Infrastructure-Influence on Movement

[unmerged(12544)]

.. the speed displayed is wrong when you don't have starting modifiers for 100% (100 infra, 100 supply, plains, 100 ESE).

It is lacking 4 factors, which do NOT reduce speed, but prolong the virtual travel distance or reduce the effective speed both with the same result of a longer movetime.

What makes this so nice, is the minimum speed which seems to be 0,5 mph (135/269) no matter which unit type.

Can be lower, see my values above. But my values are theoretic ->0 org, 0 Infra, blizzard, no oil supply, etc.; simply the worst possible case.

the terrain modifier is applied twice when moving.

It is, once from modifiers.csv AND(!) from ...\map\terrain types.csv.

I think I will stop to dig around in the movement logics since I can't find them.

The first I did, is set all values in "terrain types.csv" to 1.0, this is a good starting point. Then you can use my formula, it represents 3 days of work. What you need to get the distance is: try to get perfect conditions, no CoS with school of manoeuvre, plains and at first let your unit march towards the destination, then calculate ETA - now = movetime. Then under those conditions the ONLY factor missing in effective speed is infra. Example: MT = 20 hours, infra (destination) = 50%, displayed speed = 5 mph

infra factor = 1.35
MT = distance / speed * infra
distance = MT / infra * speed
distance = 20 / 1.35 * 5 = 74 miles
effective speed = speed / infra = 3.70 mph

I would not use air ranges. I think my method is quite precise. After you have distance, you can do all kinds of calculations.

 

[Don E]

This is not the general cap for anything, but for this specific setup. As Panther II already mentioned there is an overall cap for the movement time multiplier at 5,5. This multiplier includes all modifiers (like terrain, unit type, weather, ESE, infrastructure and oil using). As soon as the multiplier reaches 5,5 it is capped. So an oil using unit will never take longer than 5,5-times the time than a compareable non-oil user will take and in many cases less.

There seems to be something wrong with your model here, or perhaps it has just not been put to the most extreme cases.

Movement from Krivoy Rog (100% infra) to Dneprovitovsk (0% infra):
Base move 8, +10% minister mod, frozen weather, clear terrain.
Motorized: 175 hours
Cavalry: 26 hours
175/26 = 6,73

This value is probably a result of the rounding during calculations, as opposed to my guesstimate before testing of 6.6. This would be the Cav speed of 6.6 against the Mot speed of 1.0, as this appears to be the minimum speed no matter how bad conditions are.

 

[Don E]

the terrain modifier is applied twice when moving.

It is, once from modifiers.csv AND(!) from ...\map\terrain types.csv.

I noticed this too when my calculations in mountains never seemed to work out. This is a sneaky one, as the movement modifier for Mot in mountain in reality is -75%, not -50% as listed. I can't really see why Paradox made it this way, as it throws of their listed modifiers and confuse the players. Me at least .

 

[Mediator]

There seems to be something wrong with your model here, or perhaps it has just not been put to the most extreme cases.

Movement from Krivoy Rog (100% infra) to Dneprovitovsk (0% infra):
Base move 8, +10% minister mod, frozen weather, clear terrain.
Motorized: 175 hours
Cavalry: 26 hours
175/26 = 6,73

This value is probably a result of the rounding during calculations, as opposed to my guesstimate before testing of 6.6. This would be the Cav speed of 6.6 against the Mot speed of 1.0, as this appears to be the minimum speed no matter how bad conditions are.

As I already stated in my very last post, I was wrong on this. There is no multiplier which is capped, there is a minimum speed. Btw. Cavalry has speed 7 and basic motorized has speed 8.

 

[Don E]

As I already stated in my very last post, I was wrong on this. There is no multiplier which is capped, there is a minimum speed. Btw. Cavalry has speed 7 and basic motorized has speed 8.

That's what I get for spending too long writing my posts. Appologies, next time I'll make a check for new posts before submitting.

The speed was easy to edit in the unit files. That way there was only one variable, oil using vs. non-oil using. Vanilla Cav would have a effective move of 5.8 (5.775) in the above setup.

