Research: Effects of Roads

[therother]

frogbeastegg has asked for a thread to investigate the possible effects of roads, specifically to do with trade (post 1.2 patch) and movement rates, although any other feature is also relevant.

For those interested, the effect of roads on trade before the 1.2 patch is detailed in the following Guides & Ludus Magna threads. AFAIK, these figures are still to be confirmed for 1.2.

[frogbeastegg]

Thanks for opening this thread

I'll keep this very brief; I haven’t got much time just now and I’m pretty exhausted to boot, and therefore probably not too coherent. This is both research for the beginner's guide and an attempt to satisfy my own curiosity.

In 1.1 roads and trade worked like this:

Land trading is automatically established, via the road system, between neighbouring settlements. For trade to occur between settlements, they must have a common land border. Improving these routes past normal roads will increase income.

• The first level of roads adds nothing to land trade.
• Paved roads always doubles land trade,
• Highway increases it by a further 50%

So if you have a land trade route worth 100 Denarii with no roads, with will be worth:

Dirt	Roads	Paved	High
100	100	200	300

I want to know if this is still true in 1.2. I'm dyslexic and the kind of maths that gets these results is beyond me, so I'm asking for help. Actually about all I can do is take the results and copy them to the beginner's guide with credit given :embarrassed:



Part 2 of this topic deals with roads and movement for units and armies. I want to know how much of a boost each road type gives, and if all units receive the same boost. I suspect all army unit types (infantry, cavalry, artillery) will gain the same, and I expect agents will gain the same boosts as armies but you never know. Cavalry, for instance, has a larger base movement stat than artillery and if roads if a percentage based bonus then cavalry will gain more from them than slower units.

This part I can and will be working on myself, although quite how I am going to research it I have not yet decided. The campaign map does not obviously display its squares, so measuring is going to be tough.

[zhuge]

One minor issue here on the economics part. Quoting Quietus's economic guide (https://forums.totalwar.org/vb/showthread.php?t=40072):

"Highways: 50% land trade boost from Paved Road in originating city only. In my test, only 3 of 5 got the increase, the two lowest values didn’t increase. There maybe a) a distance limit how far a highway upgrade is effective (the two farthers didn’t get a boost) or b) only a fraction of the land trades get a boost (3 of the 5 were the highest values). Neither sea exports nor imports were affected by the upgrade. Currently, I’m leaning towards choice a)."


I made 1 check BEFORE patch 1.2 and it seemed to be true. IIRC, 3 out of 5 possible provinces adjacent to Patavium got the boost. Rather risky to make conclusions based on 1 observation though.
However since Quietus and I have essentially observed the same odd phenomenon, I would think that it is worth looking into. They might have 'corrected' it for the patch or perhaps it's still the way it is.

[therother]

Ah yes, I'd forgotten about that effect. IIRC, it's due to the other provinces not having a comparable level of roads. So to get the bonus from paved roads, the neighbouring provinces need at least basic roads. Similarly, to get the bonus from Highways, the neighbouring provinces need paved roads. As with zhuge, I haven't had time to check in 1.2 though. (*)

IIRC, there are also a few other trade income idiosyncrasies to do with Governor's buildings, but that's off this particular topic.

(*) Edit: Just did a very cursorily check with 1.2, using a 1.1 game, and it does seem to still be the case. Also, based on this quick check, the effect of roads on trade income does seem to be unchanged, although I recommend more testing. I do not currently have the time, I'm afraid.

[therother]

The campaign map does not obviously display its squares, so measuring is going to be tough.

There are two ways that I know of to count the tiles. The first is probably the best, especially over long distances, and certainly the most reliable. Place your cursor on the spot you want to get info on, and then enter the command show_cursorstat into RomeShell. The output should be something like this (cursor is over Tarentum):

pos 114,63 region id 63 (Apulia)

This means that the tile your cursor is over is 114 from the left and 63 from the bottom of the screen. You can measure distances this way. Also, the show_cursorstat can be used as many times as you like, unlike show other RomeShell commands.

The 2nd method is a little fiddlier. If you click and hold the right mouse button on an unoccupied map tile, an information box will pop up. It should show the type of terrain, the region name, and the region owner (via faction symbol and colour). Now if you move the cursor a little, with the mouse button still held down, the info box should move. Everytime the box moves symbolises a transition between tiles. Using this, you can count the number of squares. Note: this will only work on land, which shouldn't be a problem for your purposes.

[frogbeastegg]

The second method sounds like the safest one for a dyslexic frog, so I shall try that on a part of the map with a relatively straight stretch of road. The test will probably have to wait until the weekend though.

[therother]

I did a quick bit of research on your second topic. There seems to be some complications, but hopefully this will help you nonetheless:

Distance in tiles for different road types

This is the data from one test. I tested the distance each type of unit could travel along the vertical road from Bostra to its port.

