Dev Log Week 2026-04: State of the Game

Obviously, the most noteworthy thing to happen last week was the release of this year’s State of the Game video. If you haven’t seen it yet an got 45 minutes to spare: Go ahead and check it out to get a high-level overview of last year’s main projects was well as my plans for the coming weeks, months and beyond.

Besides the video I was busy figuring out why Location-based Demand shows quite unexpected behavior in combination with Individual Travel Request. Since I plan to eventually transition to both by default (in new game worlds), solving this is fairly important. But as it turns out, the required adjustments won’t be trivial. For example, we noticed that tourism-heavy areas show far too little demand. Which makes sense, given that demand is solely based on population numbers in the current iteration of LBD. Similarly, cargo is just as spread out across (even the smallest) destinations as pax, with the difference being that there are a lot fewer units in total, meaning it is near impossible to make cargo work right now. I’ll look into both (and other) issues together with the volunteer data team over the coming weeks. That said, I don’t expect a fast solution here, so it might take a while until LBD becomes generally available.

Towards the end of the week I also started looking into the last remaining “open items” for ITRs and LBD, but didn’t get too far. Adjusting the rating displays for ITRs across the game turned out to be surprisingly difficult and I am actually wondering whether it’s worth the effort. More on that next week, though (probably).

2 Likes

Apart from tourist heavy destinations:
How are big transition hubs like DXB/SIN doing? I would also expect way less demand with LBD than before ^^

Yes and no. Haven’t looked into it yet, but for airport-based demand, I already implemented several counter measures to deal with large hubs. Both in terms of absolute demand (reducing the total demand of large airports) and relative demand (fudge the numbers somewhat during pre-processing). The former should in fact solve itself with LBD. The latter relies on the same tuning as the current model.