Every sim I tested runs faster than the real car. Here's the data.

I have been building a lap time database for a few months, and the point of it was never to rank simulators. It was to answer a narrower question: when you set a time on your rig, what have you actually achieved?

The answer, at least in Hypercar, is uncomfortable. At four of six circuits, the sim is quicker than the real thing.

Spa is the clearest case. A real WEC Hypercar qualifying lap sits around 2:03.1. A good Le Mans Ultimate lap in the same class comes in near 1:59.2. Nearly four seconds, in a car with the same name, on the same layout, against a stopwatch that does not care about your feelings.

What the gap is made of

None of this is a defect. Every item is something a simulator either cannot model or deliberately spares you.

Track state. You get a perfect surface at a fixed temperature. Real qualifying happens on whatever rubber the support races left behind, at whatever the ambient decided to do that afternoon.

Fuel. Most hotlaps are set light. A real qualifying lap carries what the regulations demand and the strategist allows.

Consequence. This is the big one. A sim driver can carry an extra five kph into Pouhon forty times in an evening, because the fortieth mistake costs a reset. Nobody drives like that when a mistake ends a weekend and costs seven figures.

Setup freedom. Open-setup hotlaps let you build a car that would be undriveable across a stint. It only has to survive ninety seconds.

The exception is the interesting part

Silverstone runs the other way: iRacing's GTP car is around six seconds slower than the real WEC lap. That is too large to be noise, and it points somewhere specific — most likely how the tyre model handles sustained high-speed direction change. Maggotts-Becketts-Chapel is the most demanding sequence of that kind anywhere, and if a model is going to be conservative about lateral load, that is where it will show.

So this is not "sims are optimistic". It is that every sim is wrong in its own particular direction, and the direction tells you something about the model.

What to do with this

Stop using real-world lap times as a target. If you are two seconds off the real Hypercar time at Spa, you are not close — you are six seconds off what the sim will actually give a quick driver.

Compare like with like: your time against the same sim, same car, same conditions. Everything else is a category error dressed up as a benchmark.

I will keep adding data. The numbers behind every circuit on this site are visible in the circuit cards, with the session and layout attached — because a pole lap and a race fastest lap are three seconds apart and quoting one as the other is how bad claims start.