Analysis of Aero Board Racer tag sizes

[poorman]

Hello everyone. For your consideration: my submission to the New England Journal of Leather Jackets.

Title: Analysis of Aero Board Racer Tag Sizes

Background: I'm new to buying leather jackets and I'm considering buying an Aero board racer. I noticed that there seemed to be a fair amount of variability in measurements for each tag size when shopping for jackets.

Objective: To determine the variability in board racer measurements for each tag size, so that I can figure out if any particular individual jacket is "a large 38" or "a small 40".

Methods: I scraped measurement data from the Thurston Bros website for each board racer on sale / previously sold with relevant measurements listed. Then I used R to spit out some statistics and plots.

Results:
Thurston Bros has 49 different board racer jackets with measurements listed. I excluded one jacket from the analysis because it was listed as "short".
Box plots:

Descriptive statistic table

Cross table: each column corresponds to the tag size, and the rows correspond to the type of measurements. "#" is the number of jackets for each tag size, "min/max" is the biggest and smallest measurement, "std dev" is the standard deviation.


Discussion: Back length and sleeve length had quite a bit of variability with respect to their tag size, as might be expected since these are the customizable measurements via Thurston Bros custom ordering. There's a decent amount of overlap between tag sizes for many measurements.

Keep in mind that this is just a small and limited sample. There weren't many jackets available to compare. I didn't calculate confidence intervals and all that to estimate the means of all the Aero board racers that exist. But it can still be useful as a reference point. For instance, we can look at a Real Mccoy J100 tag size 40. Measurements listed on their website:
Chest: 21.3in
Sleeve: 25in
Length: 25.4in
Shoulder: 18.9in

We can compare it to our Aero BR 40 mean (standard deviation in parentheses):
Chest: 21.70 (0.23)
Sleeve: 25.70 (0.39)
Length: 25.73 (0.30)
Shoulder: 17.80 (0.16)

We can say that a RMC J100 size 40 has a smaller chest, shorter sleeves, and shorter back length than most tagged 40 BRs from our sample, and it has wider shoulders than all of them. In fact, it has shoulders that would be comparable to 44 or 46 BR.

Of course, this all assumes that we're all measuring exactly the same as Carrie at Thurston Bros.

Anyway, thank you for coming to my board racer TED talk.

 

[TooManyHatsOnlyOneHead]

data, birds of a feather. Can't tell you how much of this Beautiful Mind crazy rabbit hole crap I do too. Well done!

on your last point about how TB measures. Don't quote me on this, but I feel pretty confident about it. They pull on the leather and take the measurement. So you're basically getting a full measurement at it's max. Most of us in the classifieds are laying flat, giving a tug, and then taking a measurement. I think you can assume at least .25 inch difference, maybe even more depending on the type of leather.

 

[Carlos840]

I don't know what you do for a living, but i like the way you approach this.
I thought i was over the top, it appears i am not after all.

Welcome to TFL!

 

[Marc mndt]

I wonder whether leather thickness or stiffness acts as a moderator. (ie the thicker the leather the more variance in sleeve lenght)

Recent quote from Denny: "Regarding sleeve length we also explained that although slightly shorter than reqested 25.25" they did fall within our standard 0.5' tolerance allowance at 24.75" , some if this is due to your request for the heaviest available skins (with very heavy leather you do loose a bit more length in the seam allowance)."

 

[Monitor]

It's just a jacket...

 

[Marc mndt]

It's just a jacket...

With this line you can prematurely kill each and every discussion on this forum.

 

[Monitor]

Fair enough.

To actually add something constructive to the thread, tolerance allowance aside, Aero has been consistently inconsistent in their measurements which is in part due to the nature of their product being almost exclusively created per order so it is only natural that the numbers are all over the place. Plus, TB's stock page isn't exactly that per se, meaning many of these jackets weren't made to be sold as OTR but are often returns if, for instance, a size didn't work out, etc. These guys have a lot of such jackets in stock so of course they'll sell them as stock items as there's technically nothing wrong with them but they still do not represent a standardized pattern model.

