1,000 Square Miles

[danconnors]

That's how big the Fallout 5 game world should be. With graphics as good as Fallout 4's. Of course I know the "experts " on this forum are going to immediately proclaim "Impossible". But the "experts" are wrong. Fallout 4 takes in the vicinity of 40 gigabytes of SSD area, and the standard size of an SSD now is one terabyte. At least that was the standard size when I bought my current computer almost four years ago.

My next computer won't have a standard hard drive, and I plan to put at least a couple of 2 terabyte SSD's inside it. Fallout 5 could be made as large as one terabyte, and only take half the room on one of those 2 terabyte drives. Every other game I have could fit on less than a fifth of the same SSD, leaving a completely empty 2 terabyte drive.

Fallout 4's game world I've read estimates of 25 to 50 square miles, and each estimator thumps his chest and snarls that anyone who disagrees with him is stark, raving mad. But I would say throw in Far Harbor and Nuka World, you have at least 40 square miles. Multiply 40 square miles by 25 and you get 1,000 square miles. That's a square over 31 miles on a side. Imagine having a game world you couldn't walk across in a full day of walking.

All we need is programmers who know what they actually have to work with, and do it. Computers are orders of magnitude more powerful than what we had when Fallout 3 slid off the boards. It's time we started getting game worlds of appropriate size for the machines they're housed in. You can spread encounters out so they're not in each others's laps and still have many times the encounter we're stuck with in 40 square miles.

Let's do it.

[Salamand3r-]

Originally posted by The Inept European:

1000 square miles is 31.6 miles square. Very close to 50km square. I just had to do that calculation for my own sanity.

The Far Harbor map looks like it's allegedly about 30-32 miles square.

Far Harbor should be about 2.755 square kilometres, about 20-30%-ish of the size of the FO4 map, which is approximately 9.90 square kilometres.

Edit - this is based on it taking 11 minutes at sprint speed to cross the main map, and sprint speed is generally realistic when measured realtime (measured versus an assumed player height of about 2 meters). However, this is realtime minutes. If you adjust those to the 30:1 compression ratio used by FO4, your character has taken 330 ingame minutes, or 5.5 ingame hours to make that run. Assuming an average human runspeed of 15 miles/hr, that means a map size of 82.5mi per side for a whopping 6,800ish square miles.

The problem is that the sprint speed makes sense in realtime (again based on the units per second you are moving), but not in relation to ingame time.

What I'm saying is that Fallout 4 basically doesn't follow our laws of physics, and it's difficult to get numbers that make sense.

[The Inept European]

Wasn't 30:1 for FO3 and they reduced to 20:1 for FO4?

To be honest I was basing this entirely off the map grid and assuming the map grid scale is the same in Far Harbor as on the main FO4 map. Entirely possible that's not true.

[The Inept European]

A 5 hr sprint at 15mph is kinda ridiculous though. That's why I'm using the jog mode and assuming amateur marathon speeds of 6mph. But the numbers work out similar. I get the main map to be about 25-28 mi wide on my assumptions.

[Salamand3r-]

Originally posted by The Inept European:

A 5 hr sprint at 15mph is kinda ridiculous though. That's why I'm using the jog mode and assuming amateur marathon speeds of 6mph. But the numbers work out similar. I get the main map to be about 25-28 mi wide on my assumptions.

Yeah, but to reckon the total distance, the speed doesn't need to be realistically achievable, it just needs to be known and constant.

Since it takes 11 minutes at an assumed speed of 15mph, we are looking at a map that is 2.75mi on a side MAX.

[xybolt]

Originally posted by The Inept European:

A 5 hr sprint at 15mph is kinda ridiculous though. That's why I'm using the jog mode and assuming amateur marathon speeds of 6mph. But the numbers work out similar. I get the main map to be about 25-28 mi wide on my assumptions.

sorry for nitpick, but when running, you aren't having a steady distance per each step you have done. Even a simple walk does not grant to have a constant distance per step.

If you don't understand me, well, find a settlement with a small slope. Build a straight path of wooden floors near that slope. Walk on it; from its start to the end. Count your steps. Then walk on a little slope uphill along the path you just have built. You will see that you get bounced back for each step you have, leading to more steps you need to perform to reach the same distance.

And going downhill "glides" your character a bit downwards.

It is just hard to find a reliable means to figure out the distance of the landscape directly in game, unless you go in the engine and use the coordinates of the world space. There are other odd things that I see people doing, such as scaling out common items (like a door...) You can do this and give a rough estimation, yet don't spend too much time on that.

[hawkeye]

Originally posted by Salamand3r:

Originally posted by The Inept European:

A 5 hr sprint at 15mph is kinda ridiculous though. That's why I'm using the jog mode and assuming amateur marathon speeds of 6mph. But the numbers work out similar. I get the main map to be about 25-28 mi wide on my assumptions.

Yeah, but to reckon the total distance, the speed doesn't need to be realistically achievable, it just needs to be known and constant.

Since it takes 11 minutes at an assumed speed of 15mph, we are looking at a map that is 2.75mi on a side MAX.

