If you ride a mountain bike, you know how quickly you can go from being a downhill wannabe hero to a heap of dust-covered lycra groaning next to the trail. A 50 km/h crash on a bicycle will be a biggie, and the next destination is usually the clinic that happens to be closest to where you're riding.

My stand-out road bike crash was in the 94.7 Cycle Challenge, circa 2005, and the F-bunch pile-up became something of a talking point that year. We were southbound on the M1, heading downhill towards the Grayston Drive off-ramp. The bunch was closed up tighter than a Scotsman's wallet, and everyone was in the drops, so we were possibly cruising at 60 km/h-plus.

Suddenly, something went pear-shaped in the middle of the peleton and the image embedded (pardon the pun) in my memory is of the hood of my right-hand brake lever disappearing into the left glute of the rider directly ahead. And the next thing I remember after that was sitting up and surveying the carnage.

I grabbed two drinks bottles at random as they rolled towards me, remounted, and eventually made the finish. I'm not sure about the guy whose glute absorbed the initial impact, which also flung me towards the side of the road and out of serious harm's way.

I've had some decent car crashes too; on the road, track, and at the Gerotek test facility. The common thread is that it happens fast. The national speed limit isn't a lot when you're heading in the direction you planned, but 120 km/h is scary when you're en route to Destination F**ked (to quote a popular Aussie YouTuber). Remember, 120 is 33,33 metres per second, or to put it another way, the length of a rugby field in 3 seconds.

My point is that there isn't always time for perfect decision-making at riding speeds, never mind at 240 km/h in a single-seater. So it amuses me when armchair experts and keyboard warriors make righteous calls on who did what and when.

Do they think an experienced racer will deliberately "take out" another driver at those speeds – remembering that 200 km/h is more than 50 metres a second? For starters, orchestrating what transpires metre by metre – or as suggested in the headline; millimetre by millimetre and millisecond by millisecond – is nigh-impossible. The likelihood of scoring an own goal is huge, as are the chances of getting hurt.

Do yourself a favour and visualise Brian Habana going from the halfway line to the tryline in one second. Now imagine him doing it with Usain Bolt at his shoulder and the two of them aiming to reach not just the tryline, but a drinks table with a champagne glass on top of it. The victor is the one who gets there first, collects the glass without spilling a drop, and then runs back to the halfway line whilst drinking the contents.

Opportunities to be tripped up are, literally and figuratively, plentiful.

And so, on to Copse and that crash. Lewis Hamilton created an opportunity – he had set it up for several corners. He went for it. It didn't work out, and while some say it was 50/50 – in other words, a racing incident - the Stewards disagreed and penalised the Mercedes driver. Maybe they saw it as a 51/49 scenario, maybe 55/45, but based on what seems a reasonably lenient penalty, it wasn't the slam dunk that some "commentators" think it was.

One thing is sure: the Stewards had access to much more information than the armchair experts. They may have examined things like GPS speed traces, GPS position of one car relative to the other over time, steering angle, throttle position, brake pressure and who knows what else – access to this sort of technology has removed guesswork and conjecture.

Their decision is the one that matters.

It is no wonder that several Porsche Carrera Cup championships use the VBOX Video HD2 as a mechanism to inform decisions on driving standards. The benefit of an independent source of information is obvious: you can't claim to be trying to avoid contact when the data indicates that you were turning towards your rival and the throttle opening was increasing.

The VBOX units are in a sealed enclosure to which the organisers hold the key. Information saved to an SD card would be examined in VBOX's Circuit Tools software, essentially replaying the incident in minute detail, aided by syncronised video. Rolling the data back to the previous corner might also reveal a back-story, and maybe it was a revenge attack. On the other hand, it could be an honest mistake – drivers are human, after all.

Ending this back on the Hamilton/Max Verstappen debate, two things strike me as significant: Lewis has made his intentions to race as hard as possible crystal clear, and that data can help track officials make objective decisions.

Officials could have a busy second half of the season.

Wesbank Modified racing cars, South Africa
In an age when there wasn't much data, justice wasn't always served

Audi endurance race cars at Le Mans
Modern race cars make hiding errors and indiscretions difficult. It is a bit like rugby. Why stand on someone if, in likelihood, the world is watching?


