A driver comes off track and says the kart pushes, or feels loose, into turn three. The session notes don't say which part of the corner it happened in, what the track was doing, or what the kart was set up to that day. That's a normal starting point for a setup complaint. It's also not usable yet — a feeling with no address attached to it.

The fix isn't to guess at a cause. It's to slow down and do two things first: work out what actually applies to this kart, this event, and this driver, and write down what was observed without turning it into a diagnosis. Only after that does it make sense to change one thing and see what happens.

This article walks through that sequence and ends with a compact card you can copy and fill in yourself. By the end you'll have a completed baseline-and-test card and a bounded decision — retain the change, reverse it, or defer judgment — instead of a guess dressed up as an adjustment.

This is general process guidance for owner-operated karts everywhere. It doesn't hand you a globally legal configuration, a universal setting, or a fixed order of adjustments. You still need to identify the exact kart and components you're working with, then check current manufacturer instructions. The rules matter too — your national authority, your class or series, and whatever your organizer, event, and scrutineer require.

Establish what applies

A baseline is simply the starting configuration you wrote down before you changed anything — the thing you compare your next result against. Without a written baseline, "it feels different now" is hard to interpret, because you don't have a clear record of what it felt like before.

Before you record one, keep three separate ideas apart — they answer different questions:

  • A manufacturer's starting configuration applies to the specific product and conditions it names — nothing else automatically.
  • An event-legal configuration depends on the current authority, class or series, and whatever your organizer and scrutineering process require — regardless of what a manual says is technically possible.
  • A configuration that worked somewhere else only counts once you've confirmed it applies to your own equipment and conditions.

Keeping them separate is what stops a product- or event-specific detail from being treated as a universal rule. The next two sections show what each one looks like in practice.

Manufacturer guidance has a scope

CRG's own racing-kart manual is a useful illustration of how narrow this guidance is. It separates direct-drive configurations from geared ones, and dry conditions from wet. It also names variables like widths, geometry, chassis or stub-axle positions, and axle types among what its setup material covers — the kinds of things a manufacturer document expects you to work out for your own kart.

The Slack Axiom chassis guide makes the scoping even more explicit. It describes its own information as a reference and a starting point, and says tracks and driving styles can require variation from it. It also draws its own internal lines — dirt versus pavement, junior versus senior use — and names driver weight, track size, grip level, and tyres as variables relevant to that particular chassis.

Treat both of these as a manufacturer-specific example of what setup documentation looks like, not as a source of numbers to copy. Whatever a document says is a starting point, not a universal setting, and what applies to your kart depends on the named chassis and conditions in your own manufacturer material — not on what worked for a different chassis, a different surface, or a different driver.

Legality belongs to the event

A few terms are worth defining before this gets confusing. A scrutineer is the event official who checks that your kart complies with the technical rules — check your event documents for exactly when and how that happens. An ASN is the term the cited regulations use for the recognized national sporting authority in a given country. Homologation, in the 2026 CIK-FIA regulations cited below, means the applicable homologation form identifies the approved product or specification, and homologated parts must be used as represented on that form.

The 2026 CIK-FIA technical regulations are a useful illustration of how these pieces fit together. They define their own international categories, use "ASN" for the national authority they recognize, and set out how valid homologated material and conformity with homologation forms apply within those categories — including the conditions under which an ASN may permit previously homologated equipment to keep being used.

Treat all of that as a 2026 CIK-FIA international-category example. It does not establish legality elsewhere, and it isn't a substitute for checking your own situation. Before you treat any configuration as permitted, go to the current national authority, series or class, organizer or event, and scrutineer that actually apply to you. A rule that's right for one category, one season, or one country can be wrong for yours.

Describe before diagnosing

Once you know what applies, the next job is to write down what the driver felt. One avoidable failure mode looks like this: someone says "it's loose" and someone else reaches for a wrench before anyone has agreed on where "loose" happened.

A short observation framework helps here. It's Motorsigma editorial guidance rather than a manufacturer method, giving you a repeatable shape for the description:

  • Where it happened — on the approach or corner entry, somewhere in the middle of the corner, or on exit.
  • What the driver felt at the controls — more steering input needed, the rear stepping out, the kart not rotating, and so on, in the driver's own words.
  • Whether it repeated — once, every lap, only under certain conditions.
  • What might have interfered — traffic, a changing track surface, weather moving in.

That's it. Write down what you noticed, not what you think it means. This is an observation, not a diagnosis — describing the corner phase or location and what the driver felt tells another person what happened without asserting why. A mechanic or a chassis representative can use that description as one input if you bring it to them later, but this article doesn't supply a cause, and neither should your notes at this stage.

Stop when the prerequisites are uncertain

Before you go any further, this next boundary matters more than anything else in this article.

Stop if the card's prerequisite status shows mechanical condition, fitment, safety, or legality/scrutineering as unresolved. Don't run a comparison test on a kart you're not sure is sound or permitted. Check with whichever of the applicable manufacturer instructions, the organizer or governing authority, the scrutineer, and qualified local mechanical support actually bears on the unresolved field, and leave it unresolved on the card until it's genuinely completed or not applicable.

