Mobility quirk, speed differences (Bug?)

[ww2steel]

Hi,

Not that it makes an real difference, but I was running a ton of mobility tests and I noticed something weird... units on the northern end of the map move slower than on the southern end! This is a very small difference, but it's there.

I use a 3040m * 1040m (height * width) map for these tests. Exact layout is irrelevant, but take a terrain that takes a long time to get through, scattered trees for example (that's what I used for most of today's tests). Add 100 vehicles to the map and put a single enemy somewhere that the vehicles will not encounter it(such as a gun so it won't move). I was testing the Jpz I. Drive straight west to east.

{numbers vary very some on identical attempts}

Normal fastest time: (removing the 7 second command lag, remmeber that the beginning of turn 12 is 11 minutes into the game)

tanks 1-34: 11:09

tanks 35-(about)68: 11:28

tanks 69-100: 11:39

When you run this you will find that tanks 1 - 34 are fastest, 35-(about)68 are medium, and 70-100 are the slowest. I tried different map sizes, different total numbers of vehicles, different densities of vehicles, wind direction, going the opposite direction (W to E), flipping tank 1 to the northern end, and more. Many hours, 127 immobilized panzerjäger, and 1166+km later I finally am pretty sure I have discovered what it is.

The tanks within 1024 of the southern end of the map will be fastest, and for each 1040m the map extends to the north, the vehicles in that section will be about 2% slower. A 900m tall map for example would all have 'fastest' vehicles. If a tanks is 2500m up a 3000m map, it goes 'slowest'. I don't know if the vehicle speeds up if it crosses south of this magic boundary.

Really, who cares, because 2% on most terrain and short distances is irrelevant. This certainly shouldn't effect anyone's tactics. I merely post it for novelty. It's not even big enough for me to consider a bug, but I do have to take it into account when doing my mobility tests. The effect on immobilizations was inconclusive, as it takes me about ten minutes to get about ten immobilizations. The slower ones were slightly higher (fast to slow = 23, 23, 27), but this is certainly not very significant.

Sorry this is so boring, but I'm bored!

Mike

[Redwolf]

Great catch, man.

Do they also turn slower on the south edge?

[ww2steel]

Slower = north. I'll check rotation as soon as I finish these damn R-35s. It took them over 46 minutes to cross scat, deep mud. Now to count and calculate distance travelled by all of the stuck ones.

BRB,

Mike

[ww2steel]

Rotation seems the same, but I'm not sure I could detect a 4% difference. As it took them 2:16 to turn just under 180° I feel I could. As there is no command lag (and no associated variance) they all seemed to stop rotating at exactly the same instant.

For kicks I backed them all up. Got 94km of data. 12 Immobilizations, took the southern tanks 13:56 to cover 1km averaging 2.7mph (down from the 3.3mph 'fast'). I'll have to do MANY more tests to determine if there is a difference in the distance covered/ immobilization ratio. It seems from this one test that reverse immobilizes about 10% more often than fast but this is NOT statistically significant.

Mike

[Mike Summers]

I continue to be amazed at the patience some of you guys show in running these tests. That's some really, uh, interesting stuff WW2steel. Thanks.

[HarryInk]

Bless you for finding this out. We'll start a collection for your therapy presently...

[Ace Pilot]

I'll bet BFC took into account how the earth's acceleration due to gravity changes as you move from the equator to the poles.

[Kanonier Reichmann]

I believe that the Oxford Dictionary has now had to reconsider its definition of a gaming geek. It now simply reads: Gaming Geek- Refer ww2steel.

:eek:

Regards

Jim R.

[GreenAsJade]

How the frick does an effect like that even get _into_ software?

Many bugs after a while you can figure out how they happened (like the "reversing vehicles don't bog" bug). You can get in the head of the programmer and understand how it came to be.

But this one? Beats me!!!???

[ww2steel]

You guys love to pick on me. :cool:

I even considered that they were taking into account the smaller diameter of the earth taken at more northern latitudes, but that should make the northern vehicles faster (shorter distance travelled) not slower.

HUGE CORRECTION FROM SOMEONE ELSES PREVIOUS THREAD. I have said this before, and believe me, I've done lots of tests- REVERSING ABSOLUTELY DOES NOT PREVENT BOGGING/ IMMOBILIZING.

