Hacker News Reader

Comment by xxpor | Hacker News Reader

>It highlights that liner companies (e.g. Maersk) are leaving cash on the table by selling capacity at rates which don't reflect demand, although perhaps they prefer long term stability.

It seems like there's an opportunity here for a bidding based shipping company.

I can't even begin to fathom the time or capital required to start a new container shipping company from scratch though these days.

View on HN

Pretty much the entire ocean transportation industry, with the exception of container shipping, works this way. Prices are largely based on predictions of market conditions in your load port at the time of loading, and the the same wherever you’re sending the cargo.

If you’re trying to load cargo in an area with an undersupply of vessels, you’ll pay a lot for the opportunity. If you’re sending a ship to an area with an oversupply of vessels, you pay more for that because the company renting it to you doesn’t have good confidence in its ability to get a contract once your cargo is offloaded.

Despite the significant increase in commercial intelligence products from satellite imagery, ripoffs etc, the information available is still quite limited. And the MILP problems that result are extremely challenging to solve. A lot of the shipping companies still seem to use heuristics instead of models.

But shipping lines seem able to control capacity because of consolidation, and hence maintain high utilisation:

https://www.freightwaves.com/news/qa-flexport-on-2021-contai...

I helped start an organisation which builds data science models for the shipping and commodity markets, and has had great success trading on the results of those models.

Within most shipping companies, there continues to be a significant hubris amongst charterers of "I know what I know" - a lot of people who have worked for decades in the industry and still trust their gut instinct over anything else.

As a result, since so few decisions are significantly data-driven, it's actually not that complex to build predictive models that take advantage of arbitrage opportunities in the market.

The capital requirements are certainly no joke, but the ships themselves are cheaper than I expected. The Ever Given, for example, supposedly cost around 130 mio. dollars to build. For comparison, a new Boeing 747 costs around 230 to 260 mio.

Ships don’t last as long.

Reply on news.ycombinator.com