I know, because I quickly checked, that there is some variance here, as I got 17 tiles for Calvary on Paved roads at some positions. I didn't find any difference between 1.1 and 1.2, although I only checked quickly. Also, I didn't notice any differences between the terrain types, although again I didn't check very thoroughly.

I don't see any obvious pattern:

% Increase from dirt

Progressive increase

Anyway, hope that is of some use...

[frogbeastegg]

Hmmm, thanks. Most interesting. My own tests were delayed until today (I haven't done highways yet though) but my findings for no roads, simple roads and paved roads mostly match yours, except my set is not nearly so neat looking. The discrepancy comes with paved roads; I found they offered a slightly larger boost than simple roads; 4 tiles instead of your 3. I suspect this may be down to the variance effect you yourself mentioned, and I would trust your set of results over mine simply because I know how poor at this kind of thing I am.

So ... highways are useful, but not really worth crying over if you are playing a non-Roman faction. They certainly aren't such a high priority item as the first two road types, as they cost a significant amount but the increase they provide is not so dramatic. Paved roads and simple roads deserve to be high priority items, IMO, but highways are now slipping down to a mid level priority, depending on location, trade income and the need for fast moving armies.

[frogbeastegg]

Highways have now been tested, and again we have a concordance between our findings. So I suppose that is it, job done. I do wonder what causes the variance in movement though, but I have no real ideas on how to investiagte that further. It's also not beginner's guide material, or at least not for the main 1.2 update, and the guide update has to be my main RTW priority now. Maybe in the future I will be able to come back to this.

Anyway, thanks for your help

[therother]

I was going to research a complementary area, that of the effect of terrain on movement, but it quickly became apparent to me that there was little or no effect. My plan was then to bring this data together with the data for the effect of roads, to see if a desert road allowed the same level of movement as a road through hills. I've still to check what effect height has on movement - it could be responsible for the variance. Hopefully I can get some time soon.

So ... highways are useful, but not really worth crying over if you are playing a non-Roman faction. They certainly aren't such a high priority item as the first two road types, as they cost a significant amount but the increase they provide is not so dramatic. Paved roads and simple roads deserve to be high priority items, IMO, but highways are now slipping down to a mid level priority, depending on location, trade income and the need for fast moving armies.

Good analysis - I agree with your conclusions.

[Suladan]

Thanks, both of you. Your study was informative and very useful.

[Oaty]

Slaves need roads to be traded. So early in the game when you want that enslavement growth bonus you need to have a road going out of that town for the bonus. even when the town shows .5 percent bonus pop growth it will not get that bonus until you place that road.

As far as highways being profitable it definately has to do with distance. Lepcis magna can make a ludicrous profit from highways, whereas Arretium with highways does not get that great of trade bonus from them.

[Simetrical]

Are you sure that's the reason, oaty? I seem to recall somebody else hypothesizing that it might have to do with how good neighboring roads are.

-Simetrical

[frogbeastegg]

Thanks, Oaty. The information about slaves and roads should definitely be in the guide; if it isn't now it will be after the update. Very important little detail.

Distance and neighbouring roads! Another important facet to investigate, and one which tweaks my curiosity something rotten. I shall have to come back to this later.

therother, those are some very good ideas to poke about with; if nothing else it will rule those possibilities out and inform us that the variance comes from something less logical.

Based on real life, and on how most other games handle movement, I would expect a slightly higher movement cost for hills, mountains, deserts and forests. Grasslands and steppe should be either easy or average to navigate. However as each road has a predefined position on the map which avoids running right through forests, instead using little clearings, this might be something that requires no roads at all for a good test. Also certain units are logically more likely to have difficulties with certain terrain types, for example infantry should do better in forests than artillery. Infantry would be a good basic all rounder for these tests; there is no particular terrain they should have undue difficulty with.

[Simetrical]

I would expect a vastly higher movement cost for mountains, deserts, and thick woods, actually, along with snow (although not in that order).

-Simetrical

[therother]

Hmm, in the midst of a bit of testing at the moment, and a few questions have been raised.

Firstly, the range of agents on fertile ground is approximately double that reported above (the increase for the others is much smaller); this results in agents being the fastest ground moving units in 1.2, instead of the slowest after artillery. I'm not sure why this should be so different to the wilderness test I did on a modified 1.1 game. This time I'm using a modded 1.2 campaign map, based on Duke John's minimal map, instead of setting things up with a 1.1 game and then converting to 1.2. That may be the reason - I don't know.

Second thing is that there are definitely movement modifiers for certain terrain types. I suspect that I couldn't detect them before because the effect for the terrain types I tested was small. This time I used whole blocks of each particular terrain type, i.e. the unit moves only over each terrain, instead of only over a couple of squares. Anyway, I've tested hills, sparse forest, and swamps. The movement modifiers appear to be around 0.8, 0.8 and 0.5 respectively, i.e. if you can move 10 tiles on fertile land, you can only move 5 through a swamp.

