Manifold

Ultimately, there's no way to verify the 'accuracy' of the analysis of a particular incident with respect to what was going on in the driver's mind, but if the analysis is ballpark accurate, that may be enough to make the analysis useful in preventing future incidents, and the very act of analyzing an incident and having the driver be part of the analysis may be sufficient to make a positive difference. You just need to be careful in the analysis to not direct too much attention to secondary factors while giving too little attention to primary factors. If you anchor the analysis to objective data like braking points, entry speed, etc., that should help reduce the need to do mindreading.

Overall, I think that what you're doing is a good thing, and I encourage you to keep it up!

Dionvm24

Data is a big emphasis for us and I don’t mean actual data logger systems, but you can use multiple laps to be as exact and specific as possible. That’s true for all of our coaching services and what makes video such a strong learning tool.

I appreciate that!! We’re really excited about it!

ProCoach

As someone with decades of experience performing forensic analysis after the fact using video and data, it's important to realize, as DvM does, that you look at several measures to better validate what the decision making was that went into the driver's mind. Multiple measures allow a much more, nearly full context of what happened, and there is no set "set" of measures needed to do this well.

You suggest that one cannot divine what a driver is thinking, but I can tell you that there are some prevalent behaviors common among those vulnerable to make questionable decisions. With pretty predictible outcomes, too.

DvM is doing a fine job, and he is accumulating experience at a great pace.

Manifold

I agree that, when the analyst has experience and a range of info to triangulate from, the analyst will often get it at least approximately right. As I said, I think DvM is on the right track and should keep it up. We just have to recognize that there's always some uncertainty and speculation involved in this type of analysis, so a degree of humility is warranted.

I was involved in a forensic investigation not long ago that I spent over 1,000 hours on and the failure consequences were over $1 billion. I interviewed many dozens of people at length and wanted to understand what had been going through people's minds, but I saw clearly the limits in being able to do that. And research in this field generally shows that our confidence that we got it right often isn't a very good indicator of whether we got it right.

RobertR1

The context isn't similar.

The racer is driving on habit and instincts (learned behavior) which is easier to detect through patterns and repetitions. There's a much higher success percentage of root cause analysis in the driver scenario given a large data set of laps being limited to 1 individual who is likely very cooperative in asking for guidance towards remediation.

You're attempting to uncover intent behind an anomalous event. This is where the similarities end as in your scenario cooperation and transparency might be lacking due to potential ramifications. You also mentioned "people" which if accurate assume a series of actions across numerous individuals where you're set to determine at which point there was a deviation from the norm and then the intent element behind it.

Manifold

In that forensic investigation, I looked not only at the overall incident and the events leading to it, but also how particular individuals made key decisions. I've also done many forensic investigations of track incidents, including my own incidents (never crashed, but have gone off a few times). The challenge of trying to infer what an individual was thinking and feeling (why they did what they did) is similar across domains.

Let's consider the scenario of a driver who brakes too late, carries too much entry speed, goes off at the exit, and crashes. Was he getting desperate because of his race position and trying to step it up too much? Was he not desperate, but just decided to try stepping it up and see if he could gain some time? Was he distracted because at the moment he was too focused on checking the status of other cars? Was he distracted by something else? Was he getting physically or mentally tired? Did he have a momentary lapse of attention? Other factors I didn't mention? A combination of these factors?

You can interview the driver and ask him what was going through his mind, but he'll be working from memory, and memory is known to be somewhat unreliable; moreover, we don't really know what's going through our subconscious minds and tend to confabulate about that. You can try to make inferences from patterns and repetitions, and deviations from them, but there's still some speculation involved, and confidence that you got it right doesn't mean you got it right. Again, I'm not saying that this type of incident analysis is worthless, I'm just saying that we should recognize the speculation involved and not be overconfident about our ability to read the driver's mind (by contrast, we can often be very confident about our understanding of the physics of what the car did and the driver's observable behavior).

For those who're interested in delving further into these cognitive aspects, I highly recommend the book "Strangers to Ourselves: Discovering the Adaptive Unconscious" by Timothy Wilson.

esscape26

Education is always good and the idea is innovative and to be commended, but I have been in the tower for about 20 PCA steward interactions with drivers on incidents over my years of volunteering with PCA. While I cannot be sure what the final decisions were, there was little consistency in what I heard being espoused and meted out during the conversations. So I am not a fan of adding another layer of "education" on top of what might be poor decisions by stewards.

