Logistics net: Less research needed

[jonatkins]

Some dude launched the rocket in under 8 hours using Burner Inserters with Logistic Network ban

That's one great thing about Factorio - once you've mastered the gameplay with all available features there's endless replayability in there by setting personal restrictions - complete the game without using feature X and/or Y.

And that's why I don't like this change in 0.15 - it *forces* requester chests into a much later part of the game than they were before.

In 0.14 and earlier, requester chests were already needed blue science, which was not a requirement in any other logistics research. Logistics bots also require advanced circuits, therefore needing oil+plastic in place for those too.

And.. (correct me if I'm wrong) no-one complained . It was fine - there was a reasonable progression through the research + production requirements for setting up construction + logistics.

Now, this ONE SINGLE PART of the logistics system has been moved far ahead of the rest of the system in terms of research requirements. It just doesn't make sense.

I would argue that the game would be more fun if Construction Bots + Personal Roboport were available sooner.

I've seen this comment come up a fair few times - even from the devs themselves. Personally, I disagree. Personal roboport should be further ahead in the research, not sooner. (static roboport + construction bots is fine as it is though)

But, I'm not going to suggest the personal roboport is made harder to get - I don't want it early, so I don't research/build it until needed. Now, can you (the general 'you' of all commenters happy with the change) accept that some people liked where requester chests were in the research tree before, so the extreme change to research needed for them should be reverted?

[dragontamer5788]

Now, this ONE SINGLE PART of the logistics system has been moved far ahead of the rest of the system in terms of research requirements. It just doesn't make sense.

Why not? Its clearly the most powerful system in the game. Requester Chests literally solves all the problems in the game to such a gross degree that they trivialize design. There's always going to be players who want to play with overpowered items and stuff that honestly shouldn't exist. That's why there's demand for gamebreaking mods like Factorissimo. Players also never like nerfs, but I think this nerf is very much warranted.

I understand that people don't like it when things are nerfed. But I honestly think that Logistic Bots are so powerful that they deserve the nerf. Its only a delay of a couple of hours anyway. Tell me, how long does it really take for you to set up the last few science packs? We're talking about a nerf that pushes Logistic Bots out by no more than a couple of hours forward.

And.. (correct me if I'm wrong) no-one complained . It was fine - there was a reasonable progression through the research + production requirements for setting up construction + logistics.

Factorio got harder in 0.15, extending the game. That's not a bad thing IMO. The game is basically over as soon as you get Requester Chests. The "main-belt" becomes completely obsolete and it becomes better to spam Logistic Bots. The main difficulty bump is the need to get blue, purple, and yellow science automated without the help of Requester Chests.

That... only makes sense IMO. As they are set up right now, Logistic Bots are basically there for the final push for the rocket and for the infinite post-game. If Requester Chests were added before Yellow Science, then "solving" Blue, Purple, and Yellow Science would become ridiculously trivial.

I've seen this comment come up a fair few times - even from the devs themselves. Personally, I disagree. Personal roboport should be further ahead in the research, not sooner. (static roboport + construction bots is fine as it is though)

But, I'm not going to suggest the personal roboport is made harder to get - I don't want it early, so I don't research/build it until needed. Now, can you (the general 'you' of all commenters happy with the change) accept that some people liked where requester chests were in the research tree before, so the extreme change to research needed for them should be reverted?

This isn't quid-quo-pro. This is a discussion. You have your opinion, I have mine. I'm fine with whatever the Devs decide. But I'm strongly on the side of the Devs in this case. I believe 0.15's balance change (effectively nerfing Logistic Bots by making them a later research item) improves gameplay.

You clearly think the opposite. And that's fine. We're all entitled to our own opinions here.

[Kelderek]

The more I watch this discussion the more I think the main reason I find this change to be well balanced is in how cheap the requester chest technology is at only 150 science packs. You can easily slap down a handful of assemblers and maybe a couple chem labs and have a rudimentary yellow science automation setup. Go off and do something else for a little bit, maybe extend a rail line or build a mining outpost and before you know it you have finished enough of these packs to buy the requester chest tech and are free to build a full bot based design from there. In my current game I managed to get this tech with only 2 assemblers each for blue, purple and yellow science and in most cases I just used the basic automation I had already built for my "shopping mall" where I was already making things like the batteries and modules to use in other stuff.