Cheers,
E

 

[Don E]

Ok, I have looked over all my results, and have come to the conculsion that both Mediator and me were correct in our calculations, although in slightly different ways. My speed calcs were correct, and Mediator's calculations on infrastructure effect were rather a modification of travel distance than of speed. In the end one have to apply both to get the correct travel time.

Travel Time (TT):
TT = Effective Distance / Effective Speed

Effective Speed (ES):
ES = Base Speed * Minister Mod * Terrain Mod * Weather Mod * ESE * Target Infra

1) Terrain Mod is not only the listed mod from 'modifiers.csv', but divided by the value from 'terrain types.csv' as well. E.g. Mot in mountains = (1 - 50%)/2 = 25%
2) Target Infra is only used for oil consuming units.
3) The ES has a minmum of 1mph
4) There does not appear to be any effect from the nighttime movement modifier.

Effective Distance (ED):
ED = Base Distance * 1.24 * (1 + 0.7*(1 - Target Infra))

1) Base Distance is the straight distance between the XY coordinates of the army icon in the start and target province.
2) The 1.24 factor is empirical based on the experimentation with provinces in the Dneprovitovsk area. This might vary in other areas, and I hope soembody could verify this.
3) The last term is the same as from Mediator's formula.
4) Entering an enemy province adds another 25%
5) Crossing a river adds another 25%

I'm starting to get confused by my own numbers now, but I think this should be close to the correct formula for travel time. I have not tested it extensively in all types of terrain and regions, but it has so far been correct where tried. Hope somebody can give feedback on this.

Cheers,
E

 

[Ramses]

A few facts I remember

Infantry in 1914-1918 actually WALKED through France. So did my grandfather in 1940 from Strasburg to Metz.

Tactical movement differs from strategic redeployment, that allows for higher speeds when the "actual" move is too bad. You trade speed for fighting ability and vice-versa.

Remember Ferdinan, the heavy gun stuck in Russia.

Remember the north African soldiers of the French Army, solving the Monte Cassino trouble by WALKING through mountains with only passes. In fact trucks were not only absent there, they were useless.

Such facts give me the feeling the spirit is OK in HOI2, even if the figures are a little overestimated.

Ramses

 

[unmerged(12544)]

... movement modifier for Mot in mountain in reality is -75%, not -50% as listed.

I was confused by this, too. The value is for speed, not time. So speed is -50%, time can be double.

ES = Base Speed * Minister Mod * Terrain Mod * Weather Mod * ESE * Target Infra

This is not complete. E.g. Org influence is missing. Again, look here for more details
http://www.europa-universalis.com/forum/showthread.php?t=191553

2) Target Infra is only used for oil consuming units.

Depends on what you define as "Target Infra".

3) The ES has a minmum of 1mph

This wrong. ES is much lower.

4) There does not appear to be any effect from the nighttime movement modifier.

Yes.

ED = Base Distance * 1.24 * (1 + 0.7*(1 - Target Infra))

In this formula is also a lot missing.

2) The 1.24 factor is empirical based on the experimentation with provinces in the Dneprovitovsk area. This might vary in other areas, and I hope soembody could verify this.

It is the result of several factors multiplied on each other.

4) Entering an enemy province adds another 25%
5) Crossing a river adds another 25%

Exactly. Don E, if you have an Example, you want to calculate, you can present it here or in the other thread.

 

[unmerged(18738)]

I noticed this too when my calculations in mountains never seemed to work out. This is a sneaky one, as the movement modifier for Mot in mountain in reality is -75%, not -50% as listed. I can't really see why Paradox made it this way, as it throws of their listed modifiers and confuse the players. Me at least .

First, I am really glad I am following this thread and that you are all working so hard on this. I think how Paradox deals with this will say something strong about their commitment to the community of customers.

Anyhow, this thread makes me realize why Mithel (Starfire) adjusted the terrain penalties in TerrainTypes.csv all to _ONE_. That way you only get hammered by the modifiers.csv penalites. So that at least takes 1/2 of the penalties out, now if they can only deal with the oil penalty.

 

[Don E]

This is not complete. E.g. Org influence is missing.

Yes, the zero org was an omission. Can't see much else there is to it though. Unfortunately lack of oil is something that in my model falls into the distance category, as it isn't affected by the minimum speed.

