I doubt those lowballs were your first picks for the numbers here, but nonetheless I respect that you put forth some elbow grease so I'll reciprocate by grabbing some better data. The first Duck Tour business started in 1946 [1] - they're still in business. You can read the NTSB report on the 1999 incident here [2]. The PDF (on the right side panel) has far more information. It includes information such as each year having more than 250 ducks transport more than a million passengers on tours averaging 90 minutes. It's reasonable to assume that that number has been increasing over time, but the millions in the 90s at least gives you a decent ballpark. And so like I expect you were finding with more reasonable estimates, these vehicles are hardly dangerous.

And the NTSB also did not find the boats themselves unsafe, though that they could be made to be safer in case of accidents. They stated the cause of the accident was "inadequate maintenance." The other issues, which the media is obsessing on today, were listed as contributing to the magnitude of the incident, not causes of the incident. The cause then, like now, was human error. The 1999 boat was sent out to sea with with a 4.5" hull access plug removed. The mechanic, who was new to the job, forgot to put it back in. The operator did nothing to evacuate the craft, and so on.

[1] - http://amusementtoday.com/2015/05/original-wisconsin-ducks-c...

[2] - https://www.ntsb.gov/investigations/AccidentReports/Pages/MA...

>having more than 250 ducks transport more than a million passengers

That wasn't the number for DUKWs that was the number for all amphibious passenger vehicles.

>It's reasonable to assume that that number has been increasing over time

That's really a reasonable assumption considering the vehicles are 70+ years old.

>but the millions in the 90s at least gives you a decent ballpark

Again that's all amphibious vehicles, nothing about DUKWs, but let's assume they are.

1 million passengers. The DUKW in Missouri had 30 passengers, so let's assume the average trip is half full. 1 million passengers / 15 ~= 67k trips * 1.5 hours per trip ~= 100k hours per year.

There have about 40 fatalities caused by DUKWs in the last 20 years. 100k hours * 20 years = 2 million hours.

20 deaths per million hours operated

For cars we have ~0.625 deaths per million hours operated.

32 times more deadly per hour. Remarkably close to my first estimate.

Your #1 link estimates that there are 300 operational DUKWs in the US today. That's going to include reserves that are only operating during busy seasons/weekends etc... But let's assume they are all active for 6 hours per day on average (which is way too high). You still get a fatality rate per hour about 5 times higher than cars.

And again, tour boats and tour buses are much safer than cars, so the comparison will be much worse.

>And the NTSB also did not find the boats themselves unsafe

They did. Here is their recommendation:

“Without delay, alter your amphibious passenger vessels to provide reserve buoyancy through passive means, such as watertight compartmentalization, builtin flotation, or equivalent measures, so that they will remain afloat and upright in the event of flooding, even when carrying a full complement of passengers and crew.”

“The Safety Board considers that the major consideration in assessing the ability of passengers to escape from a sinking DUKW is the overhead canopy.”

>The other issues, which the media is obsessing on today, were listed as contributing to the magnitude of the incident, not causes of the incident.

Of course they aren't the cause. A canopy isn't going to cause a boat to sink.

These vehicles have been operating since 1946. That's 72 years, not 20. The reason I gave you the hours and population is because you then figure out 'person hours' without having to make any assumptions other than average rate of travel to go from deaths/mile to deaths/hour - which I expect is a stat you can actually probably also find out.

This was the NTSB's conclusion on the accident.

"The National Transportation Safety Board determines that the probable cause of the flooding and sinking of the DUKW No.1 was a missing access plug, which, in turn, was caused by inadequate supervision of company personnel and inadequate management oversight of amphibious passenger vehicle maintenance. Contributing to the sinking was a flaw in the design of DUKWs, that is, the lack of adequate reserve buoyancy that would have allowed the vehicle to remain afloat in a flooded condition."

The accident was not caused because the vehicles were unsafe. It was caused because of human error. The things you're talking about are factors that contributed to the sinking. Without the main problem (human error) there would be no problem. And, lo and behold, human error was also the cause in this most recent incident as well.