(Jpz I data on Deep Mud, Scattered Trees)

Orders: MPH: Avg, dist, between Imm:

Fast 3.3 9184m

Move 2.9 16282m

Hunt 2.9 16021m

Reverse 2.7 7190m

Obviously, the higher the distance between immobilizations, combined with higher speed = good off road. CLEARLY REVERSE SUCKS WORSE THAN ANY OTHER FORM. ALso, interestingly, in some faster terrains, Hunt can actually be FASTER than Move.

This has been tested on multiple armored vehicles.

Hope this helps,

Mike

[dieseltaylor]

Mike I think I have a vindication for you that I actually picked up from a company I am interested in. It is quite funny and strangely relevant:

"If you look at the presentation document used by Ray at the last AGM, you’ll see pictures of the three vehicles that have so far been fitted out with Transense tpm systems for demonstrations to prospective OEM purchasers of our technology.

In the US, there is a General Motors Saturn Vue and a Daimler Chrysler PT Cruiser. The car in the middle – the 3-series BMW has been fitted out for demonstrations in Europe.

There is an interesting story in connection with this. The car has been driven for several thousand miles, in rain and snow, on motorways at speed and up and down mountains along steep and winding roads - all to test our tpms under the most gruelling conditions.

Our system appeared to pass every test with flying colours except one. At extremely high speeds, it appeared to be registering a tyre pressure slightly higher than expected by the client. When I say high speeds, I mean speeds a lot higher than you could legally drive at in the UK and when I say slightly higher I mean something like a 1psi variation when the normal reading might be expected to be around 33psi.

The client queried this result and our chief scientist Victor Kalenin wrote a paper demonstrating that when you rotate a car wheel at high speed, “centrifugal force” comes into action and the molecules of air react rather like people on the Centrifuge fairground ride (if you ever went on that as a child – I remember well being stuck to the wall like a fly on flypaper as the floor dropped away).

Anyway, the point is that there is a real air pressure gradient from the centre of the revolving wheel to the edge and that at high speeds the pressure is indeed slightly higher on the outside than it is at the centre. Our measurements were correct – it was the expectation of the client that was at fault."

So it IS the rotation of the earth that does it as vehicles on the south of the map are accelerated quicker. Of course the question is then has anyone done this test for the Southern Hemisphere - it should be reversed. Volunteers please : )

[Jamm0r]

Our system appeared to pass every test with flying colours except one. At extremely high speeds, it appeared to be registering a tyre pressure slightly higher than expected by the client. When I say high speeds, I mean speeds a lot higher than you could legally drive at in the UK and when I say slightly higher I mean something like a 1psi variation when the normal reading might be expected to be around 33psi.

Erm... centrifugal force may very well affect pressure, but I'd wager that the increase in tire temperature at those high speeds accounts for most of that 1 psi variation they measured. A tire's pressure increases as its temperature rises. Temperature rises with speed. This is why race car tires are slightly underinflated, and often need a few laps to get tire temperature and pressure up to where the tire produces optimal grip. A race tire's pressure often increases much more than 1 psi though -- more like 5, on average.

[ August 18, 2006, 09:42 AM: Message edited by: Jamm0r ]

[dieseltaylor]

The TPMS system measures temperature and pressure and an algorithm adjusts it so I think then reports its unheated normal pressure.

Interesting point. Tires left out in the sun can increase by 5psi that is why you are meant to do your tyre pressures in the morning rather than at midday : ) I suppose better slightly over-inflated than under-inflated

[Jamm0r]

Originally posted by dieseltaylor:

The TPMS system measures temperature and pressure and an algorithm adjusts it so I think then reports its unheated normal pressure.

Ah, whelp that probably explains the small 1 psi variation, then. Compared to the 5+ psi variation one would expect from heat alone.

[ww2steel]

THAT'S WHAT IT IS!

The larger diameter of the earth (rotationally speaking) towards the equator causes more centrifugal force on the southern tanks, as they are further from the rotational axis. This causes them to sink less into the Deep Mud, thus they are faster! Mystery solved! :cool:

Mike

[GRUMLIN]

Its centripetal force guys. centrifugal is a load of balls made up for science teachers to lie to you with

G

[ww2steel]

Centripetal is the opposite of centrifugal.. in this case gravity is the centripetal force. I know I was talking about centrifugal at least.

Mike