All that is to say, there's not much point in trying to determine an average in measurements for any given size tag for any given style by Aero 'cause inconsistency with numbers is not an error in this case - It's a feature!

That said, this chart would be very useful for any of the true OTR makers, like Vanson or Schott.

 

[MrProper]

It's just a jacket...

I'm afraid it's the artist talking ;-)
A numbers, data, facts person will see it completely differently.

 

[Monitor]

I'm afraid it's the artist talking ;-)
A numbers, data, facts person will see it completely differently.

That is true... It's an admirable effort in any case. I can't even make heads or tails of any of it just by looking at the charts so I don't even dare think what I'd do if I was tasked with the same kind of analysis.

 

[Peacoat]

I agree with Monitor; I can't make heads or tails of it either, but it is an admirable effort.

In my almost 17 years here, that is the strongest first post I have seen. Welcome to the the Lounge.

 

[ton312]

Fair enough.

To actually add something constructive to the thread, tolerance allowance aside, Aero has been consistently inconsistent in their measurements which is in part due to the nature of their product being almost exclusively created per order so it is only natural that the numbers are all over the place. Plus, TB's stock page isn't exactly that per se, meaning many of these jackets weren't made to be sold as OTR but are often returns if, for instance, a size didn't work out, etc. These guys have a lot of such jackets in stock so of course they'll sell them as stock items as there's technically nothing wrong with them but they still do not represent a standardized pattern model.

All that is to say, there's not much point in trying to determine an average in measurements for any given size tag for any given style by Aero 'cause inconsistency with numbers is not an error in this case - It's a feature!

That said, this chart would be very useful for any of the true OTR makers, like Vanson or Schott.

Exactly! Even TB's "stock" is not that at all as the jackets are made to TB's desired numbers....not Aero's. So features like "added body and sleeve length" are baked into TB's spec and mentioned in each description on their sales page. While certainly a valiant effort, this dog won't hunt b/c the harvested data lends little to any information of value as these are not stock jackets from Aero, they are "custom stock" to TB. The best advice is to put in your numbers together and hold them to .5" tolerance in any direction. Ignore the tag size entirely, or let Aero help you pick one.

 

[Monitor]

@ton312, right! TB's specs. Forgot completely they prefer to elongate the jackets they spec. from MTM makers. Their Comet by Vanson is a good 3 to 4 inches longer than the regular version, for instance.

But yeah, it's absolutely essential to have the maker help you out with the size. .5"+/- won't really make any difference, anyhow because it'd have to happen all over the jacket to have any effect on the overall fit, which it won't. Plus these numbers are never 100% consistent & will change to some extent with wear & time.

Another thing, numerical size tag reflects a chest size and it's got more to do with the weight of a person, rather than height. Board Racer in a larger size won't often have longer sleeves. 27" sleeves on a sz. 46 jacket isn't all too common.

 

[poorman]

the harvested data lends little to any information of value as these are not stock jackets from Aero, they are "custom stock" to TB.

That said, this chart would be very useful for any of the true OTR makers, like Vanson or Schott.

All fair points. I agree that this information can't really be used as a means of predicting what your jacket would look like if you were order a new one from Aero. As you pointed out, these jackets were made to specific measurements instead of a reference pattern (e.g., the "official" RMC J100 size 40 referenced in the original post).

But all that being said, the Aeros WERE assigned a tag number despite being made in a nonstandard way, and the tag numbers DO convey information about the size of the jacket. The point of this exercise is to pin down what information is actually being conveyed specifically because the jackets aren't being made to a reference OTR standard. Until the time that websites like TB allow people to filter their jackets by the actual measurements, or the classifieds have measurements posted in the title, I would argue that this kind of information has value.

 

[Will Zach]

Great paper (or is it a pre-print?). I am not a referee for NEJOLJ, but I recommend a prompt publication. Awaiting more papers on topics such as: "Analytical methods of horsehide detection", "Development of sleeve creases as function of species and tanning methods", "Tensile strength and permeability of horsehide as a function of moisture content"

 

[Guppy]

Fair enough.