If there’s a common refrain from “newbie” track drivers, says Adrian Burford, it is that their car understeers.

And that’s not just from drivers of powerful front-wheel-drive hatches – it is from everything. I’ve said it before – if you don’t set up the car to point at the apex (and then hit it), you’ll never exit the corner under maximum power and thus maintain the highest possible average speed for the length of the following straight.

The worst kind of understeer is the kind where you’re forced to get out of the throttle post-apex because the nose is running off the road. Then all you can do is wait for the lateral grip to come back, which eventually allows you to reapply power with the tyres’ trajectory now heading along rather than across the tarmac. Unfortunately, the fact that you’re still using a chunk of the grip to control the lateral forces way after the apex suggests you got it wrong at the start of the turning phase.

There are plenty of signals to identify this with a VBOX datalogger but you can also see it clearly spectating from the inside of the turn. It’ll be the car that is still leaning when the corner is long gone, because the driver has turned it into a never-ending arc, rather than executing the majority of the turn earlier. Achieving this is crucial because an earlier, decisive steering input will also mean at a lower speed, when the tyres will react more willingly to a change of dynamic state.

The other way of spotting the understeer is to try to look at the driver: he’ll be the one still wrestling with the wheel when hands should be back at the 10 and 2 position.

You’d think driver-induced oversteer would be more common. Not necessarily. That’s partly because so much of the road car market is now FWD. Stability control systems have also done much to quell unruly RWD oversteer and most control systems won’t let you add power if there’s excessive steering lock applied – or if it isn’t being simultaneously reduced as throttle angle increases.

To be clear: we aren’t talking about cars that have handling issues. We’re talking about driver technique issues. There are patterns here. This kind of understeer is definitely most apparent when drivers are consciously trying to “be fast”. They’re trying too hard.

I often use a golfing analogy when trying to explain this. I haven’t hit a white ball for decades but I recall often being at my best (a relative term) when I’d gone out there without too many expectations and merely planned on just smacking a few balls about.

Unfortunately, after making decent work of the first few holes, expectation reared its ugly head and that’s when things went wrong. Overthinking it is a great reminder why Golf is such a mental game. Driving fast can be the same.

Once again, I’m going to use what we do at the AMG Academy as an example. Those tentative early sessions are usually quite predictable, and we see the speed build. Then the driver starts trying to race. They consciously hurry the process and it usually manifests in arriving at a turn too fast (having not braked decisively enough and/or at the right point).

Turning in while still braking is a sure-fire way of overloading the outside front tyre and it doesn’t take much to end up beyond its total grip envelope, even momentarily. Momentarily is enough, and that’s when the understeer starts, shifting the nose outward by a foot or two, while any chance of hitting the apex goes abegging. So now you must wait for the fronts to recover, and then coax the nose back in again.

This phase is clearly visible on a VBOX speed trace – there’s a flat line, often until way past the apex. The speed curve ends up being shaped like an L and not like a V – which is the signature letter you should get if you bring the speed down to the correct level, unsqueeze the brake, turn in, and start accelerating again almost immediately. Remembering, of course, that the throttle is an analogue instrument and not binary – good drivers use the whole spectrum of the loud pedal, from 1 to 100 percent and everything inbetween.

Simply going in too fast has a similar result and normally needs great armfuls of lock and the use of lateral G to reduce longitudinal G (if you get my drift) before going anywhere near the right hand pedal. It is remarkable what going in just four or five km/h too fast at Zwartkops’ hairpin does – it is enough to push the tyre over the limit and make the fronts ”blur” sideways.

The longitudinal G trace is a great tool for reinforcing all this when back in the “lab” after the session. Just where and when does that signal go into positive territory after braking and turning? And does it stay positive? And does the positive value increase steadily?

Looking at laps driven by Academy Top Gun, Clint Weston, this tends to happen sooner rather than later and when overlaying teacher and student at the slower corners (Turn Eight onto the pit straight is also good for illustrating this) there is invariably a Eureka moment. When we also have a throttle position graph available, it is another useful way of conveying the same message, and not only about where the throttle is first applied but also the rate of application.