Again, look here for more details
http://www.europa-universalis.com/forum/showthread.php?t=191553

Sigh, another thread to look up :wacko: Why did you post it on the mod forum, well hidded from my sore eyes?

Depends on what you define as "Target Infra".

The infrastructure value of the province the unit is moving to.

This wrong. ES is much lower.

I have yet to find a case where these formula give anything lower. If you insist on lumping all the mods onto the speed as opposed to the distance moved this is of course correct, but I see no reason why. It actually makes sense that a unit have to walk a bit longer in a province where the roads are not ruler straight, but that it can still move with the same speed.

It is the result of several factors multiplied on each other.

I'd love to find out which, as a random fudge factor always leaves me a bit sceptical. I'm inclined to believe it is a conversion between the pixel scale of the map and the real world mileage.

Exactly. Don E, if you have an Example, you want to calculate, you can present it here or in the other thread.

Try my usual setup from Kirvoy Rog(#236) to Dneprovitovsk(#235), both 100% Infra. Modify Motorized 1 to have move 1, and make sure there is no School of Manoeuvre minister. No matter what terrain and weather you put in now, from Clear in Plains to Blizzard in Mountains, you end up with the same travel time, 102 hours. This points in the direction of a speed minimum of 1. Change the target infra and your path will wind longer (or the effective speed will be reduced in your terminology).

I don't think there is any real difference to our formulas, but you have all the modifiers lumped in with speed rather than distance. I guess distance might not be the best choice, as some factors that clearly doesn't affect the distance have to be lumped into the bracket. It is perhaps easist to call it 'shown' and 'hidden' modifiers, with the shown speed being subject to a 1mph lower limit.
I would guess your 0.376 speed limit should really be 0.375, and comes from 1/1.7/1.25/1.25, and the 0.125 speed limit comes from this divided by the no oil penalty of 3.

The one error I can find in your formula is that you have lumped the terrain and weather mods into one factor. I've found this to be wrong, especially since a blizzard in bad terrain would give negative move.

Cheers,
E

 

[unmerged(12544)]

ES = Base Speed * Minister Mod * Terrain Mod * Weather Mod * ESE * Target Infra
Yes, the zero org was an omission. Can't see much else there is to it though. Unfortunately lack of oil is something that in my model falls into the distance category, as it isn't affected by the minimum speed.

The oil penalty is missing for oil consuming units.

I have yet to find a case where these formula give anything lower.
I would guess your 0.376 speed limit should really be 0.375, and comes from 1/1.7/1.25/1.25, and the 0.125 speed limit comes from this divided by the no oil penalty of 3.

They are not part of DISTANCE or SPEED:
* INFRA_F * CONTROL_F * RIVER_F * OIL_F

so
INFRA_F <= 1.7
CONTROL_F = 1.25
RIVER_F = 1.25
OIL_F = 3.0

can play a role.
Effective Min. speed for non oil con. = 1.0 mph / 1.7 / 1.25 / 1.25 = 0.37647
Effective Min. speed for oil con. = 1.0 mph / 1.7 / 1.25 / 1.25 / 3 = 0.12549

until we discover further factors, then it might be even lower.

If you insist on lumping all the mods onto the speed as opposed to the distance moved this is of course correct, but I see no reason why. It actually makes sense that a unit have to walk a bit longer in a province where the roads are not ruler straight, but that it can still move with the same speed.

It is hard to believe for me, that a distance between point A and point B gets longer, because you move into enemy territory or infra is low. I would rather imagine, that speed goes down, because of this. I don't insist on that, it is just my thought.

I'd love to find out which, as a random fudge factor always leaves me a bit sceptical. I'm inclined to believe it is a conversion between the pixel scale of the map and the real world mileage.

I am quite sure, that there is a distance grid world wide, where all connected provinces have a distance to each other stored somewhere.

This points in the direction of a speed minimum of 1. Change the target infra and your path will wind longer (or the effective speed will be reduced in your terminology).

Yes, because INFRA_F always influences and is never caught by the minimum speed.

I guess distance might not be the best choice, as some factors that clearly doesn't affect the distance have to be lumped into the bracket. It is perhaps easist to call it 'shown' and 'hidden' modifiers, with the shown speed being subject to a 1mph lower limit.

