A BMX History Library reference page

How This Library Is Made

The sourcing rules, the conflict of interest, the corrections, and the contradictions we leave in

Legend Bike Co · BMX heritage series

This page exists so that a reader who does not trust any of this can check it.

The BMX History Library runs to more than three hundred chapters, cross-linked to each other and to the origin article, The History of BMX. It makes claims about who founded what, who welded what, who won what and when. Some of those claims rest on printed pages from the period. Some rest on institutional records. Some rest on collector research of varying quality. And some rest on the memory of one man who was there, who also happens to own several of the companies in the story.

What follows is how the work is actually done, including the places it is weak. None of it is an argument that the library is right. It is a description of the method, so that anyone can go and test it.

The conflict of interest, stated first

The person who edits this library is also a subject in it.

Bill Ryan founded Supercross BMX in 1989 and still runs it. He owns Torker Racing, TECH BMX Products, Max Racing Pants, Speedline Parts and Apple Valley BMX Moto Park, and he co-founded Legend Bike Co — the company whose website you are reading — with Eddie Fiola and Pete Loncarevich. Supercross, Torker, TECH and Legend Bike Co all appear in this history. So does Bill Ryan, who has a chapter of his own.

There is a live commercial relationship in here too. Alans BMX of Wigan, England has its own chapter, and since 2023 Alans has been the United Kingdom distributor for Supercross BMX and Speedline Parts. That is trade moving between the subject of a chapter and the owner of the site the chapter sits on.

None of that is a reason to leave the material out. BMX is a small industry and it has always been written about by people inside it. A rule that barred anyone with a stake from writing the history would hand the history to people who were not there. But it is a reason to be exact about how those chapters are handled, so here is the handling in full.

• The same sourcing standard, with no exceptions. The chapters on Bill Ryan's own companies are held to the same evidence rules as the chapters on any other brand. Where the sourcing on one of his brands is thin, the page says so, in the same box and the same words used everywhere else. The Supercross BMX chapter carries a list of open questions about its own frame-building history — which years the frames were welded in house, when each builder handed over — because those dates are not documented and nobody's memory is a document.
• His claims are testimony, like anyone else's. When Bill Ryan is the source for something, he is named and dated in the text as a source. He is not written into the record as established fact, and he gets no standing that the evidence does not give him. Where his account and the paper disagree, the disagreement is printed.
• Commercial relationships are disclosed where they occur, not here. The Alans BMX chapter carries its own disclosure in the body of the page, next to the fact it applies to: “In the interest of full disclosure: Supercross BMX and Speedline Parts are sister companies to Legend Bike Co. That relationship is why we know this story well, and it is stated here rather than left for a reader to find out.” A disclosure parked on a standards page most readers never open is not a disclosure.
• No brand gets elevated. That rule is printed at the top of the origin article and it is the rule here: no brand gets elevated, no rider gets shorted, and where the record is contested the library says so. A reader is entitled to check whether the Supercross chapter reads any softer than the chapters on brands nobody here owns. That check is the point of publishing the standard.

What this page cannot do is make anybody trust the arrangement, and nobody has to. The chapters are there, the sources are named at the foot of every one of them, and the disclosures sit on the pages themselves. Go and look.

What this is, and what it is not

This is a connected history. The chapters are written to be read against each other: a rider chapter names the frames he rode and links to them, a brand chapter names the riders and the shops and links to those, and the origin article ties the whole run together. When a fact turns up in two places it has to agree in two places, and when it does not, that is how errors get caught. Several of the corrections described below were found that way and no other.

It is not a database. There is no complete results ledger, no serial number registry, no catalog scan archive. It is not a forum, and it is not a substitute for one — the community discussion held on the BMX forums is a different and often better kind of record, and this library cites it rather than replacing it. It is not peer-reviewed. It has no editorial board, no fact-checking desk and no institutional backing. It is one project, written in the open, with its sources listed so that its mistakes can be found.

It is also not finished, and not everything written gets published. Some chapters are drafted and held back because the sourcing under them is not good enough to put a person's name to. A page that is not up yet is usually a page that could not be stood behind, not a page nobody got to.

Where the material comes from

Four tiers, in descending order of weight. Every chapter names its own sources at the foot of the page, and the tier each one belongs to should be obvious from the naming.

1. Period print — the strongest thing available

Magazines, catalogs, advertisements, price lists and spec sheets printed in the years in question. A dated page beats every kind of recollection, including a good one, because it was made before anybody knew what the story would turn out to be.

The clearest example in this library is a single line in a component list. BMX Action's Probe column for March 1989 ran a bike check on Kevin Hull's customized Brackens under the title The Sheep Dogmobile, photographed the previous October. Printed in the spec list is the line:

Six words in small type settle a question that no amount of remembering could: a Belfast-made number plate was bolted to an American factory pro's race bike in Southern California in October 1988, and the biggest BMX magazine in the country printed it. That is what period print does.

2. Institutional records

The National BMX Hall of Fame, USA BMX, the UCI, British Cycling, the British BMX Hall of Fame, the Texas Bicycle Motocross Hall of Fame, and the IBMXF archive kept by the federation's general secretary of the period. These are reliable for results, dates, inductions and rosters, and they are the right authority for anything a sanctioning body would have recorded at the time.