Of course, knowing what you’re doing wrong isn’t a guarantee that you’ll magically fix it from the next session, but it’s a start. Patience is a virtue, especially exiting slow corners. It is another reason why the calm, progressive guys are also invariably the fast guys – and the slow guys keep blaming the understeer on the machinery.


I’ve coached many drivers who are fast driving away from the apex, and able to reach the exit point of a curve at a good speed (ie, using most of the available longitudinal/lateral grip). But the drivers who are fast up to the apex are less plentiful.

That’s because transitioning from the initial application of the brake pedal until that point where longitudinal G force changes from a negative value to a positive one is the zone where the great drivers hang out.

There’s no doubt that it takes much more skill and finesse to brake well, as opposed to accelerate well.

Over my years of coaching with VBOX dataloggers I’ve come to realise there are four distinct (but paradoxically, with blurred edges) phases in the braking process. I call them Preload, The Big Squeeze, Modulation, and Unsqueeze. I’ve been lucky to have access to data from Clint Weston of the AMG Driver Academy and observing what he does and how he does it at the wheel of an AMG GT has been educational. However, it has only been possible to make accurate assessments thanks to the VBOX HD2 supplying me with G forces (calculated from GPS) and throttle/brake position (this information being extracted from the vehicle’s CANbus).

Let’s dissect them.

· Preload. These nanoseconds are where you’ve moved from accelerator to the brake and taken up the slack. The duration where there is neither brake nor throttle being applied is as little as three-hundredths of a second in Clint’s case, when braking for Turn Two at Zwartkops. The slack in question is in the pedal but also the rest of the car. When talking stock road cars, it includes every rubber bush in the suspension and steering. It is, ultimately, a heads-up to the various mechanicals and systems that a big change is coming to the dynamic state of the vehicle. On the pedal pressure data, you can usually see this as a little up-tick when the pedal position goes from nothing to about five percent - and then very rapidly into The Big Squeeze.

· The Big Squeeze. This is where 90 percent of the action happens. The name sums it up. When I look at the negative G curve and pedal pressure data of a skilled driver, this reveals a rapid build up to maximum forces – in other words, they create a near-vertical G-force line, with the pads squeezed hard against the discs. It is still a squeeze and not a stomp though – compressing the suspension as evenly as possible is still an important goal in the first part of the process. You want the car to hunker down against the tarmac so that each brake can deliver maximum contribution.

· Modulation. Some way into The Big Squeeze you’ll see the G-force start to level off- or even reduce slightly. This is the driver controlling his input as the car slows and the risk of a lock-up or ABS intervention increases. It is a subtle process but you can see it clearly when you overlay Clint and Joe Average. You’ll see a high plateau in the former instance, with an occasional dip. Lesser drivers often have a series of jagged spires, suggesting multiple ABS interventions. Also, the phasing of the process differs, with Clint doing the hardest braking earlier in the entire process.

· The Unsqueeze This phase is critical. More commonly known as trail braking, this is the release of the pedal in a controlled manner up to the point where the foot can be transferred to the gas. The objective is to keep the front end of the car sufficiently loaded to aid steering grip, the first, subtle movement of the steering already having been made. With the car starting to pivot/swivel around its front axle (aided by the rear being slightly unloaded), the final release of the brake pedal can be made. Get the "Unsqueeze" wrong and you’ll overload the front tyres at some point on your way to the inside of the corner, creating an initial understeer and moving the car away from the target enough to make hitting the apex impossible. Miss the apex and it is virtually impossible to get on the gas soon enough and hard enough to get perfect drive onto the next straight.

Further around the lap, the process starts again but it isn’t identical. Turn Four at Zwartkops, with an apex speed of 110/115 km/h - versus just on 60 km/h for the hairpin - is a momentum corner and the G-force curves Clint creates are very different - understandably so, when you consider they have little in common other than being right-handers.

What he does at two totally different corners has convinced me that braking is of the essence and that it becomes even more of a delicate art when driving a traditional manual, and a car that doesn't have ABS. I'm keen to get the VBOX onto something like a Backdraft Cobra with a Reghard Roets at the wheel, or a Lotus Challenge car with Jeffrey Kruger driving and take a close look at their data when they're working the middle pedal...