With the bracket [] I tried to visualize, that this stands for SPEED and speed has the cap/limit. The other factors have no limit, so you can not put them into the bracket. My SPEED should exactly be the speed shown in game, which is exclusive those factors. They should be part of speed in my opinion.

The one error I can find in your formula is that you have lumped the terrain and weather mods into one factor. I've found this to be wrong, especially since a blizzard in bad terrain would give negative move.

You are right, I have to do some further testing on that. It might look like that

MODIFIER_F = 100 / (100 + weather_dest) + 100 / (100 + terrain_dest) - 1

 

[unmerged(22114)]

Actually the US had GREAT succes with using tanks/APCs in vietnam even in jungle areas.
I remember reading an article in Armor Magazine(From US Army: Armor School)
where the procedures and tactics from an armored cav reg where discussed.
First they had great lethality by using canister rounds for tanks. Also the vehicles were not that unmanouvrable. They didn't have to slow down to watch for booby traps and so on. Of course this was mostly tactical mobility, which mostly has an effect on combat power, not moving province to province...
Bottomline of the article was that armor was usefull even in jungle areas...

That is, I must say, mighty interesting. Was this done large-scale as well or not, and if so, why not? Was it plain arrogance that they needn't use tanks or did they have a reason for not deploying more armored vehicles?

 

[unmerged(15260)]

That is, I must say, mighty interesting. Was this done large-scale as well or not, and if so, why not? Was it plain arrogance that they needn't use tanks or did they have a reason for not deploying more armored vehicles?

Well it was an Arm Cav reg which is about 3000-4000 men, so fairly large scale (must have been the 11 ACR). Of course infantry really have an edge in jungle, but walking on foot is so nasty in a jungle! I guess it also depends on what kind of jungle it is... And PROCEDURES!!!
Found these articles about armor in Nam:
http://www.knox.army.mil/center/ocoa/armormag/mj00/3suoitre00.pdf
http://www.knox.army.mil/center/ocoa/armormag/mj97/3vietnam97.pdf
And check out this tank from Nam: http://www.knox.army.mil/center/ocoa/armormag/mj98/3reaper.pdf

A quote from http://www.knox.army.mil/center/ocoa/armormag/mj98/5trends3.html
(Vietnam):
"The purpose of the operation was to attack enemy sanctuaries in Cambodia, which had been previously off limits. U.S. forces involved in the operation included 1st Cavalry Division (Air Assault), 25th Division, and the 11th ACR. Brilliant use of aviation and armor in mobile warfare led to success at the tactical level. Surprised enemy units were encircled and annihilated. Huge stocks of individual weapons, crew-served weapons, ammunition, and rice were captured. The penetrating forces overran an extensive logistics base with a fully equipped motor park, complete with grease racks and spare parts.4 The 11th ACR was assigned two additional engineer companies to handle all the added demolition work. By the end of the operation, almost 10,000 tons of materiel and food had been destroyed and over 11,000 enemy soldiers killed."

Even in WW2 tanks were used in jungles:
From: http://www.knox.army.mil/center/ocoa/armormag/ja97/4pacific.pdf (WW2)
"While prewar planning foresaw no
important role for armor in the jungles
of SWPA or any other Pacific Theater,
American soldiers discovered the necessity
of tank support for their numerous
offensives against the skillful defensive
tactics of the Japanese Army
even before the Wakde-Sarmi campaign.
Captain Richard J. Satran, commander
of Company A wrote: “The
success of the recent operations on
Wakde... has opened up a new and unexplored
field for tank warfare in the
Southwest Pacific Area.”

And an article about tanks in Hurtgen forest(germany):
http://www.knox.army.mil/center/ocoa/armormag/mj02/3huertgen02.pdf

 

[Don E]

The oil penalty is missing for oil consuming units.

That is the term I named Target Infra, as it is a multiplier equal to the infrastructure in the target province.

It is hard to believe for me, that a distance between point A and point B gets longer, because you move into enemy territory or infra is low. I would rather imagine, that speed goes down, because of this. I don't insist on that, it is just my thought.

