dwscrimshaw@bti

While trying to date my Holdsworth i have compiled the available data on the web. Basic info below with the rest on my Google drive if anyone's interested. Hope it helps.

https://drive.google.com/drive/folde...UZYSTNKZkN4TmM

allend

Dave
Fantastic work for all to see. Will be a great resource for folks. I've also brought it to more attention by posting a link to this thread on https://www.bikeforums.net/classic-vi...s-here-11.html
Doug

unworthy1

That's truly a prayer answered, many thanks for this compilation!
I'm also happy to see that my (now ID'd as 1975) Super Mistral 001240 is listed
as well as my "shop-built" Touring model (1969) 69226

dwscrimshaw@bti

Thanks for the comments, glad it helps.

Just one thing the 6 Digit numbers, of which yours is an early example, started in 1976 I believe which would make your one from that year. That still fits with the mid 70's description I found of your bike. I report the actual details as found on page one of the spread sheet and my interpretation of it on the second page, others may hold different views.

As I did not cover this number range initially I have now added it and the graph attached.

Can i just confirm, are both your bikes Super Mistrals's with one a touring version?

Link to updated files
https://drive.google.com/drive/folde...UZYSTNKZkN4TmM

unworthy1

Dave, My black-with-red '76 is definitely a Super Mistral and has the decal on the top tube that says so, still in original paint as far as I can tell, it is NOT a fastback seatstay model, but has wrap-over stay tops.
The "shop" frame was never identified by model except it's clearly intended for touring and came into my hands with early Blackburn racks, front and rear. It has Campagnolo vertical dropouts and cantilever brake bosses F&R, plus a lower-than-road BB height (I measured it once but now long forgotten). Original paint was what I call "pea soup green" with red decals and the seat cluster features the "chunkier" style wrap-over that "shop-builts" have.
All this reminds me I really need to take new pics of both bikes cause the old ones have gone into the ether of digital limbo...

jetboy 

cool! some of the data points may be .. what do you officially call it I am not sure.. but for example my #037416 was used as a 1982 point but I only dated it as a 1982 from extrapolation from other data points in the same set, but then it has been put back into the set as a point in a recursive way...

still, a good best guess!

allend

Hi Dave,

I think the line for the six digit numbering system should not be linear but more a sigmoid shape. This would reflect the gradual ramping up of production at new factory between 76ish to 85. As you acquire more accurate data points likely this will be apparent.

We know that production of Holdsworth frames in 1972 was at approx 1500 as stated by Kilgariff. This level of production would reconcile with the older numbering system progression continuing on at same rate to 1976. From the 16 Claud Butler numbers (different sequence) I have collected from 65 to '75 ish it appears there were approx 2000 Claud frames built per year. Thus a total of 3500 Holdsworth and Claud frames per year, not including MacLeans and Grubbs which were ceased around time of move in 76. Following move in 76, according to Kilgariff, all numbers for Holdsworth and "Claud" frames were lumped together. He also says total production reached 7000 by 1980 and for what it's worth I chatted on line with a former Holdsworth painter who states total production was at 200 per week by mid 80's.

So.. Collecting and using Claud's in this graph would increase firmish data points. FYI I have about a dozen Claud numbers in the 6 digit period. Lowest number I've seen is a probable "mid 70's" Claud Olympic Road number 000017. See:
https://forum.cyclinguk.org/viewtopic.php?t=24130

Hope this helps.

Doug

dwscrimshaw@bti

I have done bit more work on this. Added a few more CB numbers and set the date range inline with the known production volumes. I have left all but the very early frames and two of the 84 frames as is. I need to go over the info again to adjust the obvious ones that are out of place.

dwscrimshaw@bti

You might be interested in this thread from another forum discussing frame numbers and the style of seat clusters. think this was on specials though.