That was just one way to do it and everyone plays different. You're not forced to do anything too elaborate if you don't want to. This tech is right where it belongs in relation to how powerful it is and it is hard to complain when it it is one of the cheaper techs based on number of packs needed.

[jonatkins]

Now, this ONE SINGLE PART of the logistics system has been moved far ahead of the rest of the system in terms of research requirements. It just doesn't make sense.

Why not? Its clearly the most powerful system in the game. Requester Chests literally solves all the problems in the game to such a gross degree that they trivialize design.

That's something I just don't get. Yes, they trivialise a low-volume production route, but anything with large volumes is far easier to keep on belts I find.

The things I use logistics bots to deliver supplies for are making such low-volume items - chemical plants, refineries, assemblers, power poles, combinators, regular+logistics chests, roboports, etc. Things that are needed when you're building new stuff, not for things needed in continuous production for research.

This seems, to me, the ideal use for them. I can hand-build the start of a production line, then blueprint to extend it out for as long as needed - and without worry know I can rely on my base logistics system to make then required items and construct the rest. No need to stay in the area and wait for my personal roboport to do the work, or wait around for my character to craft the items needed.

I can't imagine how I'd convert my current long-term game (160+ hours) from belt+train to logistics driven. Infinite research is progressing with a production rate of just over two of each science pack per second, which in the high-volume production area involves 6 blue belts of basic circuits (which consume 6 belts of iron + 9 belts of copper), with two basic circuit belts + two plastic belts going into advanced curcuits, and.. well, you get the idea. (I would go for higher rates, but FPS/UPS drops too far beyond this point)

And, with all of this production rate - there's 500 logistics bots, and most of them are idle unlless I've ordered construction or return to the base for re-supply.

Factorio got harder in 0.15, extending the game

Yep, it did. Mainly by the additional science pack types and the resources needed to produce these at a good rate. I generally play by doing all research I can before launching a rocket, and this has gone from about 30 hours to about 50 hours of play. But now I find I'm at the point I want to start major rebuilding of the base to support the volumes needed for purple + gold science, and this change takes away a key tool that makes it easy to build this infrastructure. I've ended up so frustrated that I do end up setting up single purple + gold science assemblers, and manually carrying ingredients to these where belting wasn't practical - and it ends up a tedious couple of hours of gameplay. Once the research is done I can then get on with the needed base upgrades for purple + gold science at sensible production rates.

[dragontamer5788]

Now, this ONE SINGLE PART of the logistics system has been moved far ahead of the rest of the system in terms of research requirements. It just doesn't make sense.

Why not? Its clearly the most powerful system in the game. Requester Chests literally solves all the problems in the game to such a gross degree that they trivialize design.

That's something I just don't get. Yes, they trivialise a low-volume production route, but anything with large volumes is far easier to keep on belts I find.

A blue-belt can only move 40-items per second. You can fit an infinite number of bots into any particular space, meaning bots are only limited by the charging stations (IE: The density of 4MW Roboports). Each roboport can support roughly 800-tiles of movement per second. If you have multiple Roboports charging in parallel, you can achieve the most throughput in the game in the smallest amount of space. Ex: four roboports support 3200-tiles of movement per second (16MWs used). As such, Logistic Bots are THE BEST way to load and unload from Trains... especially since Stack Inserters are far faster from Chest -> Wagon than Belt -> Wagon.

Tell me, have you ever put 4000 Logistic robots in a Roboport network? Because that basically solves every problem in the game. There are a few issues that come up (including the ridiculously dumb AI that visits chests even though they're full...), but once you get used to the quirks, its clear that Logistic Networks solve literally every problem in the game.

