GolfIssues
Search

Menu

Sign in or create a free account

Optional. Every page reads in full without one.

Contact

Methodology

How a diagnosis is made

This site has no byline, no headshot, and no list of credentials. That is deliberate, and it places an obligation on this page: if you cannot check who wrote something, you should be able to check how it was reasoned. So here is the whole method. What a diagnosis actually is, how it is produced, the physics it rests on, and precisely where it stops being reliable.

Why there is no author here

Authority on a golf site is usually a person: a name, a tour credit, a certification, a photograph mid-swing. There is none of that here, so the work has to be done differently.

  • Every claim traces to one of four ball-flight relationships, or to a setup principle nobody seriously argues with. Anything that cannot be traced to those is not published, however popular it is.
  • Every symptom, cause, and fix records the sources it rests on, so a claim can be followed back rather than taken on trust. The full list is at the bottom of this page.
  • Nothing is generated. There is no model anywhere in this product, at build time or at the moment you press an answer. The content is written by hand and the engine only chooses between pieces of it.
  • Nothing is dressed up. No seeded comments, no invented usernames, no counts that were never earned.

None of that makes this better than a coach who can watch you swing. It makes it checkable, which is a different and smaller claim. The limitations are part of the method, not a footnote to it.

How a diagnosis is produced

Eight steps, in order. Nothing happens between them that is not described here.

  1. You name a symptom, not a fault

    You start from something you can actually observe: what the ball does, or what the strike felt like. “The ball starts left and keeps going left” is a symptom. “I am coming over the top” is a diagnosis, and it is the thing you came here to find out, so the site never asks you to supply it.

  2. The symptom carries a fixed shortlist of causes

    Each symptom is authored with the handful of root causes it can plausibly resolve to, ordered by how often each one turns out to be the real fault. That list is hand-written and reviewed. The engine cannot invent a cause that is not on it, and cannot return one from another symptom.

  3. You answer between three and six questions

    A question is written only if it separates those causes. If every golfer with this symptom would answer it the same way, it tells the engine nothing and it does not get asked. That is why the questions feel oddly specific: which way your divot points, how many knuckles you can see at address, where your weight ends up. Those are the answers that split one cause from another.

  4. Each answer adds weight to specific causes

    Every option carries weights toward one or more causes on a nought-to-five scale, relative within its own question. Answers only ever add weight; nothing is ever subtracted. A question you are unsure about therefore contributes less, but it cannot actively push you away from the right answer.

  5. The causes are ranked by share, not by score

    When the questions are done, the weights are totalled for each cause and converted into that cause’s share of all the weight collected. Share rather than raw score, because a raw total only tells you how many questions mentioned a cause; the share tells you how much of the evidence actually pointed at it.

  6. If the top two are within 15%, it says so

    A gap of less than fifteen per cent between the leading cause and the runner-up means the questionnaire has not separated them. Picking one anyway would be the easiest way on this site to be confidently wrong, so the result is marked as a close call: the order the causes were authored in, most likely first, decides which is shown as primary, and the runner-up appears immediately alongside it rather than buried under it. Work through the primary first. If a fortnight of honest practice changes nothing, the runner-up is where to go next, and that is a real answer rather than a hedge.

  7. Your answers select writing that already exists

    Nothing is generated for you. Your answers decide which pre-written fix you land on and which pre-written passages appear on it. A fix page is complete on its own: arrive cold from a search result without answering anything and you still get the whole explanation, because the personalised passages are additive and never replace the core of it.

  8. Paths that do not resolve are logged, not faked

    Some combinations of answers do not land cleanly on one cause. When that happens you get a page that says so and points at the likeliest candidates, and the combination is recorded so the gap can be written properly. A tool that always produces a confident answer is not more accurate than one that admits a gap; it is just less honest about the same uncertainty.

The physics this rests on

Four relationships carry almost everything on this site. They are not controversial, they are not a theory of the golf swing, and they are the reason a questionnaire can say anything useful at all.

Start direction and curvature have different causesA golf shot seen from above. A dashed line runs from the ball straight ahead to the pin, marking the target line. The shot leaves the ball along a straight line angled slightly left of that target line, then bends steadily to the right and finishes right of the pin.
Seen from above. The straight line leaving the ball is the start line, set by where the face points at impact. The bend away from it is the face-to-path relationship. Two different causes, which is why they need two different fixes.

  • Start direction is governed primarily by face angle

    Where the ball begins its flight is set almost entirely by where the clubface is pointing at impact. The path it is swung on influences the start line, but far less. This is why the first question in most flows asks where the ball starts before it curves: it is the single most informative thing you can report.