Login | Retrobike

dwscrimshaw@bti

I have re-done the graph working on stated production levels in the factory and then added a year by year ramp up between the known points. This gives a steadily rising production rate in contrast to how I estimated it before by splitting the difference between highest and lowest between the years. On this new graph, above, yours sits right in the middle of 1982 production numbers. I guess 82 is a good date for this frame, what do you think?

dwscrimshaw@bti

Yes good way of proceeding, as you can see I have done this above and added the chart. I have also added some more CB numbers, but not gone looking for any more yet. do you have any that could be added?

dwscrimshaw@bti

More up dates on the 76-85 numbers, files also on cloud.

https://drive.google.com/drive/folde...UZYSTNKZkN4TmM

dwscrimshaw@bti

Another update taking into account know number progression. hope it helps
https://drive.google.com/drive/folde...UZYSTNKZkN4TmM

realsteel

Thanks for posting all your meticulous research! I did something similar for the 1973 Holdsworth build ( https://www.bikeforums.net/18723467-post4.html ) but nowhere near as extensive.

Have you tried entering the year numbers as 19xx.5 in your plots? It may give a more accurate fit than using 19xx.0.

Also, there is a noticeable decrease in the rate of frame production between 1960 and 1967. Any idea why? Maybe problems at the company or an economic slow-down.

dwscrimshaw@bti

Thanks, no I haven't tried the 19xx.5 approach, may give that some thought.

The apparent drop in 1960 could be due to my juggling of the numbers to get close to 26400 in 1960 and not evening things out before this enough. Might have another look although there are some big jumps at the year boundaries so I don't think it would change much. More numbers would help confirm the dating.

The 60's numbers are tricky and not so many of them, it is possible to make this progression linear similar to the pre-60's, but I am not happy with the amount of date juggling involve. After the return to 5 digits in 1965 the production rate does look very low and we seem to have a lot in 1975 pushing the rate right up. I tend to think this is due lack of info as I could make it look linear. The shop used their own numbers incorporating a 2 digit year code, this means that the lack shop builds would drop the production rate a bit. Also when the 6 digit system started the production steadily climbed. I have assumed that prior to this they also tried to ramp up production out of the old facility. Not sure this is totally satisfactory so i hope I can come across some more numbers

Thanks, I'll keep working on it

allend

This apparent drop in production in early 60's is largely accounted for by the fact that between ~ 62 to 64 the Holdsworth factory for some reason gave up on their usual sequence, converting to a 4 digit system for this period and by 65 reverted back again to former sequence. This why the sometime confusion re 4 digit numbers on early 60s looking bikes. Also shop numbers because allotted by factory in groups often lagged behind factory numbers as built in much smaller numbers. In 1965 of course is when shop started own system beginning with 2 digit year of build. This is all info gleaned from Kilgariff.

https://retroride.blogspot.ca/2007/05/holdsworth-italia_1672.html

dwscrimshaw@bti

I have tried some Box plots on the numbers to give an idea of range in each year. The first is the raw info ordered and obvious date problems corrected, the second is the info as I have it in the Excel file. I guess it is a matter of opinion how much modification of the dates you prefer. What is clear is the distinct grouping of numbers for each year. I think I would expect more the tales of each year to be closer on the basis that we are taking a fairly random sample throughout a roughly steady production period.

I am still of the view that working on the known yearly volume of ~1500 frames and trying to hit the know fixed points is as good as it is going to get. Also need to bear in mind that there should be roughly 40K+ numbers and I am only working with a handful of them. This is probable the only explanation for the fluctuation caused by the higher number of frames found in one year compared to the rest.

I would welcome any more thoughts form anyone on how to present this.

dwscrimshaw@bti

Just plotted a constant production rate for 65 to 76 using the highest and lowest know numbers and dividing by number of years. You can see the only frame numbers that fit are the first and last, where as pre-65 the known numbers largely fall naturally into a constant yearly rate of ~1500/yr and also hit the known fixed points.