To actually add something constructive to the thread, tolerance allowance aside, Aero has been consistently inconsistent in their measurements which is in part due to the nature of their product being almost exclusively created per order so it is only natural that the numbers are all over the place. Plus, TB's stock page isn't exactly that per se, meaning many of these jackets weren't made to be sold as OTR but are often returns if, for instance, a size didn't work out, etc. These guys have a lot of such jackets in stock so of course they'll sell them as stock items as there's technically nothing wrong with them but they still do not represent a standardized pattern model.

All that is to say, there's not much point in trying to determine an average in measurements for any given size tag for any given style by Aero 'cause inconsistency with numbers is not an error in this case - It's a feature!

That said, this chart would be very useful for any of the true OTR makers, like Vanson or Schott.

Right the average measurements don't matter. The measurements of the specific jacket you're looking at are very important. You can narrow things down by filtering for tag size, and these numbers can help identify tag sizes that may likely work for you. But you probably could have guessed that anyway. Then you just have to account for the variation of the measurements due to error from inconsistent methods and people just being imprecise. Finally, you should just try it on.

 

[ton312]

Finally, you should just try it on.

Bingo. Esentially the whole point of TB's stock pages.

 

[Canuck Panda]

Excellent work!

I would agree with some of what Monitor said about Aero. They make one jacket at a time, one machinist, using more than one machine, just one jacket. This is very different than batch runs where one man one machine and it takes two to three men / machines to make one jacket in the same batch run.

So essentially even a stock Aero jacket is still a custom jacket by default.

Take my recent board racer for examples, some measurements would fall outside of the range from TB numbers. And all I asked for is widened shoulder and add sleeve length 0.5". The first which I think isn't real just machinist trying to sew at minimum seam allowance vs going whatever, and the sleeve length might be from the cutter, but again I am not sure.

What I am trying to say is that every Aero jacket is gonna be different, even if they are ordered stock, because the nature of one machinist production. I know this because I have the same jacket in five different leathers and same stock size, and they are slightly different somewhat. Aero that is. I also have the same jacket in two different leather from RC which is batch made and they are identical except the leather. Very different production methods will yield very different results.

Here is my BR measure pics. Stock size 42. The shoulder and sleeve falls outside of the TB range. There is a margin of error in measuring too. For torso measurements I try to do the TB way now. So that's the biggest possible size which actually makes a lot of sense to me now.

I am extremely impressed by your statistical work. The way I database my jackets seems kindergarten compared to what you've done. Great work.

 

[Aloysius]

Exactly! Even TB's "stock" is not that at all as the jackets are made to TB's desired numbers....not Aero's. So features like "added body and sleeve length" are baked into TB's spec and mentioned in each description on their sales page. While certainly a valiant effort, this dog won't hunt b/c the harvested data lends little to any information of value as these are not stock jackets from Aero, they are "custom stock" to TB. The best advice is to put in your numbers together and hold them to .5" tolerance in any direction. Ignore the tag size entirely, or let Aero help you pick one.

This was my immediate thought when I saw the post last night but I didn't have the heart to say it.

 

[poorman]

Sure, that could limit the generalizability. Aspects of the sample may reflect all the board racers out there (e.g., those not made [edit—ordered, I should have said] by TB).

But that being said, while TB gives Aero their desired measurements, the tag is determined by the Aero craftsman when it the jacket gets made (I assume). Aero is still the one looking at a jacket and saying “yup, let’s call this a 40.” I imagine TB orders them in a similar manner compared with the individual consumer, just in a much bigger volume (than some anyway—I’ve seen the size of some of your collections here). So these jackets, even those made to TB spec, still reflect what Aero deems within the suitable tag size range, presumably. I expect that the TB sample closely mirrors the general population in PTP, hem size and shoulder length, though I don’t have the data to actually show it.

Like I said above, the purpose of the exercise is to help answer “what does a tag 40 mean?” and related questions. I don’t think the limitations totally negate the findings.

 

[Jasonindenver]

My brain hurts.