Exposure compensation is probably the solution.... knowing how much comp to dial in, the art.
Matrix metering on modern cameras is an awful lot smarter than the pretty crude Center Weighted Average of old Auto-Exposure systems, that were based on the presumption that 'most' scenes average out to 18% grey, hence would tend to under-expose lighter/brighter/whiter scenes or subjects... [bisto-moment] It was probably my first real lesson in photography, aged about 11, when I went skiing for the first time, and chap at the min-lab explained how to use the ASA selector to 'fool' the meter into thinking that the camera needed more light by telling it I had loaded slower film than I actually had.. ie exposure compensation! Top tip, b-u-t... took a LOT of film before I learned where/when/how much compensation was best...
Suggestions of metering off grass, or tarmac or grey-card... are an old trick to make a through-the-lens 'reflected' light meter, that is seeing the light that is reflected off the subject, record something 'like' you would get from an 'ambient' light incident meter reading taken from the light falling 'on' the subject at the subject, with a hand-held incident light meter.... and the trick is pretty reliable... b-u-t you have to know the difference between incident metering and reflected metering, and where either may be more or less useful... which leads into such techniques as spot-metering, and getting bogged down in an awful lot of technical, that modern evaluative matrix systems are attempting to mimic with expert-programming.
Essentially, an incident reading measures the light falling on the subject. Take that reading, make shutter/aperture/ISO settings based on it, and you 'should' get a pretty 'faithful' exposure, in which dark records dark, light records light, and mid-tomes in the middle, regardless of how light or dark they actually are.... b-u-t... a black cat in a coal-hole will come out very very dark.... as pretty much all of your scene is in the shadows, and very little of the light falling on the scene you point the camera at will be being reflected at the camera. Conversely; a white rabbit on a ski-slope will go the other way; same ambient light, same exposure settings, but subject will reflect a heck of a lot more light, so your picture will come out very bright. Now, taking 'reflected' light readings, at or in the camera of the actual light that's getting to it, will give you a less 'faithful' image, as the meter will offer an exposure value or suggest shutter/aperture/ISO settings, based not on how much light is falling on the scene, but being reflected.. now your black-cat-in-a-coal-hole will tend to come out grey, the meter seeing less light, so brightening up the overall exposure assuming that it should, on average be aprox 18% grey.. again conversely, your white rabbit on a ski-slope will reflect a lot more light, and the suggested exposure will ted to tone that too down to grey, on the assumption that that's what it should average out to.. back-to-bisto, that's where my first Ski-trip shots went, lots of very bright white snow being rendered down to mid-grey and people on the slopes being darkened into shadow, by meter seeing a lot more light than normal, and assuming it 'should' all be grey...... B-U-T..... turn slightly to the left, now the woods are the back-ground, not the slope, and dark woods make the meter go the other way., and white slopes start blowing....
Which begs consideration; or exposure centering, and the dynamic range.... ie not just getting an exposure thats good for the mid-tones, but one that retains shaddow detail and keeps high-lights from 'blowing' which may not be possible.. you may have to sacrifice some shadow detail to top highlights blowing out, or accept some highlights blowing out to get some shadow detail, and in either case, the mid-tones going off center a bit.... this was a 'problem' with film, but is even more so with digital that doesn't often offer the same 'latitude' that film does, to cover both ends and the middle.. but still.... point i, there s no such thing as a 'correct' exposure, just good-enough and not so good ones, and in between 'better' or 'worse' ones; it's very subjective, and down to what you thought 'best' not the meter, which is merely a guide!
Shooting rally-cars..... you don't say whether they are in action or not.... 'walking around' does imply static padock shots, rather than track-side though. In 'action' I would tend to take meter readings 'through the lens' before a car comes through the stage; those meter readings, would be TTL 'reflected' readings, but of the scene before a car slid into it, and with likely grass, mud gravel as the main scenery, mimick a grey-card reading, and/or be close enough to the metering needed when the car came through... I would be aiming to drop shutter speed and 'pan' with the vehicle to get the back-ground and wheels to show some blur, while I got the car crisp, so would probably take the reading I got as pretty indicative of the scene as a whole; pick the shutter wanted to get some panning blur, and an aperture to keep the car in the DoF zone as it came through, and 'lock' the settings there.. if it was a black car... it may come out a tad dark... if a white one, a tad bright..... but it would be reasonably 'faithful'... and I wouldn't expect there to be uch difference if I shot i Aperture or Shutter priority or even full-auto, and let the TTL meter adjust for reflected light as the car came through.... unless..... the scene was peculiarly bright or dark, and or I was framing very tight on the car.hence a large expanse of back-ground threw the metering off, or filling frame with subject, excluding that back-ground did likewise.... and we are back to the 'art' of knowing where/when/how-uh.
Walking around the parc-ferme/paddoc; panning shouldn't be an issue, car shouldn't be moving, but you cant take reflected readings off the scene without the car in it so easily, and then 'lock'. You are probably also framing tighter and filling more frame with subject, so black or white cars will likely kid the metering more into over/under exposure... you may meter off subject on a more neutral back-ground exposure lock and recompose, or guess how much exposure comp to dial in.. or you could start taking spot meter readings off the car and trying to assess the exposure range or looking at the histogram and trying to shoot to the right to avoid blowing high-lights, or or or.... faff as much as you care really!!! BUT, the answer is there is STILL no 'correvt' exposure, less still a 'perfect' one, just what you think more or less pleasing, and the meter is very very dumm.. it aint got a clue what its looking at, you do! So, if you have the luxury, Chimp it! Take a test shot, review on screen and adjust settings if you think its too dark or too bright!
There really is no one-size-fits-all answer to the question here; there are so many metering methods, incident and reflected being the two foundations, then so many more based on them, like spot metering, like CWA like 'evaluatve'; or matrix, like shooting to the right, before you translate the meter reading to exposure settings! And ALL have merit in different situations and circumstances... a-n-d to some degree, to achieve different results in the same situation/circumstance...
BUT Keep-It-Simple-Silly... you can easily disappear up your own f-stop, fretting about all the technicalities, complexities and subtleties of the matter.... remember there's no correct or perfect exposure, it comes down to what you prefer.... so f you aren't getting what you'd like, tweek it on the exposure comp dial ether way.... but, remember its subjective, and the art is in knowing where/when/how much... and just like the meter in the camera.. WE aren't looking at the scene.. YOU are, so YO have to decide how far to trust or tweek the meter's assessment.. we cant tell you.... even less, which means of tweeking is best.... Exp-Cop, Exp-Lock, off-subject metering, etc etc etc... you have to figure that out for yourself by practice, I'm afraid!