Opportunity cost and bulk measures fit here, so math talk a bit.

JIT (Just In Time) is a delivery method that focuses on anticipating needs in highly complex production chains. The automotive industry is a perfect example of this. A manifold has a dozen steps to create, and the first one takes place months before it’s actually going to be used. Assuming a stable logistics chain, JIT ensures you have only what you need, when you need it. Saves overhead and unnecessary storage costs..

Saturation is a delivery method that quite simply fills up the storage containers as the production chain is simple and volatile. An ice cream shop is an example, where the product can stay frozen for an extremely long period of time. One rainy day you sell nothing, the next sunny day you have a run on double chocolate. This allows for stock protections, but does have a cost for storage. For smaller items, this is less of an issue. Larger volumes – issue.

For what it’s worth, Amazon uses the saturation model. Next day delivery of a pair of shorts only works if they have them in the giant warehouse after all.

Factorio

For the wide majority of the game, Factorio uses saturation methods. ‘Fill a belt’ as it were, and when the belt looks empty, fill it some more. There’s very little math involved in the raw material processing, aside from knowing how many machines it takes to fill a belt. Vanilla Factorio had mini-factories, but that’s long gone now. You build big, and you fill that box at the end. Stack inserters broke the older model.

DSP

Also a game that focuses on saturation as mini-factories have very little benefit. Even with the best belts, you’re going to need to stack items. It’s maybe 10 seconds to build a line of 30 smelters that will produce a hundred or more plates. It’s meant for scale.

Satisfactory

While I prefer saturation, or perhaps simply more accustomed to it, Satisfactory is designed for JIT. A full belt feeding a dedicated crafting station will put out a pittance of material. This gets worse and worse as you build more complex items. There’s a reason there are offline tools to help you figure out the math on production chains. Heck, powering a basic Nuclear plant takes nearly a thousand raw items per minute.

Saturation works if your goal is simply progression. The challenge is that you need bulk material in order to unlock specific research / space elevator tiers. While a production chain will only ever use 1 Nuclear Pasta, you will need a collection of 100 to meet a given milestone. If you only apply JIT, you’ll never have that extra 100. Now, if you use the basic tools, such as a 48 stack container, you’re wasting storage. The balance here isn’t super clear, so your mileage may vary (I put in a mod to reduce storage, I set it to 5 stacks per container). The end result for progress is that you visually see a belt is full, and have minimal storage for when you need to unlock something new. You’d be surprised at how few machines you actually need.

If your goal is aesthetics and end-game scaling, then JIT is the only way to play. A nuclear plant takes 240water per minute, which is 2 extractors. If you want 40 nuclear plants.. well 80 extractors. It’s a pain enough that you don’t want to build 200. (It’s also 10 supercomputers per Nuclear Plant, and the base creation rate is 2 per minute. So 40 nuclear plants = 400 supercomputers = 3.5 hours of base crafting.)

Note: Don’t build nuclear plants for power, build them for looks. Rocket Fuel plants are a good 90% more efficient.

Basic Materials

The exception to JIT is for basic materials: Iron Ingots, Copper Ingots, Concrete, Caterium Ingots, Plastic, Rubber, and Aluminum. Trigons and Diamonds too, but that won’t matter much by the time you unlock them. Saturate the crud out of that stuff, it’s simple enough to do and in the case of Copper Ingots, you are going to need way more than you ever thought.

Sommersloops & Power Shards

This part messes up the math. Sommersloops double the output of a machine – it costs no extra resources but does require substantially more power. Power Shards increase the output of a machine, up to 2.5x the amount, for a commensurate amount of intake materials + power. Combine both, and you get 5x the output for 2.5x the input. What’s neat in Satisfactory is that you can set the output to a math formula, and it automatically resolves – like 3.5/2 will set itself up to 1.75 just fine. When you’re building a unique factory for say, a Thermal Propulsion Rocket, you’re going to see a lot of weird math. Having flexibility in and out is a must.