>These vehicles have been operating since 1946. That's 72 years, not 20

I only have fatalities for the last 20 years if that wasn't clear. If you want to go back into non-digitized local newspaper archives across the country to find out if there were any other deaths as well as try to figure out how many were operating 70 years ago, be my guest.

That doesn't change that fact that over the last 20 years, DUKWs are much deadlier per hour than cars, and it's even worse when compared to buses and tour boats.

>The reason I gave you the hours and population is because you then figure out 'person hours'

Fine. 1.5 million person hours * 20 years = 30 million person hours

40 deaths / 30 million person hours = ~1.3 deaths per million person hours

The best answer I can find for how long the Average America spends in a car per day times number of drivers times days give me 199,685 million person hours per year, and 37k vehicle fatalities last year = 37k/199685 = 0.19 deaths per million person hours

About 6 times deadlier.

Tour boats and buses are much safer than cars so adjust accordingly. In the UK (only place I could find the data in hours) water transportation has 50 deaths per billion hours. 0.05 deaths per million person hours, Making DUKWs 26 times deadlier than the average UK passenger boat.

>The accident was not caused because the vehicles were unsafe. It was caused because of human error. The things you're talking about are factors that contributed to the sinking. Without the main problem (human error) there would be no problem.

This is completely meaningless. Human error is the cause of almost all traffic accidents as well, yet some cars are still unsafe because of the behavior of the car once human error takes place.

Assume that the NTSB finds that a defect in a new bus called the Explodo will trap all it's occupants inside and cause it to catch fire in the event the driver hits an obstacle at greater than 20 miles per hour. Are you going to argue that the NTSB isn't really saying the bus is unsafe because human error is necessary to trigger the defect? No you're not.

Look, I get it. You pride yourself on making evidence based judgments. However, in this case you did some quick math and misjudged the danger, but now you're digging in and sticking to an indefensible position. You're letting your emotions influence your position--behaving exactly like the journalists you were original railing against.

>If there were significant fatal accidents before 1999 - especially if the fault lay with the vehicles' design, they would have dug them up. So it's safe to assume there were none.

That's not even remotely safe to assume. You can't seriously think the LA times has the manpower to comb through 70 years of non-digitized records and local newspaper stories for 1 article.

Since they included traffic incidents after 1999, and you think they went through 70 years of data. Why no traffic incidents before 1999? The simplest explanation is that they only included what they could find using digital records that were easy to access.

> looks like you missed some decimal points.

I looked it over and the data is just wrong. It was based on self reported data, and it looks like people just over report the amount of time they spend in cars.

Since there is no way to accurately calculate person hours for cars. I'll just refer you back to the estimation based on operating hours.

Or you can compare it to the person hour safety data for passenger boats in the UK. Either way based on my estimations, DUKWs are likely statistically much more dangerous.

But who cares. I'm just some guy on the interne with very limited access to the relevant data. You don't need to take my word for it.

The NTSB found that the boats were unsafe. Their recommendation to New York and Wisconsin was that they require DUKWs to be modified or cease operations. Their recommendation was that unmodified DUKWs are too dangerous to continue carrying passengers.

>Your explodo example is not really reasonable.

So you agree that safety flaws that magnify human error can make a vehicle unsafe? Because it seemed like you were arguing the opposite for the last several posts.

>And they've also been in operation for millions of hours. There's way more than enough of a sample for lots of things to have gone wrong. Yet of all the incidents that have happened in these millions of hours, and the likely 40,000,000+ people that have used these vehicles, we only had 2 severe accidents - both caused by extreme human incompetence. So you're left with two choices, neither of which I expect you'd really like:

This entire paragraph is an appeal to emotion. You've disguised it by throwing in some very dubious numbers that you have no way of verifying. Your argument essentially boils down to: We've had lots of people ride our boats and only a couple really bad accidents--that means they're safe.

