Hacker News Reader

Comment by relix | Hacker News Reader

I may be misremembering, but from how I understood container shipping works, that is always the case with lost / damaged containers, no matter the size of the event. If only one container falls off the boat, all entities who were shipping things on that boat are required to chip in proportionally to the amount of containers they were shipping, to make the owner of that container whole.

The idea being that otherwise it's not fair to the owners of the top level containers, which have a higher chance of falling off than the bottom level containers. Everyone's share of the boat is equal.

For this reason you always want to take the insurance that is offered against this, just in case that one container contained something worth a few million, and suddenly you're on the hook to pay an amount that is a multiple of what you were shipping in the first place.

View on HN

From some rudimenatary googling it's based on the value of the goods. So unless your cargo is expensive you wouldn't pay nearly as much. Unless there were expenditures there's no way you would be out more than its worth. Also because if this they'd likely load extremely valuable cargo in positions were it's not likely to be lost or jettisoned.

Shows quite nice how important it is to really understand your Incoterms. And to procure insurance to cover any risk left open.

> And to procure insurance to cover any risk left open.

Or alternatively, to be aware of which risks you are still bearing. Not all risks have to be insured.

True. But than full, or almost full, transport insurance is cheap enough even for small businesses.

> The idea being that otherwise it's not fair to the owners of the top level containers, which have a higher chance of falling off than the bottom level containers. Everyone's share of the boat is equal.

In general, being able to charge different prices to different customers is very beneficial to the seller. So should we expect this to change as ship operators figure out more sophisticated pricing models?

I always thought part of the advantage that shipping containers brought was that they were more or less interchangeable. And that meant that they could be loaded in whatever manner best suited the whole ship. The details of which might not be known at the time when the container was booked. The carrier might be able to reserve “interior” or “top level” spaces for a different prices, but with shipments being loaded/offloaded at different ports, the logistics of loading can be quite complicated. If your container need to be offloaded at the first of three ports, you probably couldn’t have the option of an interior space. Or if you had an extremely heavy container, that should be loaded in the lower part of the ship for safety.

My guess is that the orchestration of how containers are loaded is complex enough that the order of stacking is best left up to the shipper and adding cost/location optimization to the equation might make things too complicated. The containers should be loaded in the manner that is the safest and most efficient for unloading.

I'm sure it's just a level of optimization that's hard to solve because of a few reasons. a) it's a fairly hard technology problem. Large container ships can carry 20k containers. If each of those has a spot, and you have to optimize, the search space is suddenly looking like 20,000!, which is just a wildly big number. So assuming you have some sort of technological solution and know where every container should go, b) you need to move all the logistics around so that containers arrive in the correct order to get loaded in the correct spots.

You need to do all this just to massively increase the complexity of your billing model to offer slightly different rates of insurance. I'm sure there are teams looking at ways to make this better (just stipulate in the contract that any super valuable containers arrive early so they're loaded lower), but this is a HARD problem to solve completely.

This is such a HN thread..

The same effort is better invested into making sure you are not losing containers.

It's like saying 'we can optimise policing so that stabbings only happen to the right people'

Do you think that doesn’t happen? Police departments absolutely prioritize certain areas over others - Places where crime is unlikely to happen but should be dealt with swiftly, places where crime is very likely to happen and they have resources on-hand. The net losers tend to be “The Average Person”, for whom police response times in event of a burglary or domestic incident is measured in tens and fifteens of minutes rather than fives.

Even “Even spread” would be a serious topic of debate - Even spread over what? Geographic area? Residency? Where people are likely to be at a given time? Those are three radically different answers in any commuter city.

I'm not an expert on ocean shipping by any means, but I would assume that cargo value is such a low priority when it comes to determine loading strategies that you would only need to optimize around a local area of maybe 100 containers. Which would not make it a hard problem at all but it could be pretty pointless depending on how many containers go overboard in a typical "lost container" event.

Also, consider that from the point of view of the people loading the ship, it's all insured. And if you didn't get the insurance then you understand the risk. Why would they care? They aren't the insurance company. What's their incentive?

A big search space is not a detriment to optimization. It just keeps you from reaching provable optimality, but doesn't keep you from saving on costs compared to doing nothing.

Same for the rest of what you describe: yes, a perfect system would be a lot of hassle, but making any effort at all is sure to yield benefits.

You don't need to change the billing model necessarily. The customers don't have to know that you are doing this kind of optimization.

Of course, the more you can expose incentives to your customers, the more efficient you can get.

But even if your containers arrive in arbitrary order and you have to make your decisions on where to place them 'online', ie before the next container arrives; as long as you have any choice left at all, you can try your hand at optimization techniques.

> Or if you had an extremely heavy container, that should be loaded in the lower part of the ship for safety.

By the way the standard ISO 668 intermodal container has a total limit of about three tonnes; the unusual, heaviest size is about ten tonnes. A container port won’t be prepared to handle anything heavier or even, in practice, the 10 tonne version. These constraints apply to the rest of the system (roads, trucks etc — they are “intermodal” after all) so anything outside these constraints (wind turbine planes, major ship engine blocks etc) isn’t containerized anyway.

In other words the point of containerization is to be able to ignore these distinctions. If you care, you ship as if it weren’t in a container

Oh, just because you can ignore some distinctions doesn't mean that you will.

If a shipping operator can eke out a few extra percent of efficiency by taking container weight into account, they will try to do so.

Of course, they will have to balance the efficiency gains against extra complexity.

But it's not so different from eg Amazon taking what's going on inside the AWS VMs into account when deciding on how to run their data centre.

Eg you might have booked a VM with four cores, but if you are typically only using one, they'll take that into account during placement.

3 tons is tiny. An empty shipping container weighs at least 2 tons.

The actual ISO 668 limit seems to be 30 tons.

https://en.m.wikipedia.org/wiki/Intermodal_container

I guess that would mean they have to distinguish between containers when loading them onto the ship. That's probably difficult and more expensive for everyone.

Wow, this is sobering information. I shipped a car from the UK to the US several years ago. I purchased shipping insurance on the container and car. And watched nervously from afar as I tracked it online when possible. I had no idea there was any possibility I could have any liability for others losses. If I understand this correctly...

I wonder when it becomes more economical to load top level containers with cheap goods and keep the multi-million-dollar containers somewhere safer.

You'd need some significant automation and integration with all the port-side trucks, forklifts, and cranes that unload and load your ship to make sure that containers arrive and depart in the right order. You could then offer customers lower insurance premiums (or let them pay more for interior space) because average loss would be reduced. But that would take a huge amount of vertical integration or an industry-wide cooperation effort and cost sensitivity that I don't think is present.

Loading is already very complex and takes special knowledge and more and more software to balance the ship and keep the loading-process efficient. Dangerous goods are separated from the quarters as much as possible for example. I'm sure the process already includes variables like value.

But I also assume that overall, unless we talk about unique items like artwork, not enough containers are lost to put much weight on the value of the contents compared to everything else.

Reply on news.ycombinator.com