Note: I consider power ‘free’ past a given point. A simple rocket fuel plant can give you 50GW+. Power Shards are also free, but only once you reach tier 9. Until then, you need to harvest slugs in the world… still, it’s quite easy to have 300+ by the time you unlock nuclear. You’ll need 240 of them for a Rocket Fuel plant.

Starting this on a different footing. I really like the tools in Factorio. Space Age added a bunch of different puzzles, and through a serious amount of applied logic, you can do almost anything. The tools present mean that you spend very little time with finicky placement and instead solve math problems. The friction points have all but been removed. Dyson Sphere Program is 2.5D and has a very similar set of tools.

Satisfactory does have blueprints, which aid in growing scale, but they operate independently. You need to manually connect them together. At scale, this is super annoying.

1.1 Changes

Sweet baby carrots is this a massive quality of life boost! To get the auto-connect to work, you need to be within 2 spaces between the in-world item and your blueprint… not really an issue for belts but much harder for rails. This is entirely due to the Z axis (vertical) and how items need to be touching something. Belts work best when on the ground as they need to connect a machine… which should also be on the ground. Rails… they are meant to not be on the ground as you want to move across complicated terrain.

The general rule of thumb is that if only 1 axis moves, it looks really good. If 2 axes move, then it looks ok. If all 3 axes move, it looks like a 2 year old drew with crayons.

The impact for normal production blueprints is small. There’s a minimum distance for belts to run, so it is possible that blueprints need to be adjusted to ensure connections are larger than the limit. In my testing so far, belts set up to the limit of the blueprint box work just fine.

There are now 2 new types of valid blueprints – one for a belt bus and another for rails.

Belt Bus Blueprint

The idea is simple but building it is a bit weird. My take on this is to use a 4×4 blueprint and put in double belt holder equally spaced. Belts need to be on something. I added temporary poles next to them, ran the minimum belt length and then deleted the temp poles.

Placing this blueprint brought a tear to my eye. In the time it took to lay 1 extension, I can lay 20+.

Rail Blueprint

Rails should not be on the ground, and rails also have a minimum length. That means a small pillar with foundations on top that have 2 parallel rails. A very simple blueprint. It won’t look good as you’ll have rails floating in the air, but it is extremely quick and practical.

Placing this blueprint is not fun. You need to ‘lock it’ (press H), then nudge it close enough to the existing rails so that it connects. Then you click. And then you make a choice.

If you want to manually place the rail extension because you have terrain nearby, press H again, rotate and place the extension. This works for about 50% of the time you’ll use rails.

If you do not have terrain nearby (say you are trying to elevate a rail), then you need to manually nudge the rail blueprint (up/down, left/right, pgUp/pgDwn). The downside to this is that you cannot rotate the blueprint to make turns, so you may want to build a 45/90 degree turn blueprint as well. The crayon downside here is that your rails will have ‘humps’ rather than a smooth incline.

The net result is rails that work, placed in a tiny fraction of the time. This will save hours and hours and hours. And a few more hours.

Summary

I’ve been playing Satisfactory for a long time, and I’ve hit numerous quit walls over the years. All of those were related to a perception of wasted time. I had goals and would find myself stuck in minutiae for 80%+ of the time rather than the fun stuff. Blueprints were a HUGE boost in 0.7, if limited. Dimensional Storage removed the need for central storage in 1.0, saving dozens of hours in the late game running around for materials. Connecting Blueprints, to me, is the final QoL change which will dramatically remove friction in building factories.

While there are other tweaks I could suggest here that would aid in QoL, they are generally low friction events. (This is aside from the HD/Sommersloop/Mercer Sphere 4-hour scavenger hunt.) Finicky tweaks, for sure, but big system changes are pretty much all there now. Satisfactory is essentially ‘complete’. A heck of a journey to get here.

When I was younger I enjoyed TT games, though less 40K given the length of the gameplay and well, smells. It’s certainly better now with more adults. I’ve generally shied away from the video games. The lore is interesting but the IP is the selling point here, not the actual gameplay. Feels more like a Disney approach to milk a franchise.

Space Marines 2 was on sale and my gaming news feeds generally had high praises. The campaign for one, the associated coop and PvP. I dislike PvP for a multitude of reasons – mouthbreathers, try hards and bots notably. But a decent campaign is worth a shot.