The rational way to judge their safety is to compare the fatality rate to other vehicles, and in the absence of the ability to accurately do this, defer to a trusted expert. But you just want to focus on what feels right to you.

The NTSB thinks that they are dangerous, back of the envelope math shows they are probably more dangerous than other vehicles. Just admit that your gut reaction was wrong in this one case.

40 million assumes that in the 50 years prior to 1999 average usage was less than 400k per year. Going from 0 in 1946 to 1 million in 1999 with nothing but a linear increase would of course be 500k per year on average. My number was more than 20% below that. It was intentionally an extremely conservative estimate. I do not believe that you believe that organizations such as the LA Times and NY Times are incapable of digging up incident reports in the decades before 1999 - that's cognitive dissonance.

And on that note, I feel as though you're projecting here. The numbers we've come up for here look okayish even when you are making crazy assumptions like pretending these things never existed before 1999, or were using some rather creative math to get that person hours estimate. Reality is rather friendlier, meaning these numbers are going to go from "okayish" to something substantially better. The reason we were interested in the rates was not to prove they are safer than cars, but to show that when you look at the incident rates that these calls for bans on them are grossly sensationalistic and emotional. I'm not sure what level of accident rate does make one start pondering a ban, but we're obviously not even in the remotely right ballpark for that sort of discussion.

The NTSB investigates literally thousands of incidents, offers probable cause and solutions where available. The presence of solutions that could lessen the dangers of catastrophic failure is not the same thing as the boats themselves being unsafe. For instance, if I were to locate a report from the NTSB proposing some form of vehicular regulation or another, would you then insist that all vehicles are unsafe, simply for the fact that they could be made more safe? I would not have to search long, as you can see from their reports [1] yourself.

[1] - https://www.ntsb.gov/investigations/AccidentReports/Pages/Ac...

> Going from 0 in 1946 to 1 million in 1999 with nothing but a linear increase would of course be 500k per year on average. My number was more than 20% below that. It was intentionally an extremely conservative estimate.

Here's another example of why these extrapolations are useless without real data.

1. There is absolutely nothing that would allow you to say with any confidence whatsoever what the growth curve looked like. Maybe it was quadratic and there were only a million from 1946 to 1998. From looking at the history of the original Wisconsin Ducks, it looks like they operated with only a handful of DUKWs for decades. Maybe they stayed small until bigger companies got involved and spread to multiple cities in only a few years before 1999? You don't know, which is why trying to use numbers before 1999 is useless.

Exponential growth is a very likely explanation for the appearance of an increase in fatalities over the last 20 years (Vehicle aging and easier access to modern records are others)

2. The 1 million number you keep quoting was for all amphibious tourist vehicles, not only DUKWs. There were other amphibious vehicles in operation in 1999, and there are more now. There is no breakdown for the number of riders by vehicle type. Given that newer purpose built vehicles are more expensive, they are more likely to be used in high traffic areas thus they are likely to carry a higher than expected proportion of the total riders.

3. These vehicles are now over 70 years old. It is entirely possible that there are fewer of them in operation than there were in 1999.

>I do not believe that you believe that organizations such as the LA Times and NY Times are incapable of digging up incident reports in the decades before 1999 - that's cognitive dissonance.

Do you think there were zero safety incidents involving DUKWs before 1999? You stated they included recent traffic incidents. If they were going back through 70 years of records, why didn't they include any previous traffic incidents? The most likely explanation is that they found the low hanging fruit and didn't bother going back any further.

There were 4 days between the incident and the report you mention. What kind of world are you living in where a newspaper has the resources to comb through 70 years of local news reports and accident reports from every relevant municipality for 1 article in 4 under days.

>crazy assumptions like pretending these things never existed before 1999

This is absurd. We clearly lack data to make accurate estimations from before 1999. We can try to look at the fatality rate over the last 20 years, or we can spend time arguing over why we can't find 70 year old accident reports from a small hamlet in Wisconsin over the internet.