You just gotta split up logistic networks in an intelligent manner to mitigate the dumb-quirk issues. Its a good idea to have an isolated network for train offloading, so that the "visits already full chests" problem is mitigated. The "visits full chests" issue only really comes up in train-stations (when big deliveries of ~8000 or 16000 ore from 4-cargo or 8-cargo trains arrive. The Bots all visit the same chest, even if chests can only hold 2400 ore). And aside from that, you need to have an understanding of the MW-limitations of bots. Bots use energy every tile they move, so the power-requirements of logistic bots can be large (requiring a large number of Roboports to deliver power effectively).

The only time I use belts are when I'm sure an area won't expand beyond 40-items per second... maybe 80-items per second if I'm willing to lay down double-wide belts. But if I'm building an area with more traffic than that... Bots are the only solution available in the endgame.

And, with all of this production rate - there's 500 logistics bots, and most of them are idle unlless I've ordered construction or return to the base for re-supply.

500 is a ludicrously low number. I have 500 bots per logistic area, and over 4000-bots in my main area. And I don't even consider my base to be a "Megabase" yet... not compared to what everyone else around here is doing anyway.

Lets see... my major logistic areas are:

1. My main area of ~4000 bots (I was lazy and designed it poorly. Its due for a refactoring).
2. My Iron Smelters: ~500 bots that deliver from train to belt. I'm only serving 120-iron per second so... its sufficient for now
3. My Copper Smelters: ~500 bots again for the same purpose. Only 120 copper per second, so I don't need many bots.
4. My Circuit / Adv. Circuit area: ~1500 bots in a "lazy" configuration.

My "northern thrust" is composed of isolated Logistic / Construction Networks that automatically pass walls, electric poles, laser turrets, chests, inserters, logistic bots, and roboports further and further north (a train delivers raw supplies to the supply, but logistic bots deliver all items north. Long-inserters passing from Requester Chests -> Provider chests separate the logistic networks to ensure that the network doesn't suffer from "dumb AI" issues). Each of these are composed of ~20 logistic bots. (I needed the long-inserters so that they can read from the logistic network's wireless item count and perform logic on that. I only want to store ~5 laser turrets in each 50x50 construction area)

I mean seriously: with decent design powers, you can literally solve every issue in the game. Self-building laser walls (I stamp down blueprints using map-view and then logistic bots deliver the supplies... including construction bots... which then automatically build the next +50 tiles that explores northward)... offloading from trains, giant piles of assembly machines with Requester Chests / Provider Chests automatically pulling everything. Logistic bots literally solve every problem in the game in a ridiculously effective manner, obsoleting everything else by huge margins except for trains.

[British_Petroleum]

Requester chests aren't even that good. You can win the game faster without building any robots, so i think people should be able to have them earlier on if they want. Maybe "easier logistics network research" should be a game option? Or make it require gold science only if expensive recipes is selected

Also, Gold + purple science is still easy to setup compared to modded science.

[dragontamer5788]

Requester chests aren't even that good. You can win the game faster without building any robots, so i think people should be able to have them earlier on if they want. Maybe "easier logistics network research" should be a game option? Or make it require gold science only if expensive recipes is selected

By that logic, Laser Turrets (which have lower DPS than Gun Turrets at the same research level), Solar Panels, Nuclear Power, Trains, Beacons, Tier2 / Tier3 modules are all bad too. Based on my speedrunning attempts, you pretty much want to go belts all the way (Primarily Logistic Level 2: Red Belts for a good balance of cost efficiency / space savings), Steel Furnaces for smelting (moved close to a coal mine), Steam Power... and Gun-Turrets for defense (manually fed Gun Turrets. With 200-bullet capacity, that's more than enough to survive till rocket-launch)

But I don't think we should be using speed-running strategy as the sole indicator of "good" strategy vs "bad" strategy. Don't get me wrong: speed running is a great metric that should be discussed, but most players aren't speedrunners. The speedrunning game is a grossly different mindset than what most people are used to. Ultimately, I do have some agreement with your sentiment. I just want to make sure that its taken with the right mindset. Speedrunning is just one viewpoint in the world of Factorio. Megabase building / post-game is another viewpoint (perhaps one that I'm overemphasizing).