Clearly something was going on in this period, it is possible production ramped up prior to the move to the new factory, but not sure why it was so low in early 60's. that said I only have one or two numbers for each year at the start of this. Very difficult to draw any conclusions.

allend

Hi Dave,

I am not sure if you did already, but I think it would be helpful to exclude shop builds from the graphs. As stated in my previous email, until '65, the shop used factory numbers that were allotted in batches and thus shop dates and numbers drifted out of sync with factory builds. Therefore not valid data points.

Also the sequence Holdsworth factory was using from 62 or 63 until 64 inclusive was a different system so I think we need to either use a flat "average" line for these years or take these years off the horizontal axis altogether - as if no production.

This is assuming we take Kilgariff info as gospel. What do you think?

Great work! You must have a lot of patience.

Doug

dwscrimshaw@bti

Yes, not sure if I know enough to spot the pre-60's shop built frames. All I would say is that this tracks factory 'issued' numbers. I understand that a batch was issued early 60's 28,000 to 28,300 which explains why we have some old 5 digit numbers in 62-63. I think some similar may be happening in early 50's as I have a number of dates where I have had to move them by a year. Could these be shop frames?, a lot of these numbers for 1952/53 came from Classic Lightweights so I would expect them to be good dates. Not sure what you think of these numbers?

I have left out the 4 digit numbers from these charts and understand post 65 shop frames started with two digits to signify the year, these too are not included. The numbers for 65 on wards are a bit difficult as I don't have many good dates. I do have a couple of good dates at each end of this period, one in 1974 and one dated by VCC that is close. These dates support a linear progression rather the the rising one above, this progression is lower than pre-60's due to lack of shop numbers. The rest for this period are a bit vague and therefore support a linear progression.

Is it plausible that frame production was so stable over so many years? I guess these charts should be viewed as a predictor based on constant production meeting some fixed points with accuracy being only a year or so either way.

I have now changed these charts to be based on known production volumes and fitted them around dates with good pedigree and tried to hit the fixed points for 58, 60 & 61. Previously i just grouped them by years and then split the difference between the highest and lowest across the year boundary. i am also trying to check out the known good dates to make sure I have not shifted them without good reason.

The excel sheet is on the cloud, I have not updated the graphs yet. Have a look and see what you think.
https://drive.google.com/drive/folde...UZYSTNKZkN4TmM

rhm

What do the mid-1970's serial numbers look like? I ask because the number stamped on the BB of my Holdsworth 531 Special was so ineptly stamped that I had hesitated to consider it a serial number at all. No photo, I'm sorry, there was too much glare; but here's a drawing:



I would interpret that as 002808, so 1976, which is what I'd have expected. The 531 Special appears only in 1975 and 1976 catalogs, according to Kilgariff's site.

edit: The fork steerer is stamped much more neatly, 002803. I guess that works for the BB number as well....

Wileyone 

The Shop built 69226 did it have a K at the end of the Serial #?

unworthy1

Nope, no letters at all just those numerals and "stacked" in a column, although that column runs "horizontal" on the long axis of the BB shell.
There IS a small round drain hole (approx. 3/16") in the bottom of the shell which I hadn't noticed even though I painted his frame myself: I just looked at the bottom bracket again to make sure I hadn't missed anything.

dwscrimshaw@bti

Hi again,

Yes I have seen your graph and plotted this over mine for a comparison. To do this I read the numbers off the graph paper in the traditional way so the numbers would be close, but not exact. I also think you have sent me a list in the past and I have also put these alongside the numbers from your graph for a comparison.

You can see all this on the third tab if you are interested.

dwscrimshaw@bti

There is a couple of frame numbers I came across with a K at the end, one of them stated that it was stock build.

See at the bottom of this link True Adventures of the Retro Grupetto: Holdsworth Italia
And this one too Help with Holdsworth "Italia" | Retrobike

I will add the K to your number in my list.