I completed the campaign in about 6 hours, which feels really weird. The first few missions were interesting and flowed well. The second half was full of loading screens and aside from 2 specific fights, a cakewalk. Run in, melee, AE melee, execute an enemy, move on. With few exceptions, the guns in the game serve little purpose.

I get the grimdark setting. I don’t mind the gruffness of the characters where everything is dour and serious. I don’t see them as heroes at all, what given that 40K is an outright criticism of religious fanaticals. Everyone is effectively a bad guy. The setting and lore are solid in that regard, so hats off in that regard.

(Side note: there’s some irony that this game is a Gears of War clone given that Gears of War borrowed heavily from the setting. Oh, and Starcraft/Zergs a plenty.)

Mechanically the game is simple but effective. There’s weapon variety which is mostly meaningless, except for the melee options. The invulnerability from executions needs to be exploited to survive, and there are some battles where you simply get chain stunned to death. The AI companions are actually quite good here, which is nice.

But the grand total of it all is quite meh. It does nothing well, except give you the experience of playing in the setting. You certainly feel like a Space Marine, which is neat. The ‘mini-bosses’ are more complex than the final boss who is about parry/dodge timing. Quite honestly, it felt more like a 6hr interactive video than an actual game.

The game peaked at 200k players and then has stabilized at about 10-15k since, which is a fairly health multiplayer base. I have no interest in this, but for those who do, it’s good to see it still going.

Overall, I wouldn’t recommend Space Marines 2 unless you find it on a decent sale, or are a die hard 40K fan, and you likely already have it. There are many more games that have done this better.

I like experimental ideas and I’m supportive of devs trying something new. I’m back to the point I was in Foundry when I left a year ago, and have some thoughts on what Update 2 has brought. In short, they are an improvement over the prior version, but still need some time to cook.

General

There have been some QoL changes, mostly related to balancing recipe volumes. It is better. It still doesn’t support large scale, primarily due to stack sizes that are absolutely too low. Foundations stack to 200, it should be 1000. This becomes very obvious once you unlock Olumite (oil).

Recipes need a general rebalance of ingredients. Having a late game item require copper wire is dumb. You won’t have used copper wire for anything for a dozen hours by that point. You could build a dedicated offramp to build these weird one-offs, but the splitter/ramps/inserter size means you need to use about 500 foundation each time this happens. See prior point.

Research is too complicated and takes too long. For long stretches in the mid-game you will research something you will never build in order to unlock something you’ll build 1 of. Actually, it may take the right amount of time, just that you literally have nothing to do while it’s underway.

Elevators and bulk miners need some tweaks. Their throughputs are simply too low given their size. Good news is that mining base power management is generally improved. 5 Solar Panels + 10 Batteries will support 2 diggers.

Modular buildings are still a cool idea that is poorly executed. They take a pile of material to construct and take way too much power, and once built, generally can be turned off. They become visual achievements.

The Lava Caves and Firmarlite Sheet process is cool as an idea, but poorly executed. You can only place the massive buildings on open lava, which is not continuous. Think of it like connecting islands with foundation pieces and 2 belts (in and out). The production rates are so low that you need 16 of these buildings to choke a basic belt. If I could change the lava floor to make openings, just like water above, that fixes this.

Building robots (the end game ones) is still cool to see.

Galactic Market

There are two parts here.

• Building Robots and Shipping them
  • Robots are unlocked through research. You have no idea what a robot is worth unless you did through a pile of menus.
  • Building them isn’t necessarily hard, but is also isn’t fun. Get an assembler, build robots, belt them to a shipping pad and put them in space.
  • Selling the robots is too complicated. You can simply sell on the market for a 30% loss, or invest in the incremental game (see lower) to unlock the possibility to sell them on a planet. Each planet requires a license and a dedicated ship.
• Upgrading the Spaceport
  • This is an incremental, plain and simple. Build up 2 numbers (money + material), then press a button for a timer that adds 10% to some function.
  • To unlock some features you need to research them on the planet. Not clear why.
  • It is possible to make a mistake in an upgrade and I am not sure how to revert a choice (e.g. unlocking a useless planet, or upgrading the wrong thing).
  • The orbital laser is friggin’ cool! Wow!