From the "Megabase" perspective, Trains, Logistics, Beacons and Nuclear are the primary drivers to making a huge and efficient endgame base. Almost nothing else matters. Use Nukes to take out biters and expand, while using Trains + Logistics to solve every major problem elsewhere... with Beacons + Productivity3 to create smaller factories that require less outposts to support.

From the Speedrunning aspect, something like 80% of this game is irrelevant that should be skipped to get to the Rocket Launch faster.

[jonatkins]

Tell me, have you ever put 4000 Logistic robots in a Roboport network? Because that basically solves every problem in the game. There are a few issues that come up (including the ridiculously dumb AI that visits chests even though they're full...), but once you get used to the quirks, its clear that Logistic Networks solve literally every problem in the game.

With Bob's mods, yes - had 6000+ and them doing all the work, no belts anywhere. But that's with MK4 bots + MK4 roboports and occasional MK4 charging stations.

But in stock/mildly modded? No. I have, once, tried using bots at a mining outpost, after reading about the idea somewhere. And although I did get it working eventually, I didn't like it. With the number of bots needed, and the number of roboports to keep those bots charged, it seemed far more sensible to stick with belts.

As for managing large-scale unloading of trains - well, that's where I do use mods nowadays to improve things. In particular, Warehousing to give a large 'chest' that has multuple access points per side, and loaders for a compact way of filling a belt from the warehouses. There's also compound splitters for compact belt balancing. Limiting factor now is stack inserters unloading from trains to warehouses.

I could probably manage similar throughput completely stock, but it'd need more platforms to make up for the chest->belt limits, and far more space for belt balancing.

500 is a ludicrously low number. I have 500 bots per logistic area, and over 4000-bots in my main area

Not for me - it's far more than I need, in fact. With the bots jsut fueling trains and nuclear reactors, and delivering barrels of sulfuric acid/lubricant, there's no more than 20 active at once. I only needed more while I was expanding the base, to keep construction supplied, and even then 500 was overkill.

Logistic bots literally solve every problem in the game in a ridiculously effective manner, obsoleting everything else by huge margins except for trains.

At the end-game, they can - once the stack size + speed have been researched to the max. But when you first get them? No - they're so slow they're only usefull for low volume production.

One thing I could suggest - I wouldn't object to low limits on the total number of active bots, with research needed to increase this. It could start crazy low - 10 bots max active, with red, then red+green science resrarch needed to get this to 50-100 range, blue science then for up to 500, then purple, gold then finally space science needed for an infinite research series unlocking higher and higher numbers. What do people think of this idea to limit the usefullness of logistics bots in the early/mid game as an alternative to the current locking of requester chests behind purple+gold science?

[dragontamer5788]

As for managing large-scale unloading of trains - well, that's where I do use mods nowadays to improve things. In particular, Warehousing to give a large 'chest' that has multuple access points per side, and loaders for a compact way of filling a belt from the warehouses. There's also compound splitters for compact belt balancing. Limiting factor now is stack inserters unloading from trains to warehouses.

FYI: Loaders were removed from Factorio stock because they were considered overpowered. I personally think they're needed to be counterplay against the strength of Logistic Bots... but hey... its not my game to balance.

I could probably manage similar throughput completely stock, but it'd need more platforms to make up for the chest->belt limits, and far more space for belt balancing.

I disagree. Consider this: 12-stack inserters to chest per wagon yields 270 Items per second, while 12-stack inserters to blue belt yields only 153 items per second. Since Bots automatically pick up items from an active provider chest, we're looking at 270-items per second (per wagon) on a bot-based train unloader. To replicate that with belts, you have to use 7-parallel blue belts, nearly perfectly balanced (96% balanced), when Wagons are only 6-belts wide. This is already a difficult task... (it might be possible with some braiding however)

But it goes from "difficult" to "impossible" when you realize that achieving 270 items per second requires 21-stack inserters (to blue belts) per wagon. Wagons only have enough room to hold 12-stack inserters. The only way to achieve an unloading system as fast as bots is to use the banned "Loaders" from ages past.