(NASA has a different philisophy on handling and meting out decisions BTW. So perhaps that platform or others is a better platform for this service to reduce penalties, at least at this time. Or save it for obvious at-fault incidents. Or perhaps some other form of education that serves to continue and advance learning proactively, without assigning fault.)

I have heard (again can't confirm but I trust my sources) that at least one racer involved in a single-car incident has borrowed or rented another car only to cause another single-car incident during a weekend, making races have unnecessary yellow laps. Now, these racers don't get any penalties, but they certainly could use some education, too, one would think.

Steve113

Its not as hard to analyze as people think .

I can narrow it down very fast :

60% turned in to early ran out of road on exit - hit wall on outside or yanked it back on track hooked up hit inner wall (inner wall =biggest hits)
30% carried to much speed at entrance lost it mid corner spun off and hit or some how hit inner wall (some reports of a red type mist in helmet before crash)
2% mechanical related
4% weather related hydroplane
1%- only God and the driver will ever figure this out no logical explanation (maybe an Alien thing-I never believed before but in 8 years of doing DE insurance I might believe)
3% - open for analyzing

Just my 2 cents

Manifold

I think pretty much everyone agrees that we can usually accurately analyze what the car and driver physically did. Being able to accurately analyze what was going through the driver's mind is where we may have some disagreement.

RobertR1

The fundamental aspect you're trying to piecemeal in the driver scenario isn't intent but reason.

Don't get me wrong, intent serves the purpose of building out the scenario, as you said, were they fighting for position? distracted? and so on.

So now you have a situation where the vehicle has performed an unintended change in direction and has created a loss of control scenario. We're not into the granularity of determing the reasoning of their inputs that created the end result.

Let's also assume the driver is capable thus generally aware if the car is understeering/oversteering and other such feedback mechanisms. We get extra lucky and the driver has telemetry from the incident. Now we can align the data points and actually ask the driver why they didn't take the appropriate measure for correction. Make them think. Did they think staying on the gas while you have exit understeer was going to fix itself? why didn't they open up the wheel sooner when the car started over rotating? did they forget that they were offline the corner before and might have dirty tires?

The intent can help build the grand scenario that lead to the incident but what's the reason they let the loss of control take place at that point in time? Where I agree with you on speculation is how do we believe what the driver is saying? even if their cooperation is beneficial to themselves, it still stands to reason that things like ego, or confusion or even a general lack of awareness could lead to speculation on the part of the driver as to their reasoning.

This is where the person performing the analysis needs to have the necessary experience to sniff it out. Like a good seasoned detective, they will need to have the experience across a wide array of individuals and incidents to determine the accuracy of the reasoning provided. After all the goal is to identify the reason (taking into account the intent of the overall scenario) and build a remediation plan to avoid such in the future.

Thanks for the book recommendation. I love learning about such topics so already used my audible credit on it I enjoy these conversations so don't think I'm trying to win some war of words here or anything.

fatbillybob

I just come off as a "hater." Intuitively, I find this "ed after incidents" flawed. Your two posts are a cogent voice of reason. Looking at incidents is not a bad thing but we have to take conclusions with a grain of salt.

fatbillybob

One thing not considered is that being an involved party is punishment enough. No one wants the time and expense of hitting a wall. A few points on a license or mandated school just adds insult to injury. We all want to learn but we need need to be receptive to it too and educators what to teach a receptive student. We also don't want to wreck again so the conscientious will explore the fault that caused the problem. The bad people are going to be bad until you take their license away. I think SCCA has it right. I have been "collected" in two wrecks. Personally, my car is still damaged even if blame is assigned. So assigning blame just isn't worth my time or energy. When no one protests the wreck SCCA says out of it. I'm asked if I am physically OK then if I'm OK with the other driver. If I make a stink they act and investigate. If I just let it go everyone goes home to lick their own wounds.

Finally, this post puts in the other dynamic the stewarts decisions which can be right or wrong. I remember a hotly contested Ferrari Challenge wreck where a driver was tried convicted and hung for video that clearly appeared that the turned into the other driver. However, his steering trace showed virtually zero steering angle. I guess it is hard to overlook what the eye sees even when data tells you otherwise.