Evolution models

01The number in the rule

The Group B regulations required two hundred road-legal examples to be built and homologated before a car could compete. That threshold was deliberately low — low enough to admit serious manufacturers but too low for mass production — and it defined the character of the class from the start. What it did not do, on its own, was produce the cars that actually won.

Buried in Appendix J was a second provision: once the two hundred were certified, a manufacturer could homologate an évolution — an evolution variant — by building just twenty additional examples. Those twenty cars had to be offered for sale, had to carry a recognisable relationship to the base model, and had to be documented on a revised homologation fiche. Beyond that, almost anything was open.

The practical consequence was decisive. The production run defined what the car looked like to a customs officer; the evolution run defined what it looked like on a stage. Two hundred cars set the platform. Twenty cars set the performance ceiling — and manufacturers discovered very quickly that the ceiling was effectively the sky.

02What twenty cars could change

The fiche governed every measurable attribute of the car. Wheel arch width, suspension geometry, engine displacement, forced induction arrangement, weight distribution — all of it had to match the filed paperwork. An evolution submission was a new fiche, which meant a manufacturer could change almost any of those attributes provided twenty physical examples confirmed the specification.

Audi Sport used this route to develop the Quattro through its Sport Quattro variant: a shortened wheelbase, a revised five-cylinder engine, wider bodywork. The Sport Quattro was a genuine road car in the same sense the original was — you could buy one — but its proportions and aerodynamic treatment were driven entirely by stage requirements. The homologation form had done its job; the question was only what the engineers were willing to put on it.

Lancia went further. The Delta S4, developed at Abarth in Turin, used a twin-forced-induction layout — a supercharger for instant low-rev response and a turbocharger for sustained power — that was extremely rare in production cars. The base S4 was the two-hundred-car requirement; the evolution variants pushed the power output and aerodynamic package to a point where calling the result a derivative of a road car required a charitable reading of the rule. The FIA filed it and the car raced.

Those twenty cars had to be offered for sale, had to carry a recognisable relationship to the base model, and had to be documented on a revised homologation fiche.

Peugeot Talbot Sport adopted the same logic with the 205 T16. The road version shared its name and its skin with a front-wheel-drive shopping hatchback. The racing version was a mid-engined, four-wheel-drive machine whose mechanical relationship to that hatchback was essentially the badge on the nose. The E2 evolution stretched the concept to its geometric limit: the bodywork grew, the power climbed, and the weight came down to whatever the minimum permitted.

03The problem twenty cars created

The evolution provision did exactly what its authors intended: it allowed genuine development, kept manufacturers competitive across a multi-year homologation period, and prevented a class from being decided entirely by who had been first to file. What it had not anticipated was the production minimum acting as a one-time gateway rather than a continuing constraint.

Once the two hundred were sold, the manufacturer had permanent access to the evolution route. A car that had been homologated as a production-based machine could be developed, iteration by iteration across successive twenty-car filings, into something that a scrutineer from 1981 would not have recognised as belonging to the same class. By 1985 and 1986 the top Group B cars were effectively prototypes with nominally road-legal siblings — fast enough, light enough and aerodynamically complex enough that the safety infrastructure of public-road rallying could not contain them. The Tour de Corse in 1986 made that visible in the worst way.

The class was cancelled at the end of that season. The evolution provision was not what killed it, but it was the mechanism through which a rulebook written to admit sporting road cars ended up fielding machines for which that description had become purely administrative.