Затонушие корабли

[leblon]

Правильный ответ 20 000. Я рассуждал так. Начитавшись Патрика О'Брайена, понимаешь, что моряцкое дело очень ненадежное, даже при самом умелом подходе. Сотни или даже тысячи кораблей находятся в море каждый день, неужели раз в неделю ничего не утонет? 50 кораблей в год, умножить на 200 лет активного мореплавания, получается 10 000.  (Почему раз в месяц или 50 за год? Можно еще так рассуждать: десяток крупных портов, как минимум. Несколько крупных штормов в год, Если в каждом порту во время крупного шторма тонет хотя бы один корабль, то получится несколько десятков кораблей в год). Ну, и кроме того, во время войн топят почти каждый день, так что 10,000 это оценка снизу. Наоример, во время Второй Мировой немецкие и японские подлодки сильно бесчинствовали у берегов США. Так что в любом случае, 2000 это слишком мало, а 20,000 может чуть много, но по порядку величины вроде разумно.

Comments

[leblon.livejournal.com]

Раз уж нашли ссылку, поделитесь, не мучайте :) Я в машине радио слушал, деталей уже не помню.

[shkrobius.livejournal.com]

From this NOAA report
http://sanctuaries.noaa.gov/protect/ppw/
RULET database counts items of "sunken material" rather than shipwrecks. Somewhere along the way it became "wrecks" then "shipwrecks". Actually, their own site claims 30,000 such items, so even this number 20,000 was not updated.

[leblon.livejournal.com]

Зря придираетесь, по-моему. В этом отчете NOAA написано, что в базе данных затонувшего материала 30,000 targets из которых 20,000 vessels. См. подпись к самой первой карте ES-1.

[shkrobius.livejournal.com]

Ships = vessels that have steel or concrete hull and exceed 150gt in tonnage, by their criteria.

Otherwise you'd be counting motorboats, canoes, etc. These lost vessels do not count as shipwrecks, but, as the release worked its way via different news outlets to npr it became lost ships and shipwrecks. Ships are not lost at such a high rate. That would be the rate of loss at the peak of wwii.

[leblon.livejournal.com]

According to this criterion, a wooden ship of any size is not a ship, but only a "vessel". Since I doubt that NOAA keeps track of all sunken canoes and motorboats, I suspect that that a large portion of these 20,000 sunken vessels are sufficiently large to be called ships by a reasonable person (although not necessarily according to the official criterion). You did not provide a good reason to believe otherwise. And it is not actually correct to say that 20,000 a year is equivalent to a rate of loss at the peak of WWII. Dividing 20,000 by 200 years gives a loss rate of 100 ships per year. The peak rate was 3-4 times higher. Since ships were less safe in the 19th century, I do not see any conflict between these numbers.

[shkrobius.livejournal.com]

RULET database includes only vessels from 1890s on, so it is not 200 years. If you count small boats, then you were off by an order of magnitude in your upper estimate of the total fleet, hence the discrepancy. If you use their own selection criterion (NOAA) counting as ships steel hull vessels longer than 200', it is only 580 shipwrecks in the coastal waters since 1891. The other paper's count was starting from 150 gt for tankers and 400 gt for non-tankers, but they do not tell specifically what fraction was lost in the coastal waters. But it is 8500 worldwide!

The thing is that your logic was flawed. Even if one assumes the rate of attrition that you assumed, there is no reason to expect that ships are lost in specifically coastal waters rather than on the high seas, as that's where cargo ships spend most of the time. It is smaller vessels that dwell in the coastal waters where a fraction of them sinks near the coast. There is bias. About the war losses. Naturally, the Germans avoided attacking convoys in the coastal waters, so again most of the ships were lost in the Atlantic and did not make it into the statistics.

Luckily, you greatly underestmated the number of vessels and so you arrived at the correct estimate of the total number of sunk vessels. But NPR was mistaking this number for the number of shipwrecks, which are much, much fewer and more prone to occur on the high seas rather than in the coastal waters. That the two numbers coincided was misjudgment on both ends. You knew the number you believed to be correct and, predictably, you arrived at it through numerology.

You won't believe how typical are such "errors." As an exprimentalist I deal with such biases all the time.

[leblon.livejournal.com]

Does NOAA database contain only shipwrecks since 1891? I did not realize this and I did not see it mentioned in the report you gave link to. Plus, I did indeed try to estimate the number of all shipwrecks, not those happening in the coastal waters. So perhaps the fact that I arrived at a number closer to 20,000 than 10,000 is indeed a coincidence: I was not counting the same things that NOAA was counting.

"You knew the number you believed to be correct and, predictably, you arrived at it through numerology."
No, I did not know it in advance. I estimated it the way I described and then heard the answer on the radio. No numerology was involved.

[shkrobius.livejournal.com]

It is in another noaa report, sorry ( if you google rulet, there are several reports at the top). noaa do not have their own database for old ships, they use uk database that goes to 1873. They are interested in oil pollution, so coal fired ships are of little concern. They aren't Lloyds.

I guess journalists only look at the press release and executive summary, where it is already ambiguous. Perhaps different people narrowed the list in stages, so it is first sunk material, then lost vessels, then wrecks, and then shipwrecks that made it into the summary.