That's the thing about offloading from a train. You don't have space to play with. You have 12-squares per wagon to offload. That's it. The problem is only compounded if you're maximizing the efficiency of larger trains (like a 4-cargo train or 8-cargo train). Offloading an 8-cargo train will require 56 parallel blue-belts (if 96% balanced). Or... you know... just use Bots for delivery. Especially since bots self-balance the streams automatically.

At the end-game, they can - once the stack size + speed have been researched to the max. But when you first get them? No - they're so slow they're only usefull for low volume production.

And Blue / Purple / Yellow Science is low-volume production. So bots literally solve their own problems by trivializing science.

[jonatkins]

I could probably manage similar throughput completely stock, but it'd need more platforms to make up for the chest->belt limits, and far more space for belt balancing.
I disagree.

A misunderstanding - I meant I could probably manage the same throughput stock as I already manage with warehousing/loaders - I wasn't comparing to what can be done with logistics bots - I'd never tried to do the maths before either.

However, while you're right that one station with logistics bots can unload faster than the same station with belts, you just use multiple stations to unload the trains. With my warehousing + loader setup I've got two platforms each for copper + iron ore unloading, and this fills 12 or 14 blue belts to the furnaces. To match this without these mods would probably need a 4 platform setup to compensate for the chest->belt rate limits.

So thinking raw numbers for the stock game with the 4 platforms, you can just about fill a blue belt from 6 chests on one side of a wagon - so 4 platforms, double-sided unloading, gives a total of 8 blue belts per wagon out of 4 platforms. They won't be fully packed, so let's balance it down to 7 belts and ... well, we're now at numbers around your bot setup.

So yeah, belts/inserters can match the bot setup with 4 times more stations. It's not pretty, but I have done close to this before I started using warehousing/loaders to shrink it down from 4 platforms to 2 (or even 1, if you're careful) for a similar throughput.

Anyway, back to the right topic...

And Blue / Purple / Yellow Science is low-volume production.

It depends on your definition of 'low volume'

I'm thinking things like flying robot frames that a slow trickle of two a minute is fine. And even with a handful slow logistics bots, that's not an issue.

One thing that does come to mind, after seeing you talk about bot capacity in terms of distance travelled between charges. This means that the speed of bots is not important to throughput - just throw enough bots at the job and have a request size large enough to cover the travel time, and even the slowest bots could manage high throughput.

Perhaps better if bot energy was based on flight time rather than distance? Slow bots would have to charge more, and the speed upgrades would be required to maximise the throughput. Of course, you could just throw more roboports into the mix to manage this extra charging....

[dragontamer5788]

So thinking raw numbers for the stock game with the 4 platforms, you can just about fill a blue belt from 6 chests on one side of a wagon - so 4 platforms, double-sided unloading, gives a total of 8 blue belts per wagon out of 4 platforms. They won't be fully packed, so let's balance it down to 7 belts and ... well, we're now at numbers around your bot setup.

I don't have the time for a full response yet, but it's 3.7 blue belts for double sided unloading per platform. It's closer to 2.2x more throughout with bots than with belts.

I'm hitting the limits of 1-2 trains however with three parallel stations in my central hub. (all outposts... Stone, iron, copper, uranium, oil barrels unload at my central hub with active provider chests). I'm thinking of expanding my trains to 2-8 trains for future proofing.

I've seen a megabase that uses 30ish cargo trains for deliveries. So its not impossible. However, intersection and chain signal design gets harder and harder the bigger your trains get. In other tests, there's a practical limit of 3 or 4 parallel stations per track. This limit is because trains have acceleration. In practice... The first station will have already unpacked all the iron ore on the next set of trains before the fifth station's first train leaves.

So regardless, the size and speed of bot based unloading makes a real difference once you start having heavy traffic issues on your rail network.

[jonatkins]

I don't see why there should be much of a limit on the number of platforms for a station, unless you're using trains that reverse out?

