Wheelbase, track and centre of gravity
The three numbers that decide whether a machine stays upright.
Whether a machine tips over comes down to a simple contest between three measurements, how wide apart its wheels sit side to side, how far apart they sit front to back, and how high up its weight is carried above the ground, and a machine can be perfectly sound in every other respect and still tip over reliably in the field the moment any one of those three numbers is wrong for the job it is actually being asked to do.
Why three simple measurements decide whether a machine tips over
A machine resting on the ground is stable for as long as its own weight, acting straight down from its centre of gravity, falls somewhere inside the footprint its wheels or supports actually mark out on the ground. Tilt the machine, whether from a side slope, a sharp turn, or hard braking, and that same downward line from the centre of gravity starts to shift sideways relative to the footprint, and the machine tips the moment that line finally moves outside the footprint's edge entirely, at which point the machine's own weight stops holding it down and starts actively pulling it over instead. Nothing about the machine's structure has to fail for this to happen, no bolt shears and no bearing seizes, the tip is a purely geometric event governed entirely by where the weight sits relative to the footprint beneath it, which is exactly why a machine can be structurally sound in every respect and still roll over on a slope its own strength calculations never flagged as a problem at all. How much tilt a machine can absorb before that happens depends directly on two things working against each other, how far the footprint's edge actually sits from the centre line, set by the track width for a sideways tip and the wheelbase for a forward or backward one, and how high up the centre of gravity sits above the ground, since a higher centre of gravity swings sideways by more for the same angle of tilt and crosses the footprint's edge correspondingly sooner.
The barstool comparison
A barstool with its legs splayed generously wide apart resists a firm sideways push remarkably well, the stool barely rocking even when leaned on hard, since the legs' wide base gives a great deal of room for the stool's own weight to shift sideways before it ever approaches the edge of that base. The identical stool built instead with its legs drawn in close together underneath the seat tips with only a gentle push, the narrow base offering the stool's weight almost no room to shift sideways at all before it reaches the edge and starts falling rather than resting. Raising the seat height on either stool makes the problem worse again, since a taller stool's weight sits further above its own base and swings sideways by a larger distance for the identical lean, reaching the edge of an unchanged base sooner than the shorter stool's weight ever would. A machine's track width plays exactly the role the stool's leg spread plays, and its centre of gravity height plays exactly the role the stool's seat height plays, the same contest between a wide, low base and a narrow, tall one deciding whether either object stays upright under a lean.
Why wheelbase governs tipping the other way
Track width and centre of gravity height between them decide how much a machine can lean sideways before tipping, but a machine tips forward and backward too, under hard braking, hard acceleration, or driving straight up or down a slope, and that direction of tipping is governed instead by the wheelbase, the front-to-back distance between the axles, playing the same role the track width plays for a sideways tip. A machine with a short wheelbase and a high centre of gravity, a narrow-footprint forklift carrying a raised load, for instance, can be perfectly stable side to side while still being genuinely vulnerable to tipping forward under hard braking, since the short front-to-back footprint gives the centre of gravity very little room to shift forward before it crosses the front edge of that shorter base. A machine reviewed only for sideways stability, checking track width and centre of gravity height alone, has answered only half the real question, since the wheelbase is doing the identical job in the other direction and can be the weaker of the two limits depending entirely on the machine's own proportions. A machine long and narrow in plan, wide wheelbase and narrow track, tends to resist tipping forward and backward comfortably while remaining genuinely vulnerable sideways, and a machine short and wide, the opposite proportions, tends to show the reverse weakness, which is why the two directions have to be checked as separate questions rather than assumed to track each other simply because they both fall under the same general heading of tipping.
One figure worth keeping in mind
A machine's static stability against tipping in either direction is set by the ratio of half its relevant footprint dimension, track width for a sideways tip, wheelbase for a forward or backward one, to the height of its own centre of gravity above the ground, so doubling that footprint dimension, or halving the centre of gravity's height, each roughly doubles the amount of tilt the machine can absorb before its own weight starts working to tip it over rather than to hold it down.
What this changes in practice
Reviewing a machine's stability means checking all three numbers together rather than any one of them on its own, since a wide track paired with a dangerously high centre of gravity can look stable at a glance and still tip readily on a real slope, and a machine whose working task genuinely raises its centre of gravity, carrying a load overhead, lifting an attachment, filling a tank mounted high on the frame, needs its stability rechecked against that raised condition specifically rather than only against its own unloaded, resting configuration, since the resting machine and the working machine are, for exactly this purpose, not the same machine at all.