> Reality is rather friendlier, meaning these numbers are going to go from "okayish" to something substantially better...I'm not sure what level of accident rate does make one start pondering a ban, but we're obviously not even in the remotely right ballpark for that sort of discussion.

I think that 26 times deadlier per hour than UK passenger boats warrants looking into it, but I'm not the expert. The experts issued a report though. And it recommended banning unmodified DUKWs.

>grossly sensationalistic and emotional

I think you're the one projecting here. The NTSB called for a ban on unmodified vehicles. Newspaper reporters parroting the NTSB experts can hardly be called sensational.

>more safe

Your misreading of the NTSB report is incredibly frustrating. The NTSB issued a recommendation that operators immediately make heavy modifications. They then made recommendations to state governments that they ban unmodified vehicles. If you read the report they sent communications back and forth with dozens of individual operators begging them to make these modifications.

Here's an excerpt: "THE AMPHIBIOUS VESSELS ARE SUBJECT TO SINKING AND A TRAGEDY LIKE THE ONE THAT OCCURRED IN ARKANSAS COULD BE REPEATED. ACCORDINGLY, THE SAFETY BOARD REQUESTS THAT THE WISCONSIN DUCKS RECONSIDER THEIR POSITION ON THIS MATTER AND CONSIDER TAKING THE REQUESTED ACTION. PENDING FURTHER REPLY FROM THE WISCONSIN DUCKS, M-00-5 HAS BEEN CLASSIFIED "OPEN--UNACCEPTABLE RESPONSE." IN THE INTEREST OF ACCURACY, YOUR LETTER REFLECTS PERCEPTIONS OF THE EXPERIENCE LEVEL OF OUR MARINE STAFF AND SAFETY BOARD - U.S. COAST GUARD RELATIONS THAT ARE NOT FACTUAL. SHOULD WISCONSIN DUCKS WISH TO DISCUSS THESE ISSUES, MY MARINE STAFF CAN BE CONTACTED AT (202) 314-6450."

Nothing about the report or subsequent communication is indicative of an opinion that the boats are safe but that they could be made "more safe." A reasonable, rational person would read the NTSB be report as a proclamation that unmodified DUKWs are unsafe.

The mental gymnastics you have to go through to read that report and take from it that these vehicles are safe for passengers is truly, spectacularly, amazing.

Try to step back and objectively consider the logic of what you're saying. You are literally trying to suggest that we went from minimal usage to more than a million passengers per year in 1999 and then back to minimal usage. This is possible (well, not really - but I'll get to that in a minute), but beyond improbable and seems to be based on absolutely no logic other than 'this is what I need to try to maintain my view, so it must be reasonable.' By contrast the assumption of normal growth is based not only.. well normalcy.. but also the fact that the original operator stayed in business. Tour businesses are costly and fickle mistresses. To keep going for 70 years strongly indicates at least normal growth.

I don't say that to mock you or your latest suggestion, but because I think you might genuinely be unaware you're doing this. I'm certain you think your view is correct and naturally therefore assume that there must be logic, and thus evidence, behind it. You're not considering the possibility that your view may be incorrect in this case.

And this is causing you to make more silly mistakes. For instance you are now proposing that perhaps the number of ducks in operation declined since 1999. As you mentioned, from the writeup on the original duckboat operator, there are about 300 ducks are in operation today. That growth ranges from a substantial 20% if we assume every one of the 250 vehicles the NTSB was referred to as 'amphibious vehicle' was a duck. If you genuinely believe that to not be the case (and some of those 250 were something else), then the growth since 1999 could potentially be vastly larger.

The NTSB did not call for a ban on ducks. They even specifically suggested an alternative list of (much cheaper and easier) criteria states could require in lieu of providing the additional reserve buoyancy. In fact, check out page 54 of the report. There was infighting within the agency even about the relatively minimal 'alternative' suggestions they put forth. Internally it was felt that the additional buoyancy requirement was unable to be able to be effectively implemented, and that the canopies were not a danger. In particular, some of the NTSB staff felt that basic requirements (double clamps on drive shaft boots, additional bilge alarms, etc) would be sufficient to prevent this sort of accident in the future. You're conflating media sensationalism with the quite level headed and reasonable report.