If you use single-direction trains, there should be no issues. Having signals between the inserters unloading each wagon even allows the next train to start moving in as the empty train is pulling out. Beyond 3 platforms, trains can be reluctant to pull into a station that's empty sometimes, but making each station not only have space for an unloading train, but for one queued behind it, after the chain signal that holds the queue before the split, will help here.

And, of course, each separate resource has it's own station to unload at, sized as needed - so multi platforms + long queues for iron and copper, while the others are smaller. I do plan to give a large bot-based setup a go, as I've heard it said that bots can have a significant performance improvement over belts.

Anyway, we're heading off-topic again here - the original discussion is about the research requirements to unlock requester chests, not the throughput of bots vs belts for train unloading.

So back on that original topic... I guess it's possible that, with a carefully set up bot system with enough charging points and crazy numbers of bots, that even when they're slow and have low cargo capacity, high throughput is possible. However, is that alone a good reason to greately increase the research requirements for requester chests in 0.15? Should it just not matter - if someone wants to build like that, let them? Or consider alternative limitations on logistics bots early on, such as a limit on the number of active bots with research to increase it?

There's a fair few people who've posted in this thread who don't like the 0.15 change - and a couple of mods that revert to 0.14 levels (e.g. https://mods.factorio.com/mods/ItsTheKa ... search-fix). But with no response from the devs here leaving us guessing as to their intention with the change, it limits us to saying "I don't like it, change it back" and "it's a good change, keep it" with limited progress beyond that.

[Koub]

Old people always prefer it the way it was before. And sometimes, it's not only old, but just people who are not flexible and adaptable enough in their mindset.

In France, currency has changed from Old Francs, to "New" Francs, then to Euros, and there are still people speaking half the time in old Francs (that don't exist since 1960). Even I have tho dislikes in the changes that occured since 0.9 : the way inserters facing vertically place things on belts and the soundtrack change. But the game has got so much better in the meanwhile that I fint it a small price to pay.

Logistic network has always seemed to me a "I win" button. You need some random stuff on the other side of the map compared to where you produce it ? No need to think, plop a requester chest, and voilà, you have it.
Now, you have to prove you can solve the logistic issue, to "deserve" the "easy" way. I find it right.

[Aeternus]

But ask yourself this: "Is it more fun?"
It's more challenging, yes. But if more challenge is equal to more fun, then why do we even have blue belts, red belts, faster inserters and whatnot?

I'll agree that using the logistics net for mass-transportation is not what it seems to be designed for. If your arguement is that it becomes the go-to for all forms of transport, then that's a balance issue that should (and easily can) be addressed. Simplest solution to forcing stuff low volume is to reduce the amount of items a logistics bot can carry, by eliminating the capacity upgrades, or making -those- top tier tech. This should not affect construction bots much since those deliver one item per flight anyway...

[dragontamer5788]

I don't see why there should be much of a limit on the number of platforms for a station, unless you're using trains that reverse out?

If you use single-direction trains, there should be no issues. Having signals between the inserters unloading each wagon even allows the next train to start moving in as the empty train is pulling out. Beyond 3 platforms, trains can be reluctant to pull into a station that's empty sometimes, but making each station not only have space for an unloading train, but for one queued behind it, after the chain signal that holds the queue before the split, will help here.

Because a train is unloaded in roughly 7 seconds. Which means that every 7 seconds, a new train should be pulling into the station.

Station 1: Begin unloading at time 0 seconds
Station 2: Begins unloading at time 2 seconds
Station 3: Begins unloading at time 4 seconds
Station 4: Begins unloading at time 6 seconds
Station 5: Begins unloading at time 8 seconds... except Station #1 is already ready (because the train in station #1 already pulled out at T= 7-seconds)

I don't have the exact numbers, but that's the general concept. It takes very little time to unload a train, even a 4-train or 8 train gets unloaded in roughly the same amount of time. Additional stations beyond 5 will be redundant with the above numbers. In 0.14, 1-2 trains were practically limited to 4-stations per track.

But ask yourself this: "Is it more fun?"

Yes.