  • Curvature comes from the face-to-path relationship

    Which way the ball bends, and how much, comes from where the face points relative to the direction the club is travelling. Open to the path bends it one way, closed to the path bends it the other, square to the path is straight, whichever direction that path happens to be going. It follows that closing the face on a path that still travels left does not cure a slice; it converts it into a pull.

  • Fat and thin are both low-point problems

    They are not two faults. They are one: the lowest point of the swing arc is in the wrong place relative to the ball. That is why they alternate in the same round, and why treating them separately produces two fixes that fight each other.

  • Strike location explains distance loss and gear effect

    Where on the face you make contact accounts for distance that disappears without any change in effort, and, on a driver, for curvature that the swing alone cannot explain, because a strike away from the centre imparts spin of its own. A face tape or a dry erase marker settles in one shot what a golfer can argue about for a season.

  • Setup principles, where they change one of the above

    Grip, aim, ball position, posture, stance width. These appear only where they move one of the four relationships: a weak grip does not cause a slice by itself, but it makes an open face the default outcome rather than something you have to produce.

Anything that cannot be traced to one of those does not get published here. No opinions about the plane of the backswing, no positions asserted to be correct because a tour player has them.

What this site deliberately will not do

  • No swing-theory opinion

    There is no position of the arms at the top that this site will tell you is correct. The physics spine above is the whole of what is claimed, and it stops well short of an opinion about how a swing should look.

  • No named methods

    No system is being sold here, so there is no reason to argue about anybody else’s. Where a widely taught idea is genuinely relevant it is described in plain terms rather than by the name of its author.

  • No equipment recommendations

    No brands, no shaft advice, no affiliate links, no sponsored anything. Equipment genuinely matters, but it cannot be judged through a questionnaire, and a site that earns money from recommendations acquires a reason to make them.

  • No medical claims

    If something hurts, that is not a swing fault for a questionnaire to diagnose. Nothing here is medical advice, and a diagnosis that ignores pain is worse than useless. See someone qualified to look at it.

  • No manufactured popularity

    No seeded comments, no invented usernames, no votes that nobody cast. A count appears on a page only once it reaches 10; below that the element is shown without a number rather than advertising how quiet it is.

Limitations

These are not disclaimers added at the end by a lawyer. They are the actual boundary of what a questionnaire can do, and they are worth reading before you act on anything here.

  • It cannot see your swing

    Everything above is inference from what you report. A phone on the ground behind you, a launch monitor, or an hour with someone who can watch you all beat it outright, and none of them is in competition with this site. Use this to work out what to look for, then go and look.

  • Some causes cannot be separated by any questionnaire

    When two faults produce the same observable evidence, no set of questions will split them. That is precisely what the fifteen per cent rule exists for, and when it triggers the honest answer is that both are live, not a confident pick with a footnote.

  • Self-reported answers are imperfect

    Golfers routinely misread their own start direction, and almost nobody’s memory of a swing survives contact with the video. That is why the questions lean on evidence you can physically check: a divot on the ground, the knuckles visible at address, which foot you finish on. It is also why “I have not filmed it” is an allowed answer instead of a forced guess. An honest unsure beats a confident wrong every time, because the weights are designed to absorb it.

  • It does not know your context

    Injury, a physical restriction, a club that does not fit you, the lie you were playing from, the change you made two shots ago: none of it reaches the engine unless a question asks about it, and no questionnaire asks about everything.

  • A diagnosis is a ranking, not a verdict

    The output is the most likely cause given what you reported. Most likely is not certain. If the primary fix produces nothing after a fortnight of honest work, that is information: go to the cause listed beneath it rather than concluding that the fault is incurable.

Sources

Every symptom, root cause, and fix stores the keys of the sources it rests on, so any claim on this site can be followed back to something outside it. The list below is short on purpose: a source is added only once it has been confirmed to exist. A long list of plausible-looking citations is worse than a short list of real ones, because it invites a trust it has not earned.

Corrections

If a diagnosis sent you somewhere wrong, that is a bug in the content and it is worth reporting. So is a question that cannot be answered honestly, a fix that reads as though it was written for a different fault, and a claim that does not survive the physics above.

Send it through the contact form. It reaches whoever maintains the content, and the content files are edited in response. There is no comment thread on this page to argue in, because a methodology that changes by argument is not a methodology.

If you want to see the method working, start from a symptom on the list of issues.