They are also used in the negative, which is worth stating because it is easy to miss. When a chapter says a person is not among the inductees listed by a hall of fame, that is a check that was actually run against the index, not an assumption.

3. Collector-tier sources — used widely here, and weaker

Forums, brand reference pages, enthusiast research sites, auction and sale listings, and archived interviews on community sites. This tier carries a large share of the load in this library, and it needs saying plainly: it is the weakest tier used here.

Some of it is excellent. The deepest Torker research available anywhere was done by a collector working from period advertising and race photographs rather than memory, and this library leans on it and names it. But collector material is unversioned, unattributed as often as not, and prone to a particular failure mode: a claim gets repeated between sites until the repetition starts to look like corroboration. Three pages saying the same thing is not three sources if all three took it from the same place. Where a claim in this library rests only on this tier, the chapter says so rather than dressing it up.

4. Firsthand testimony

People who were there. Handled as set out in the next section.

The two archives that are missing

The two deepest repositories in BMX could not be reached for this work.

That is not a small gap and this page is not going to soften it. It is the single largest structural weakness in the work. Individual chapters record it at the foot of the page wherever it bites, naming what those sites are likely to hold rather than pretending the search was complete. Anyone with access to either archive can close more of this library's gaps in an afternoon than months of other research has.

How firsthand testimony is marked

Firsthand testimony is always attributed and dated in the text, in the form Bill Ryan, firsthand, 2026, and it is never blended into the documentary record. If a document and an account cover the same event, the two are set out separately, because they prove different things and neither should be made to carry the other's weight.

The JIVE number plates chapter is the worked example, and it is worth following in full because it shows the method doing something useful rather than just being careful.

For a long time that chapter had a hole in the middle of it. JIVE was a clothing and number plate label run out of Belfast in the late 1980s by Will Smyth and a handful of friends, who cut plates by hand at night in a friend's house. It was documented at the Belfast end. The American end was not. Nothing put a JIVE plate and an American rider in the same place.

Two pieces of evidence closed it, from different directions.

The document. The March 1989 BMX Action Probe spec list, quoted above. It proves the plates were here — on a real bike, at a fixed date, in print, under a name spelled out in a component list. It proves nothing at all about how they crossed the Atlantic. A plate can travel more than one way.

The account. Bill Ryan, firsthand, August 2026, supplies the route and the riders: the panels ran on Kevin Hull's bike, on his own bike, on Erick “Big E” Bartoldus's bike and on Craig “gOrk” Barrette's, and through the wider BMX Action circle — and they were imported by Michelle Sheehan of S&S Bikes in Texas, who got them out to riders. That is testimony rather than paper, and it is labeled as such on every page it appears on. It is also about as direct as testimony gets: Ryan took delivery of a plate himself and can name the people who bolted the others on.

Put together, the two strands answer a question neither could answer alone. Kept apart, a reader can see exactly which half is documented and which half is remembered. That separation is the whole rule, and it is why the JIVE, Michelle Sheehan and Brackens chapters all state the same distinction in the same terms.

How corrections work

Corrections are logged, not quietly overwritten. When a page changes on a point of substance, the change and the reason for it stay on the page, usually in the sources note at the foot. A silently edited page teaches a reader nothing about how much to trust the next one. Three worked examples, all of them live on the site now.

A name that moved three times

The first Supercross BMX frame was welded at Free Agent, in a garage in Buena Park, in 1989. Getting the welder's name right took four passes.

The Supercross chapter's own sources note records the whole sequence rather than the answer: the original message rendered the name as “Brent Should”; a first pass read it as Brent Shoup; a later message gave it as Brad Shoup; a further message settled it as Brent Shoup and set out the family — Brad and Yvonne Shoup married and founding Free Agent, Brent their son, and Brent doing the welding.

What settled it was not insistence. It was the library's own Free Agent chapter, which had the family in it and which forced the two accounts to be reconciled instead of left disagreeing. And that chapter did not simply switch over. It had previously run the roles the other way around — Brent and Yvonne as the founding couple, with Brad, described as Brent's father, welding the early frames — and rather than delete that version it now carries both, in a flagged box, leading with the firsthand account and calling neither one wrong.

The KHS sale date, and a source who told us to go with the documents

Free Agent was distributed by KHS Bicycles from 1994 and bought outright by KHS in 1996. Those are the published dates and they are the dates the chapter states.

For a while a second account ran alongside them. Bill Ryan, firsthand in 2026, put the sale at about 1992 and flagged it himself as an approximation. Shown the published dates in August 2026, he deferred to them:

That exchange is kept on the Free Agent page rather than tidied away, and it is the single most useful thing in this library for judging how much weight to give the rest of it. A firsthand source who tells you to go with the documents is worth more, not less. It is also the reason the whole project exists, in his own words: a forty-year-old memory is foggy, which is why it is being written down now, while there is still anybody around to check it against.