Its clear that you and I have different opinions on what is and isn't fun. But that's fine, its impossible for everyone to agree on the same thing generally, and what's where the developer's choice comes in. I find 0.15 to be more fun than 0.14 because of the science changes. In particular, there's a logical progression of higher-and-higher sciences before you get to the "ultimate" power: Logistic Bots. You start off with slow Yellow Belts, then upgrade to Red, then upgrade to Blue, then upgrade to Bots eventually. There's nothing wrong with that.

[jonatkins]

In particular, there's a logical progression of higher-and-higher sciences before you get to the "ultimate" power: Logistic Bots. You start off with slow Yellow Belts, then upgrade to Red, then upgrade to Blue, then upgrade to Bots eventually. There's nothing wrong with that.

There's the key point - that's YOUR particular playstyle, not mine.

Multi-lane blue belt busses for bulk transfer for me are the ultimate goal - and the logistics system is just a very useful convenience feature for things that don't even need half a yellow belt. It removes the tedium of manual crafting of things, and avoids the need for complex messy belt systems for things that only need an occasionally active assembler. That's how I use it, and I suspect many others posting in this thread asking for the old research level.

Also, I feel this is what the devs had in mind when they first implemted logistics, but ways were found to push it beyond the capabilities of belts.

As things stand, there's the research for bot speed and cargo capacity, and the higher levels of this need purple, then gold, then finally space science packs. THIS is the progression in capability that should be there - it starts off poor, and can be improved, and at the same time the base is transforming from yellow->red->blue->multi-belt setups as needed.

Are you saying that this progression of bot upgrades doesn't work? They're too powerful before upgrades? Or these upgrades need to be made more expensive?

And if so, do you have any suggestions on changes that could be made to allow the logistics system from the 0.14 blue science level, but stop it being the "ultimate" power until, say, you've launched a rocket and are on infinite research?

[jonatkins]

Because a train is unloaded in roughly 7 seconds. Which means that every 7 seconds, a new train should be pulling into the station.[...]

Oh, and just to address this point - you're forgetting one thing. I'm talking about train->buffer chest->belt unloading without mods, so, as you pointed out, the limiting factor ends up being the buffer->belt rate, not train->buffer. So given it needs 3 or 4 platforms to unload at a rate around a single logistics station, you're looking at something like 20-30 seconds to unload each train.

I will admit I've not gone beyond 4 platforms myself though - as I end up reaching the limit of what I can put up with in terms of FPS/UPS before I've needed more. Since using the warehousing/loader mods, I've not needed more than two platforms.

[bobingabout]

I must agree that is just seems pretty dumb to put everything robot related on red and green science, then hold back a rather key component, the requester chest, right up until high-tech science.

That's literally like saying "Using robots is okay, as long as you only use them to build things, or interact with the player". (Player's logistics slots are a requester chest, and trash slots an active provider chest)
And in my personal opinion, keeping track of, and interacting with a moving entity such as a PLAYER is a more advanced technology than simply moving things around from one fixed location to another. The logistics system shouldn't even be it's own research, requester and active provider chests should just be given to you right there on logistics robot research.

Okay, with that rant over, and everything else as it is, you know, without making sweeping changes to the system, the Logistics network technology should require Science pack 3, at the latest! maybe even just red and green like most of the rest of it.

My logistics mod does make the change to science packs 1, 2 and 3, but if that's the only thing you're after, then there is a mod out there somewhere that makes only this one change, I'm just not sure what it's called.

[dragontamer5788]

In particular, there's a logical progression of higher-and-higher sciences before you get to the "ultimate" power: Logistic Bots. You start off with slow Yellow Belts, then upgrade to Red, then upgrade to Blue, then upgrade to Bots eventually. There's nothing wrong with that.

There's the key point - that's YOUR particular playstyle, not mine.

Multi-lane blue belt busses for bulk transfer for me are the ultimate goal - and the logistics system is just a very useful convenience feature for things that don't even need half a yellow belt. It removes the tedium of manual crafting of things, and avoids the need for complex messy belt systems for things that only need an occasionally active assembler. That's how I use it, and I suspect many others posting in this thread asking for the old research level.