Ok, I lied. There’s a 3rd thing that completely changes this game.

• You can buy almost anything from the space station, with I think a half dozen things left over you can only construct on the planet.
• In general, it is much cheaper to buy an item than to construct and then sell it.
• On paper, and with a few spaceport investments, it appears entirely possible to have a factory built entirely on the concept of shipping down from space, building a complex robot, and turning a profit. This completely negates all mining and productivity bottlenecks, assuming your ships have the throughput required (each shipping pad has a built-in buffer).
• I’ve yet to fully test this mind you. Nor do I actually want to. Nilaus has though!

In Summary

Foundry is really trying for some interesting bits here. If I take a step back, the concepts here are really quite something. The implementation needs some serious thinking. It’s a bit like when my kids drew animals from their imagination, super cool but not practical.

The devs have stated their next major update will focus on quality of life things, that’s good. Tweaking the ideas present so that they work together in a more streamlined fashion would be great for everyone.

I still recommend buying the game if you like the genre. There’s enough good ideas here to justify the price.

The best games are nefarious and subtle. They start simple and straightforward, gradually adding complexity without it being obvious, and then at some point you’re an omnipotent god juggling fine pieces of art surrounded by a chorus of followers. Like that. Think about Minecraft. The first 15 minutes you played had you punching trees and dying to zombies once the sun went down. By the end, it’s redstone everywhere and you’re shooting a nether dragon.

Production games are about making numbers go up, and each step is more complex than the last. There’s an art to progression here, where you go from ore to ingots to plates to engines to robots to spaceships, and each step naturally flows into the next. At no point should you ask yourself ‘what’s next?’ as the factory must grow.

Foundry’s early game manages this well enough, up until you hit the steel tier. Before that point, you have 5 possible inputs to sort out and can find a way to bus it and manage crafting. It’s all smelters, crushers and assemblers. Straightforward enough and there’s always something to do.

The steel tier though, that’s where it gets complicated. Making steel required a very long belt (compared to what you have) to weave different materials and then put it on the bus, not necessarily more complicated just longer to set up. Concrete + Steam is in that tier, and now you need pipes and 3 new types of buildings that use water inputs. To get to the concrete step you need to build another mini-bus due to the conflicting materials, and eventually glass production. This is complicated, because instead of extending your main bus (and what you know), you need to build a second one, so that it doesn’t conflict with the main one. It’s a weird step back and sideways, rather than forward.

And then we get to Lava Caves / Elevators. The voxel world typically has you start the game at 150 units of height. Lava Caves are at 0. To get there you need to put an elevator and there are 2 types. One for you, one for freight, and they operate differently. The personal elevator has you select the depth, and can only dig through certain material. If you hit a single rock you need to manually remove it, and potentially don’t have the research unlocked to do so. Eventually through manual digging you reach the ground floor. The freight elevator is placed with a top (at +150) and a bottom (at 0) and it will self-connect if there are no rocks. If there are, you need to use the personal elevator to find them. This is still a baffling design choice to me. When it’s all working it’s really cool, but getting there is pure friction.

The Galactic Trade system changed a lot of the flow of the game, with a sort of side game of making literal spreadsheet numbers go up. I’ll have more on this in a bit, but it’s a significant break in game flow and an actual impediment to progress as Firmarlite (that 2nd R irks me) Bars are kept behind this mechanic. You need those for green research and access to the mid-game+.

This is a negative take on an experimental game, that comes from oodles of time spent in more mature and polished titles. I can emphatically say that Update 2 is miles better than what came before (belts and pipes for sure) and it’s clear there’s still a long ways to go. Pacing, tooling, and friction points are notoriously hard to balance, and exceptionally so if you have dev tools to skip pieces. I am really looking forward to testing more of the experimental components, there’s so much potential here.