The rule that has stopped two death claims

Nobody in this library is recorded as having died without a hard source that names the right person. Silence is recorded as silence, in either direction. A common name is not a hard source, and a search result is not an identification.

That rule has done real work. Twice during this research, material turned up that could have been read as a death notice for a man who is living, and in both cases the requirement to verify every claim caught it before anything went to print. Neither man is named here. The near-miss is not part of his story and his page does not carry it, so this page will not put it there either.

What the pages do carry is the rule operating in the open. One chapter records that no obituary, memorial or death notice for its subject turned up in any source checked, and that nothing will be implied from that silence in either direction. Another states that the weight of evidence is that its subject is living — he wrote a signed first-person piece in 2017, he is credited by name with relaunching his company in 2023, and a trade article in 2024 describes him in the present tense — and then says outright that this is an absence of evidence rather than a confirmed fact from a hard source. A third records a subject as alive and well on firsthand authority, and adds the reason it needed that authority: web searches on his name turn up unrelated people who share it.

None of that reads as confidently as a plain sentence would. It is correct, which matters more.

What is deliberately left unresolved

Where two accounts conflict and nothing available settles it, this library states both, calls neither wrong, and flags the disagreement on the page. That is a deliberate choice and it is worth explaining, because it looks like indecision and is the opposite.

Picking a winner between two accounts the evidence cannot separate does not resolve the conflict. It hides it. The reader loses the information that there was ever a question, the losing account drops out of the record entirely, and the one person alive who could settle it never learns that anything needed settling. A visible conflict is a request for help with the location marked. A smoothed one is a dead end that looks like an answer.

Live examples, all checkable now:

• The Titan founding date. The Titan BMX chapter carries 1982, the year the brand record gives. Three collector sources put Kendall Burleson leaving Black Lite in mid-to-late 1983 and starting Titan in late 1983 or early 1984. Asked in 2026, Bill Ryan said he could not confirm either date from memory. The page leaves 1982 in place and flags the disagreement, and the other half of it is set out on the Black Lite chapter.
• Michelle or Brenda Sheehan. Michelle is the name Bill Ryan knew her by and the name Supercross has always used. The Sheehan family's own memorial site for her son uses Brenda. Both may be correct — a formal first name and the name she went by are an ordinary thing to find in one family — and the Michelle Sheehan chapter records both rather than correcting the family or dropping one.
• Who funded Free Agent. Voris Dixon has said on the record that he funded the original operation, handed a thousand dollars to Yvonne Shoup to open a checking account, came up with the name, and holds paperwork showing half ownership he was never paid for. The Shoup side, as it appears in public materials, has Free Agent as the family's brand from the start. There are no filings, no transfers and no ownership documents anybody can point to on either side. The chapter reports both accounts and settles neither.
• The date of the first nac-nac caption. Recorded as December 1988 in one account and 1987 in another. The March 1989 BMX Action photograph of Will Smyth is the better-attested of the two appearances. The earlier one is unresolved and the JIVE chapter says so.

The same habit runs through the origin article. Where sources conflict on the date of the first BMX race, The History of BMX follows the recognized November 14, 1970 date at BUMS while noting that earlier competing claims exist, rather than presenting a contested date as a settled one.

How to correct it

This library is built to be corrected, and the correction rate is high enough that the mechanism matters more than the current text.

What is wanted: firsthand accounts from anyone who raced, worked, welded, sold, wrote or officiated in any of this. Period documents of any kind — catalogs, advertisements, price lists, order forms, packing slips, race programs, results sheets, team flyers, shop photographs. Dated photographs above almost anything else. Corrections to names, spellings, dates, rosters and family relationships. And plain statements that something here is wrong.

What happens to it: material that is used is attributed to whoever supplied it and dated, the same way every other source on the page is. It is published even when it contradicts what is already there — particularly then. Where it conflicts with something already on the page and cannot be reconciled, both versions go up and the conflict gets flagged rather than resolved by preference. Where it settles something, the previous version stays visible in the correction note so that a reader can see what changed and why.

It can be sent through the contact page.

The standing rules, in one place

Every chapter names its own sources at the foot of the page. Period print outranks everything else. Institutional records are used for results, dates and rosters, and are cited in the negative when a check returns nothing. Collector-tier sources are used widely and are the weakest tier here; where a claim rests only on them, the chapter says so. Firsthand testimony is named and dated in the text and is never merged into the documentary record. Documents and accounts covering the same event are set out separately. Nobody is recorded as dead without a hard source naming the right person, and silence is recorded as silence. Conflicts that cannot be settled are printed as conflicts, with both accounts stated and neither called wrong. Corrections are logged on the page rather than overwritten. Commercial relationships are disclosed on the page where they occur. Chapters covering companies the editor owns are held to the same standard as every other chapter, and the editor's own claims are treated as testimony. bmxmuseum.com blocks automated access and bmxsociety.com was unreachable throughout this research; neither is cited anywhere in this library, and both are flagged as open gaps on the chapters where they would have helped.

Last reviewed: August 2026. This page describes practice, not aspiration. Where a chapter departs from it, the chapter is wrong and we would like to be told.