>You are literally trying to suggest that we went from minimal usage to more than a million passengers per year in 1999 and then back to minimal usage

In the calculations I made I maintained a steady 1 million per year. Your emotions are causing you to misread what I'm writing.

>but beyond improbable and seems to be based on absolutely no logic other than 'this is what I need to try to maintain my view, so it must be reasonable.'

Slow steady growth in existing markets with bursts as they moved into new markets approximating an exponential curve, followed by low growth once all potential markets were saturated and the supply of surplus military vehicles dried up is "beyond improbable"?

I'm not making a strong claim here. I'm saying that such a growth curve is possible and not improbable.

>By contrast the assumption of normal growth is based not only.. well normalcy

Linear growth isn’t a safe default assumption. Rapid growth followed by a plateau is “normal”.

>You're not considering the possibility that your view may be incorrect in this case.

I’m almost certain you’re not actually reading what I’m writing. I’ve said numerous times that I can’t accurately estimate how many people rode DUKWs before 1999. I’m not attempting to provide accurate estimates for that time period. There could have been 50 million passengers before 1999, and there could have been less than a million. There could have been 0 fatalities before 1999 and there could have been a dozen. For both of those numbers the answer is almost certainly somewhere between those 2 extremes. Where we do not know because we don’t have the data.

We do know how many fatalities there have been since 1999, and we can make a better estimate of how many people have ridden them. We’re still not going to get very accurate data. But the data we have shows that it is likely that DUKWs over the last 20 years are substantially less safe than other boats.

>For instance you are now proposing that perhaps the number of ducks in operation declined since 1999. As you mentioned, from the writeup on the original duckboat operator, there are about 300 ducks are in operation today.

We’re talking about a blurb on a website from a tour boat operator. We don’t know what his definition of Duck is. Other amphibious passenger vehicles look very similar, and some DUKWs have been very heavily modified. I’d say that 300 is fairly safe to use as an upper limit--that’s it. Even then, take it with a grain of salt.

>The NTSB did not call for a ban on ducks.

Nope they didn't. I never said they did. They called for a ban on unmodified DUKWs..

“Require that amphibious passenger vehicle operators provide reserve buoyancy through passive means, such as watertight compartmentalization, built-in flotation, or equivalent measures, so that the vehicles will remain afloat and upright in the event of flooding, even when carrying a full complement of passengers and crew.”

“Until such time that owners provide sufficient reserve buoyancy in their amphibious passenger vehicles so that they will remain upright and afloat in a fully flooded condition (by M-02-1), require the following:...”

The second recommendation isn’t an alternative. It is a stopgap solution to be implemented immediately. Neither recommendation was ever required, and only a few operators complied with the request.

>Internally it was felt that the additional buoyancy requirement was unable to be able to be effectively implemented,

The report and the official recommendation were adopted as is though. So you only think we should take the recommendations of committees when the result is unanimous?

>and that the canopies were not a danger.

Just stop. It’s only you and me, you don’t need to lie about the report.

3 staff members asked that the report be changed to “canopies represent a grave risk to passenger safety.”

By the way. This

>199.685 trillion person hours would give an average base rate of travel of 0.016 miles per hour. Obviously you need to multiply that by the average number of occupants per vehicle, but even if you say there are 10 occupants you're looking at 0.16 miles per hour on average so it looks like you missed some decimal points.

Was completely wrong. I was on my phone and didn’t bother double checking the math. I assumed the study I found with average daily time spent in a car was wrong because it was self reported.

You’re actually the one who missed 3 decimal places. I said 199,685 million person hours. That is 200 billion miles not 200 trillion as you said. With the correct number, it’s an entirely reasonable estimate.