And I've suggested a solution, which is what I think is the developer's intent: Use Combinators and Mixed Belts / Sushi Belts for that situation.

As long as an area requires low-throughput, using Combinators to marshal delivery in the early game for convenience isn't a major issue. Again, the primary drawback to using Combinators for this purpose is the tedium of wiring all the logic. But since Blueprints are now free, and now that Construction Bots are Green Science... it is more than possible to set up combinator-based "convenient stores" off of a mixed-belt in the early game.

And if so, do you have any suggestions on changes that could be made to allow the logistics system from the 0.14 blue science level, but stop it being the "ultimate" power until, say, you've launched a rocket and are on infinite research?

Not really. The main benefit of logistic robots is the trivialization of layout problems. Requester Chests negate all layout problems for the rest of the game, as long as you have enough logistic bots to shuttle items around. I think the 0.15 approach where Requester Chests (the most powerful feature of the logistic network) is locked away in Yellow Science is the best solution.

Oh, and just to address this point - you're forgetting one thing. I'm talking about train->buffer chest->belt unloading without mods, so, as you pointed out, the limiting factor ends up being the buffer->belt rate, not train->buffer. So given it needs 3 or 4 platforms to unload at a rate around a single logistics station, you're looking at something like 20-30 seconds to unload each train.

I understand your point now. Yeah, that makes sense.

[jonatkins]

And I've suggested a solution, which is what I think is the developer's intent: Use Combinators and Mixed Belts / Sushi Belts for that situation.

Well, given the logistics system is far older than combinators (I checked old changelogs - logistics <= 0.6, combinators first in 0.12), I find it unlikely it was ever the original plan. I feel that logistics was never intended to be better than belts, but by overloading things with sheer numbers of bots + roboports players found ways of pushing them beyond their intended capabilities.

Changing the research requirements for this one part of the system seems so out of place, and a knee-jerk change from the devs to 'fix' this abuse. But it's ended up annoying a bunch of players like me who never even tried pushing the logistics system like that.

And if so, do you have any suggestions on changes that could be made to allow the logistics system from the 0.14 blue science level, but stop it being the "ultimate" power until, say, you've launched a rocket and are on infinite research?

Not really. The main benefit of logistic robots is the trivialization of layout problems. Requester Chests negate all layout problems for the rest of the game, as long as you have enough logistic bots to shuttle items around.

What you call the "trivialisation of layout problems" is the entire point of the logistics system. But only for low-volume production.

If just throwing sheer numbers of bots + roboports at it means the existing bot speed/capacity upgrades aren't doing enough to limit their capabilities then other things could be considered:

• Change logistics bot energy use when flying from being based on distance to being based on time. Of course, just throwing even more roboports to charge them would mean this isn't a complete solution on it's own.
• Increase the research cost of the speed + capacity upgrades, so they need purple+gold science sooner.
• Add a MK2 roboport, and make the MK1 much worse at charging. Alternatively, add research for charging time, and start it slower than it is now.
• Have a researchable limit on the number of active logistics bots at any one time. This would probably need to be a force-wide limit rather than per network, to prevent multiple small networks being a way to abuse the limit.
• Have a researchable limit on the number of active roboports allowed. Again, woudl probably need to be a total per force rather than per network.

The first two aren't complete solutions, but should be very quick and easy to implement in the current 0.15 if the devs decided. The third would also be easy to do, and combined with the first two could start to have a real impact.

It's too late to consider the last two options for the current 0.15 reoease, as it's just in bugfixing + tweaks stage now. However, while these would 'fix' the problem, it seems rather arbitary to have such limits.

As I've said, in the two games I've played through I ended up frustrated without requester chests, and ended up setting up single assembler production of the required science packs, etc and, as there was no space to belt things around, ended up spending a couple of hours manually ferrying items around until I had enough. So while tedious and not fun, the 0.15 change hasn't actually fixed anything. I then got lazy, and converted this temporary science production to use requester chests, and ended up researching things like MK2 power armor earlier than I would have otherwise done, and completely skipped using things like tanks.