Foundry hit EA about a year ago, I gave it a shot then. It is a 3D procedurally generated world, built on voxels (key point, that), that experimented with some ideas. Right up until mid-game, the idea of a mega bus remained practical. You could build mega structures in pieces, and finally robots along assembly lines. It felt a bit like Lego, where you could see the potential but were missing a few bricks.

My gripes then were about powering mining bases (high/low voltage stuff is way too complicated), and throughput logisitics are essentially capped due to lack of trains (or their equivalents), making a mega bus effectively starve itself. You couldn’t effectively ‘make numbers go up’ beyond a certain point.

Update 2 Notes

Last fall the devs surveyed folks for what they wanted to see on the roadmap. Top of list was production-related changes, balances, and similar items. They did want to see a greater expansion of the robot production chains! So we got Galactic Markets, which apparently is a system that allows you to sell your robots to space folks. From the patch notes, my selection of highlights:

• Added Galactic Commerce:
  • New Galaxy map:
    • Procedurally generated galaxy on each fresh game start.
    • Unlock galactic sectors and acquire trade licences to sell your robots to planets.
    • Set up and manage supply routes between planets to distribute your robots.
    • Buy and sell resources on the galactic market.
    • Compete for market dominance against other companies. (Hoping this is not PvE for markets)
  • New Space Station Features:
    • Dozens of new station upgrades.
    • Spaceship Management: Buy spaceships of different types and assign them to various tasks.
    • Establish trade routes to the galactic market.
    • Sales Platform: Sell your products to casual customers.
    • R&D Lab: Earn XP and levels on each produced robot and improve your products.
    • Fuel Station: Produce your own spaceship fuel to supply your spaceships instead of buying it from the market.
  • Many new robot types for you to build and sell.
  • Choose your company name and logo.
  • New Company Rank system: Increase your rank based on your lifetime earnings.
  • New Shipping pad buildings to ship items between the space station and the planet.
  • New station terminal building to contact the space station.
  • Keep track of your finances on various accounting-related charts and tables.
  • New feature that allows you to pay back your debts.
  • New research options to fit the commerce narrative.
  • Countless balancing adjustments.
• Pipe system 2.0: New and improved pipe flow simulation, including performance improvements. (I liked the old pipe system, anything to avoid Satisfactory’s version)
• Added new smart conveyor drag mode. (It’s much improved)
• Added new starting planet option which affects which biome and resource distribution. (This seems like a bad use of dev time. More later…)
• Added orbital uplink tool and space laser that can be used to terraform large areas.
• Added lava caves and lava smelters. (Caves in general are neat in concept, not neat in execution)
• Added new Tundra biome.
• Added new jungle/sandy desert/forest critters.
• Added new underwater decor/vegetation.
• Add new freight elevator III/IV. (This is a massive improvement)
• Incompatible with prior saves, meaning a fresh start is required.

Trade Interface

There really isn’t much here to be honest. You get a new building early on that enables shipping of material to the space station (this building doesn’t require inserters or power, which speaks volumes to game design choices), which makes a number go up. The space station itself has no interaction outside of a menu. What the dev stream has shown seems like a precursor to something larger, which still seems like on the edge of potential.

I am in the early portions, just having unlocked green science. Maybe there’s more to this.

Overall Thoughts

On the one hand, cool that there’s new systems and very curious as to how this will work out long term. It’s weird building end-game systems and asking every player to sink 20 hours to actually test it, but hey, that’s EA I guess. I’m looking forward to a much different set of goals to try out.

On another, there are some core balancing issues that still seem present. Great to see Freight Elevator improvements on throughput, because splitters and multiple elevators was a pain – there are no other vertical belt options. The game still keeps splitters behind alternate research paths (this feels QoL to me), the 3rd row inserter is way too late in the tree (on the edge of QoL and bad design), and power management still feels painful until you build acres of solar panels (there wasn’t occlusion before, so a tower of panels actually works).

Games like this need effective logistics… there’s still a fair chunk of work needed. Some interesting ideas here, but there’s still a ton of rough edges and strange mechanics. I remain hopeful they can figure this out and deliver it in a reasonable timeframe.