So that means I should be able to shoot enemies across half the map away.

[The Inept European]

Originally posted by Salamand3r:

Originally posted by The Inept European:

A 5 hr sprint at 15mph is kinda ridiculous though. That's why I'm using the jog mode and assuming amateur marathon speeds of 6mph. But the numbers work out similar. I get the main map to be about 25-28 mi wide on my assumptions.

Yeah, but to reckon the total distance, the speed doesn't need to be realistically achievable, it just needs to be known and constant.

Since it takes 11 minutes at an assumed speed of 15mph, we are looking at a map that is 2.75mi on a side MAX.

Agreed. My estimate based on the analogous method to yours is 1.25 mi on a side. But same ball park.

Tbh you're right that if it's only 12 minutes you can look at faster speeds. But I don't think it's logical to use the faster speeds but also scale up the distance by the time compression factor. If you see what I mean?

[xybolt]

Originally posted by hawkeye:

Originally posted by Salamand3r:

Since it takes 11 minutes at an assumed speed of 15mph, we are looking at a map that is 2.75mi on a side MAX.

So that means I should be able to shoot enemies across half the map away.

If those cells got loaded, then maybe. I am not sure what the engine does with a bullet that gets shot to afar. It is possible to shoot a NPC in an adjacent cell though (yank up ugridstoload to spawn npcs there as well). But past that cell? Not sure. Never tried to load more cells. I don't want to torture the engine :D

[The Inept European]

I'm assuming that when you run across the whole map, the uphills even out the downhills and you get a reasonably usable average.

[Sabaithal]

Originally posted by xybolt:

Originally posted by hawkeye:

So that means I should be able to shoot enemies across half the map away.

If those cells got loaded, then maybe. I am not sure what the engine does with a bullet that gets shot to afar. It is possible to shoot a NPC in an adjacent cell though (yank up ugridstoload to spawn npcs there as well). But past that cell? Not sure.

I've tried loading more than the default number of cells in Skyrim and the result was...well massive FPS drop. Not sure how Fallout 4 would handle it.

[Salamand3r-]

Originally posted by The Inept European:

Originally posted by Salamand3r:

Yeah, but to reckon the total distance, the speed doesn't need to be realistically achievable, it just needs to be known and constant.

Since it takes 11 minutes at an assumed speed of 15mph, we are looking at a map that is 2.75mi on a side MAX.

Agreed. My estimate based on the analogous method to yours is 1.25 mi on a side. But same ball park.

Tbh you're right that if it's only 12 minutes you can look at faster speeds. But I don't think it's logical to use the faster speeds but also scale up the distance by the time compression factor. If you see what I mean?

The thing is, the movement speed in game doesn't make sense if you apply the time compression - we can say the movement speed is accurate by comparing it to a known assumption like player height of ~2 meters.

The ingame movement speed makes sense with real time, but not with game time.

[The Inept European]

Would be interesting to find out whether weapon distances are scaled (down) using the timescale compression or are naturalistic and match what you actually see.

[Salamand3r-]

Originally posted by The Inept European:

I'm assuming that when you run across the whole map, the uphills even out the downhills and you get a reasonably usable average.

You use tcl make the run a hundred feet in the air, using the autorun key and sprint. No geometry in your way. Use tai and tcai to make sure no encounters happen.

Originally posted by Sabaithal:

Originally posted by xybolt:

If those cells got loaded, then maybe. I am not sure what the engine does with a bullet that gets shot to afar. It is possible to shoot a NPC in an adjacent cell though (yank up ugridstoload to spawn npcs there as well). But past that cell? Not sure.

I've tried loading more than the default number of cells in Skyrim and the result was...well massive FPS drop. Not sure how Fallout 4 would handle it.

Same thing, although it depends on the area. You can rebake some previs to effectively fix it, but you'd have to rebuild the entire map yourself in the CK.

Originally posted by hawkeye:

Originally posted by Salamand3r:

Yeah, but to reckon the total distance, the speed doesn't need to be realistically achievable, it just needs to be known and constant.

Since it takes 11 minutes at an assumed speed of 15mph, we are looking at a map that is 2.75mi on a side MAX.

So that means I should be able to shoot enemies across half the map away.

If the weapon ranges were anything near realistic, yes.

[The Inept European]

Originally posted by Salamand3r:

The thing is, the movement speed in game doesn't make sense if you apply the time compression - we can say the movement speed is accurate by comparing it to a known assumption like player height of ~2 meters.

The ingame movement speed makes sense with real time, but not with game time.

I agree, though there are different ways of resolving (or approaching) that problem. None of them perfect of course.

[The Inept European]

Originally posted by Salamand3r:

Originally posted by The Inept European:

I'm assuming that when you run across the whole map, the uphills even out the downhills and you get a reasonably usable average.

You use tcl make the run a hundred feet in the air, using the autorun key and sprint. No geometry in your way. Use tai and tcai to make sure no encounters happen.

Well, go for it, but I doubt the result will be materially different. I'm looking for order of magnitude, not precision. The precise value doesn't matter.