Still to test wilderness, and the effect of various roads. Except that roads through small and large mountains, dense forest, and all sea and ocean types are impossible (movement modifier = 0). Also, and I suppose this is only of interest to modders, there does seem to be a limit on the length of roads, and this limit takes into account the terrain the road would have to pass. For instance, my test map has a 25-tile road. When I mod in 10 tiles of swamp, the road disappears – it is no longer displayed. This was not true of either hills or sparse forest.

Preliminary tests of height seem to indicate little effect, but I have said that before...

Will post full data later (perhaps tomorrow), hopefully with road data as well.

[Morat]

Any update on this, therother?

[therother]

Any update on this, therother?

Yes, sorry. Got a little sidetracked with the traits. Will post my results soon.

[therother]

Still a little sidetracked. I'll get back to it at the weekend.

[Quietus]

I'd just like to share this since you guys have been measuring movements. In case it is still not known yet, there is a cavalry movement bug.

If you click on the cavalry's Unit Card, it receives longer movement range as opposed to just clicking on the unit in the the actual strategic screen.

I'll do some measurements of the difference later.

[therother]

Just a clarification: do you mean that if a unit of cavalry is on its own in a stack, that it will appear to have a reduced range if you select the unit figure on the strat map than if you select the unit card? If so, then that is strange, as it's not been my experience. I have, however, seen that if you select a general with infantry and artillery, say, that the slowest member will determine the range.

[Quietus]

No, when the cavalry is separate from the stack. It it possible though that it is opposite (it receives less when you click on a cavalry unit outright).

Have a single unit of cavalry (or an all-cavalry stack for that matter). Now click on the stack (you get the range). Now, click on the unit cards (or select all), the range is bigger.

It doesn't happen all the time though. I'll see if the difference of roads in provinces have something to do with (but not likely, from initial observations).

[Phase]

You don't happen to have a general with a movement bonus (trait or follower) and so get longer range from select all?

[Quietus]

You don't happen to have a general with a movement bonus (trait or follower) and so get longer range from select all?

No. The cavalries are separate from any general, infantries or others. All you have to do is click on the unit on the screen. Then click the same unit on the unit card, you get an extra movement range.

Or otherwise, the cavalry is getting less that what it should and clicking on the unit card confers the correct movement value. It doesn't happen on infantry though.

[Quietus]

About the 3 of 5 getting the road increase, I believe this is due to the other provinces not having a comparable level of roads. To get the bonus from paved roads, the neighbouring provinces need at least basic roads. Similarly, to get the bonus from Highways, the neighbouring provinces need paved roads.

(Resuming from the guide discussion). Ok, I did 4 tests for this. One checked ok, with three inconclusive thx to rebels and enemies (they are always on the roads that I'm starting to think building structures such as roads and trade buildings trigger appearances of rebels ).

I'm also out of save game options.

I just need highway vs. highways data or confirmation, preferably 4-5 land trade partners to confirm (and they should get all the bonus).

Thx

[Morat]

From my experience therother is right about the roads needing comparable levels to get the added benefit.

How's the research in this area coming along anyhow? This and the trade thread are areas of interest for me.

[Severous]

Hi

Hope its OK to bump this old thread. I have some observations from my current campaign that are on topic...just a bit separated from the previous remarks.

RTW V1.5 No Mods. H/H Difficulty.

Thessolonica was enslaved. It is exporting Slaves to several adjacent regions.

Byzantium was occupied. So it doesnt have slaves and imports slaves via land from Thessolonica.

Thessolonica has dirt roads. Byzantium has none.

Income in both cities from this one route is 14. This rose to 15 as towns grew a bit. Its just slaves. No other resource is traded between the two cities.

Thessolonica constructed Paved Roads.
Its land income doubles on all other routes. But not Byzantium. Both cities continue to receive 15 income from the slave trade.Incidentally building a paved road did not increase Thessolonicas sea trade income.
Edit: Image added. Showing paved roads in Thessolonica but no roads in Byzantium



Byzantium then built dirt roads.
Byzantiums income from the slave trade with Thessolonica remained unchanged at 15. Thessolonica's income from this same route doubled to 32 (assume a fraction rounding up)

[ThePianist]

I am very impressed by reading the studies (and even equations) of RTW pros like you, and it makes RTW even more interesting. I have a question, since I read here that building roads from a city not only increases the income of the city, but also the income of the provinces around it. The question is, and I wondered about that before, in the description of the market/agora type buildings, it says that it increases the number of economic items available for trade. When you click on building a market/agora, and look in the city details, you'd see an expected increase in city income. Since it adds more item for trade, does building a market/agora in one province also increases the income of the nearby provinces? (in the same way as building roads)

[Ludens]

This is probably the wrong thread for the question. Have a look at this one. With sea trade, the importing city gains 20% of the value of the rout to the exporting city. I guess it works the same with land trade.