is the significance? <<

I suppose the significance is there are 16 LEs that have been on the market longer than then the typical LE is on the market. 13 of these have not retired after over 3 years of availability. Those pieces issued in 1995 are gettin kinda long in the tooth. Though they're not close to Cape Blanco which was available for over 5-1/2 years.

>>one should expect about half of the models to be 'above average'.<<

Expectation is, as they say, the mother of dissapointment.

>> Lake Wobegone <<
Groundsquirrels, sigh ...

Rgds,
__
/im

In Chart 1 below you can see all of the retired models (not current -- missing Wind Pt., Round Island & Panama set data).

Chart 1
[img]http://albums.photopoint.com/j/View?u=42289&a=613183&p=11988961&Sequence=1[/img]

You may note that the models that *really* took a long time to retire were the earliest ones. We might attribute this to the newness of the line at the time, the relative unattractiveness of the models, the alignment of the planets or whatever. For the sake of the argument, let's throw them out.

This brings us to Chart 2. These models depicted here are from the 'HL heyday' -- Beginning with Ponce de Leon and ending with Tawas Point. We can quibble about the end point, but it will not change the character of the chart materially. I don't think anyone capable of reading Chart 1 (if you had all the detail data) will quibble over Ponce de Leon as being the start of a new era in terms of retirement rate.

Chart 2
[img]http://albums.photopoint.com/j/View?u=42289&a=613183&p=11988967&Sequence=1[/img]

Here we see an average age at retirement of 672 days with a +3 standard deviation limit of 1052 days. The numbers above and below the average have no significance. As I said before, in any group of data, about half will be above the average, and half below, by definition. From this elite group of HLs, anything up to 1052 days old should be considered normal.

Finally, let's look at Chart 3. This runs from Ponce de Leon through Hillsboro.

Chart 3
[img]http://albums.photopoint.com/j/View?u=42289&a=613183&p=11988976&Sequence=1[/img]

The average age at retirement for this group is about 510 days. The +3 standard deviation limit is around 990 days.

From these data, it seems to me that there is nothing unusual about a piece that runs over 750 days until retirement. In fact, it would take another 240 days beyond that to be worthy of being recognized as unusual.

Haven't we done this before?

------------------
-Art [I count 13 sabre-toothed puppies]
[This message has been edited by Art (edited 09-08-99).]

Those are mighty impressive graphs, Art - looks like alot of work went into making 'em. I've tried to study your message to see what to draw from it. Since your approach opts to toss out ('for the sake of argument') the models released prior to Ponce, and you're not including Round Island or Wind Point, that pretty much leaves 1 model in the group you'd consider 'unusual'. And gee thats not much fun at all. Just one.

Fwiw, the initial post wasn't attempting to futz with the adjectives to characterize the current state of non-retirees as unusual or normal. Only with a little prodding did I offer some notion of 'significance' which wasn't really much.

All the knobs and dials tell us then is there's only that one silly old piece that falls outside your characterization of The Norm. And that, if I may be so bold, is pretty darn unusual. In fact, its more than unusual, its close to spooky.

Fortunately, I don't believe in ghosts, :-) so I'll just stick with the simple (or simplistic) notion that the models listed in the first msg. have been on the market longer than the typical model is on the market.

Oh, yes, almost forgot: Chatham. Retired early because production was cut off by HL due to inaccuracies.

Rgds,
__
/im

[This message has been edited by JTimothyA (edited 09-08-99).]

Does all this mean that the Selkirk van will soon be on the road again?

Mombo

Rumor has the Selkirk van camped out behind a Dairy Queen in Mount Kisko.

Must be "end of summer nothing to do" time.
Interesting number-crunching.

Jim

The graphs didn't take much work at all, Tim. I simply stole the data from your famous downloadable spreadsheet (not the latest version, though) and plotted them. The +/- 3 standard deviation lines were a breeze to calculate, too. Spreadsheets are great.

If you consider the limits on either the second or third chart (1052 or 991 days, respectively), then fully 13 of the 16 models you listed in your first posting qualify as 'unusual' (hence my earlier reference to 13 sabre-toothed puppies). Is that a little more fun for ya?

You can see what I mean about not needing to quibble about the precise end point of the second chart... it matters not if we stop at Tawas or Hillsboro: 13 of the 16 models you mention are 'unusual' by either test.

I discarded the early models from the data set because they are clearly different from the rest of the data and their inclusion would skew the statistics toward an artificially high retirement age. I can't say what happened at the introduction of Ponce (other than Ponce itself -- reason enough, imo, for the retirement age to drop), but since Ponce, the HL world has clearly been a different place.

I agree that the majority of the models listed by you *are* on the shelf for an unusually long time. I just disagree with the method used to determine significance. "Above average" and "below average" really have no meaning, but we persist in this abuse of statistics in this innumerate society of ours. I feel obligated to combat such abuse at every opportunity. Nothing personal.

Now, what should we do with these 13 sabre-toothed puppies?

------------------
-Art

Here's an interesting statistic: Yaquina Bay LH (the real one, not the HL) retired after only 1,063 days in service.

Shouldn't retirement age be to scale?

Beware Wackos, see what happens when you get this carried away - you get promoted to Cruise Director, or Saint!

Mombo

Abuse? My dear Art, tut tut, I decline to engage in this, this pandemonium of adjectives whereby the likes of 'significance' and 'meaning' and 'skewing' are ciphered well or poorly out of a set of numbers. Review the statements in message the first - simple as they are, they turn not on a semmantic but are capable of truth value.

But in the interest of comaraderie I'll offer that serial ordination is non-naturally occuring - the ultimate ad hominem to statistics (rofl). As Will S. avered, '...the fault lies not in our stars, but in ourselves'. But then again, there's always Harry Seldon and his statistics of manifest destiny. (more rofl)

Thought for the day: methodology is to method as scientology is to science.

Rgds,
__
/im
[This message has been edited by JTimothyA (edited 09-09-99).]

(sigh)

OK Tim. Meet me in the FSB after midnight.

216.46.163.97/forums/Forum17/HTML/000052.html

WARNING: Not for the squeamish. No whiners, please.
[This message has been edited by Art (edited 09-10-99).]

You guys have too much time on your hands!

Art, Art, Art....What am I to do with you? I work by day doing performance reporting based upon statistical process controls, creating chart after chart till my mind goes numb. I turn to HL collecting and forums as an escape. Then I come here, and to what to my wandering eyes should appear, but a miniature HL collector, and his Shewhart Charts!!! AAARRGGHH!!! Is there no escape from it??? I have this major fear that in the base of Pharos, they will find the term SPC!

Now, on the practical side, Art is correct. If it isn't outside standard deviation +3, then all is in control.

Jeff,

Are you implying Art is a miniature?

LOL, Dave

Only in mind.

I really don't know, but it certainly flowed, didn't it?

(Apologies to Clemment Moore)

1. Did we ever get a firm retirement date on
Sandy Pt. Shoals and Grand Traverse?

2. Do we yet know final edition numbers on:
HL 504 Point Fermin Ornie
HL 506 Port Sanilac
HL 509 Sea Girt
HL 520 Port Sanilac Ornament

Try this page for Grand Traverse information:

http://216.46.163.98/news/news_stories/index.htm


Saint WackoPaul'
Keep the Flame

Thanks Paul. I had that date but wondered where I got it.

Save the Forum Threads! Save the Kilobytes!