I find it perfectly reasonable that you have to drive longer in a low infra province than a high infra. Just compare a province with mostly winding country lanes with a province of ruler straight autobahns. It is quite clear that in the former example you have to move a longer distance than in the latter example. In addition, a unit with trucks and tracks will have to move slower where the roads aren't as good.
Rivers also make some sense, as you have to make detours to find bridges and/or fords (though I think all the rivers shown are to big to be forded in most cases).
Same goes for enemy territory, you move around a bit more rather than going straight into the province. Minefields, prepared artillery zones etc. all have to be avoided. Roadblocks and blown up bridges would make for even greater detours. Could just as easily be translated into a speed reduction, but I dont think it is unreasonable to increase distance.
Where the terminology really breaks down is in the panelaty for lack of oil. One would have thought this should have been included in the speed calculations, just like the zero org penalty. Don't know why this is, perhaps an oversight, perhaps the developers wanted it to bypass the 1mph minimum speed.

In the end I don't think it matters if we call it an 'increase in distance' or 'reduction of effective speed'. It boils down to factors that aren't visible in the speed of the field command in the game rather than factors that reduce the listed speed. I guess we will never know if this has been on purpouse r rather an oversight by the developers, I'm just glad we finally managed to pin down the formula.

I am quite sure, that there is a distance grid world wide, where all connected provinces have a distance to each other stored somewhere.

If that is the case, I would be very surprised if it is, then the stored numbers apperas to simply be number calculated from the province. If you try it out you'll find that changing the xy coordinates for the army will not only change the spot on the map the unit is moving to, but also change the travel time, i.e. changed the distance. You should use sprites when testing this, as the counters use different coordinates, but the sprite coordinates for calculating distance.

Another interesting observation is that air units us a Fudge factor of 0.66 instead of 1.24 :wacko: .

Cheers,
E

 

[unmerged(22114)]

That "Grim Reaper" is hilarious. :rofl:

 

[unmerged(41819)]

That is, I must say, mighty interesting. Was this done large-scale as well or not, and if so, why not? Was it plain arrogance that they needn't use tanks or did they have a reason for not deploying more armored vehicles?

Tanks were not used on as large of a scale as they might have been because by the time it was discovered that they would be effective, the force structure in Vietnam was already set. some regeminatal and battalion sized units were brought in. the 11th ACR was the largest of these. initially it was expected that tanks would be largely immobile in the jungles and although they were restricted they determined after several years that they were more effective then originally thought, as well as there was more terrain that was suitable to them then originally thought. also keep in mind the Air Cav concept was fairly new at the time and was the main focus of the military especially for dealing with adverse terrain, so i believe that drew some focus away to testing armor in vietnam back in 65

 

[unmerged(12544)]

If that is the case, I would be very surprised if it is, then the stored numbers apperas to simply be number calculated from the province. If you try it out you'll find that changing the xy coordinates for the army will not only change the spot on the map the unit is moving to, but also change the travel time, i.e. changed the distance. You should use sprites when testing this, as the counters use different coordinates, but the sprite coordinates for calculating distance.

I did not know that. A very bad thing, what would mean, that Paradox was too lazy to do it with a grid and so better distances.

 

[Don E]

I did not know that. A very bad thing, what would mean, that Paradox was too lazy to do it with a grid and so better distances.

It doesn't bother me so much really. It makes it easy to guesstimate the distance from province to province based on where your units are in the province. What I think is much worse is the game engine's inability to correctly calculate the shortest distance. While I know it would take a lot of computing power to do so, examples of units in provinces with another province between them can end up with two different path, one for each direction, with one of them being substantially longer. It is really annoying to have to pause the game so one can find the shortest route for the panzer encirclement.

Cheers,
E

 

[unmerged(12544)]

I am also worried about map distortions. And a huge difference in movetime in e.g. France compared to Australia for "real" 1000 km in both cases. I have already read about someboby complaining about that.

 

[Don E]

I am also worried about map distortions. And a huge difference in movetime in e.g. France compared to Australia for "real" 1000 km in both cases. I have already read about someboby complaining about that.

That was why I was hoping the fudge factor was not constant across the globe, perhaps increasing or reducing the weight of the yy-distance on the map depending on the latitude. This apperas not to be the case, as the factor have been te same in all cases I have tested it.