A corner should change the rest of the lap

Designing SLIPFAULT around earned speed, recoverable mistakes and mastery that comes from the player.

September 2026

When we design a corner in SLIPFAULT, we have to think beyond where the road goes. We have to ask what the player brings into it, what they can earn there, and what they carry out. Exit speed matters. So does the boost they have saved for a section they cannot see yet.

That has become a useful way to judge the whole game. We want a lap to contain decisions whose consequences last beyond the moment of input. Here are some of the commitments we have made, and the problems we still need to solve.

+A corner should pay later

The central commitment is drift, charge, boost. Drifting builds a meter. Completing it and leaving the drift banks a booster. Spending that booster is another decision, separated from the corner that earned it.

That separation matters. Imagine a turn followed by a short straight and another tight turn. Spending immediately might give you speed you cannot use. Keeping the booster for the next exit can change the value of the first corner. The racing line now includes a resource decision as well as a geometric one.

This also changes how we design tracks. We think about where a player earns speed, where they can carry it, and where they spend it. A straight needs a job in that sequence. We want demanding sections followed by room to use what the player has earned, with enough breathing space to make the next demanding section feel different.

=Depth in the same controls

One of our design rules is that the button should be easy and using it optimally should be hard. Steering, throttle, drift and boost already give us a lot to work with: how you enter a slide, how long you commit, when you redirect, and what state you leave the kart in.

The interesting difference between two drivers should be visible in those decisions. We want someone to get around a corner early in their first session and still be finding better ways through it much later.

That is an ambition with a real cost. Small controls do not automatically make a game understandable. Our own design review calls out how poorly the game can explain why one drift pays and another does not. We have work left to do on teaching and feedback. An expert technique is more valuable when a beginner can recognise that something happened and has a route towards learning it.

ΔMake the response understandable

A concrete handling decision is continuity. If a player gradually changes a steering input or slide angle, we want the forces on the kart to change smoothly. We have softened abrupt boundaries in drift payout, meter income and straightening behaviour because crossing an invisible threshold could feel like the car had glitched.

We still keep deliberate events, such as a boost ending, and clear state changes. The distinction is whether a driver can understand the response as a consequence of what they did. That gives us a concrete rule for judging changes to the handling.

We have also committed to mistakes costing time while keeping the player involved. That means designing around guiding walls, recoverable runoff and continued control. A missed line should leave a recovery to drive. The challenge is to preserve a meaningful advantage for the clean line while giving the driver who missed it another decision immediately.

Progression has a similar constraint: we do not want unlocks to make the machine faster. The current vehicle choices are cosmetic. Our longer-term interest is in styles of driving that remain competitive, but those handling styles are an ambition, not a feature we have already delivered. Knowledge of a track, a better exit and a personal best are the progress we want the game to reward.

?Keep the trade-offs visible

These goals can pull against each other. Softer wall penalties keep a beginner moving, but if the wall becomes the fastest way around a turn, we have weakened the reason to learn the line. A more generous drift reward may help a first lap while removing a choice that made the hundredth lap interesting.

We cannot resolve those tensions by saying the game should be accessible and deep. We need to identify which decision a change creates or removes, then drive it. Can the player recover from this mistake? Does avoiding it still pay? Is the faster technique something they can discover and practise?

Measurements help us catch consequences we might miss in a short drive. They do not tell us which consequences we should want. A faster lap is evidence of a faster lap; it is not, by itself, evidence of a more interesting racing game.

~What we still have to earn

Our clearest unfinished problem is making the depth legible. A player needs to connect a good result to the input that caused it. If the game rewards a particular exit but gives no useful clue about why, we have built something to memorise through frustration.

We want the next attempt to have a purpose: straighten a little earlier, save the booster for the next exit, leave more room for the recovery. That is the progression we care about, a player becoming more deliberate with the same machine.

The test we keep coming back to is practical: can a new player understand enough to enjoy a lap, can an experienced player keep finding better choices, and does a well-driven corner make the next one more interesting?

Dead Air Works is an independent studio building SLIPFAULT, an arcade racing game for PC.