This isn't a formality. A test only means something if the kart underneath it is in a condition you trust. If you're not sure a bolt is torqued right, a bearing is sound, or a part is even legal for your event, resolve that first. None of this article's testing method is a substitute for that check.

Run one bounded test

With the prerequisites settled and the observation written down, you're ready for the test itself — and it's deliberately narrow. Record the starting state, choose one documented variable to change, make that one change, and compare the next result against the baseline under conditions you've also recorded. Then decide: retain, reverse, or defer.

The discipline of one documented variable is what makes the comparison interpretable at all. Change two things at once and you no longer know which one did anything. But even a clean one-change comparison does not prove the cause — it narrows things down, it doesn't settle them. Track and tyre conditions can differ between runs, and a driver's own consistency varies lap to lap. Conditions may still confound the comparison even when you've been careful about the rest of it. Keep that in mind heading into the decision below.

Baseline-and-test card

Here's the record itself. Copy it, fill it in before you touch anything, and use it again after the change.

Identity and applicable documents

  • Kart or chassis, model or identifier, and the specific components involved.
  • Manufacturer document checked, including version or date if you can find one. This manufacturer-specific example applies only to the chassis and conditions it names — match it to your kart's current documentation rather than copying its settings elsewhere.
  • Current authority, class or series, and organizer or event source. This applies right now, at your event, for your national authority — last season's paperwork or a rule from somewhere else won't confirm it.
  • Prerequisite status: mechanical condition, fitment, safety, and legality/scrutineering, each marked completed, unresolved, or not applicable, plus who to contact next for anything not completed.

Session and starting state

  • Date, track, and surface. Weather, tyres, driver, and any other conditions you can reliably note.
  • Starting configuration for the item you're considering changing.

Observable concern and proposed change

  • Location or corner phase and driver input. What was felt, whether it repeated, and any obvious confounders (traffic, changing grip, and so on).
  • The one proposed change — and only one.

Result and decision

  • Before-and-after observations, recorded under noted conditions.
  • Decision: retain, reverse, or defer — with a short reason.

Keep it to one page. The fields here echo what shows up on real manufacturer worksheets, including the Mach1 kart test sheet and its companion setup template. Those worksheets cover session details like tyre pressures and grip or weather conditions. They also cover drivetrain and gearing details, plus driver, event, and lap-time information. That overlap is why this card keeps those fields — a compact editorial choice, not a conclusion the worksheets prove. It doesn't mean filling them in explains a result on its own, or that this card is a complete testing system.

Here's the stop-the-test case from earlier, shown on the card itself. (Invented example, not a real test day.) A driver arrives for a test session and starts filling in the prerequisite status line first: mechanical condition and fitment — completed, checked that morning; safety — completed; legality/scrutineering — unresolved, because the scrutineer hasn't seen the kart yet and the event requires that check before any track session. Contact: event scrutineer, expected on-site later that morning. Every field below the prerequisite line — starting configuration, the observation, the proposed change — stays blank. The result isn't retain, reverse, or defer. It's simpler than that: test deferred — not run.

Make the decision

Not every comparison needs a clean verdict. Retain the change when the recorded comparison, under conditions close enough to compare fairly, supports keeping it; reverse it when the comparison points back toward the starting configuration instead. Defer — decide nothing yet — when any of the following is true:

  • the baseline itself was uncertain
  • the kart's mechanical condition or fitment is in doubt
  • the change's legality hasn't been confirmed
  • conditions shifted enough between runs to make the comparison unreliable
  • the observation didn't repeat consistently enough to trust it

A borderline result belongs on that list too. If the symptom eased on some laps but still showed up on others, that's not a repeat you can trust yet — record it as inconclusive and defer, the same as the triggers above, rather than reading partial improvement as a retain.

Here's what deferral looks like in practice. Say a driver runs a session, reports the kart feels loose on corner entry, and the team makes one documented change before the next session. Between runs, though, a rain shower passes through and the track goes from dry to damp before drying again unevenly. The driver comes in reporting something different — but there's no way to tell whether that's the change working, the surface shifting, or both at once. The honest move is to write "defer" on the card, note the confounder, and plan to repeat the comparison once conditions are stable enough to trust the result. No invented lap time, no claimed improvement — just a documented reason to wait.

That's the whole point of keeping the decision bounded. One change and one comparison don't remove every confounder, and they don't hand you proof. They give you a defensible next step, recorded well enough that the next person — a teammate, a mechanic, your own future self — can pick up exactly where you left off.

Where to go next

This article sits in the kart setup hub, which covers the broader work of building and testing a chassis setup over a season. From here, two natural next stops are turning a specific complaint into one bounded diagnostic hypothesis and working through a full setup test with a copyable worksheet once you're ready to go further than a single change. It's also worth working through a driver-first handling checklist before you attribute the concern to the kart at all. All of this sits inside the wider karting section, which covers everything from getting started to race weekends.

Your next step: identify the exact kart and relevant components, then write the applicable manufacturer document and rules authority into the first lines of the baseline-and-test card before you consider any change.