Kaunitz Project [moved over to NTW engine!]

[Kaunitz]

Just some short good news: I got the videos to work. I "took over" the whole movies.pack - file, replaced the frontend and gameintro-videos with my own videos, repacked the whole thing as movie-type and called it "movies.pack". I backed up the original movie-file. So now all the glorious names of honour will show up in the main menu. And I feel like creating a new intro video this evening, soaked with 18th century spirit.

The problem with this method is that players would have to do some manual work (backing up the original file, copying the new movies.pack into thedata-folder...). Perhaps it can be overcome by means of an installer?

@ wangrin: In the Habsburg army, the first bataillon of a regiment would carry a Leibfahne (the white one) and an Ordinairfahne (the yellow one, maybe sometimes still red, judging from the paintings), whereas the second bataillon of a regiment would carry two Ordinairfahnen. There isn't really a difference in the design of the colours between different regiments. There are only some minor differences for hungarian as opposed to german regiments that I will have to look up.

PS: I renamed the faction "Austria" to "House of Habsburg", and, likewise "Prussia" to "House of Hohenzollern"

[bloody bill]

Sweet buddy I cannot wait to get some more battles in with ur mod.

[wangrin]

Can you update download link, please ?
Those is the first thread are dead

[Kaunitz]

I'm currently uploading, wangrin, sorry for the delay. I hope the version is the same as the one previously linked in the first post. The file has been changed the last time on 16.9., so I guess it should be. If there are any troubles, both players should redownload this new "old" version.

Are you going to have some online-matches or do you want to help me with the units?

I will send you the Shuvalov howitzers this evening.

[wangrin]

I want to try some LAN matches with my brother (I'm an extremely bad loser and I avoid carefully real on line battle...).
I will probably try to add some of my units too.

[Kaunitz]

Allright, it took ages but it's ready now:

http://www.gamefront.com/files/20883958/kaunitzmp3.7z

[wangrin]

Thank you very much

[Kaunitz]

Because the latest Nvidia driver didn’t crash my ETW as the previous ones did, and because I’ve now replaced gamecam with fraps, I’m able to record videos again! Here is just a very short one showing some unit cards in action. The grenadiers (perfect cohesion) advance against “very low cohesion” line infantry, suffering a bit too few casualties on their advance, and finally deliver a general discharge followed by a loud huzzah and a dash forward! J To be honest I hardly find myself going for the bayonet charge without firing at all, even now that bayonet charges work better than in previous versions. When you advance in perfect cohesion against an enemy with low cohesion, you just know that your volley is going to rout him – stupid AI doesn’t know how to keep its fire. There is no need for melee in such a situation…



Hopefully I can put up a video of a bayonet charge tomorrow.

[bloody bill]

Sweet buddy, vid looks great as do the unit cards. When will the mod be ready? Am looking forward to a battle with you.

[Kaunitz]

I'd say mid-November because I don't want to hurry. But there are a lot of things that I cannot test against the AI, which always "exhausts" its troops, so if you like, we can conduct some battles earlier! The only problem is that my internet connection is a bit weird these days. This should change in January when I'm moving. Perhaps, for November and December, you can find a second wargame buff in the same time zone with a good connection?

[bloody bill]

I got some buddies very interested. so let me know

[Didz]

Your figure scale is 1:4 isn't it, so a four gun battery is the equivalent of 16 guns, which does seem a bit large. Or do you mean 2 gun models representing 4 guns?

If I understand your test results 8 round shot hits inflicted 26 casualties, which is 3.25 casualties per hit. Presumably on a three deep close order formation. That's slightly higher than I'd expect and suggests that the richocet fire was extremely effective, unless this was enfilade fire of course.

8 hits out of 30 balls is not too excessive if the ground was perfectly favourable out to 1,200 yards. I imagine this would fall quickly in rolling or broken terrain.

[Kaunitz]

With 4-gun-batteries I was refering to units of a single gun model.

Yes, 8 shots of the gun model inflicted 26 hits, and you're correct that some shots took out 4 models, even though I let the battalion advance from a right angle. I'm pretty sure that this is caused by the big size/diameter of the cannon-ball-model which - in combination with the dense and not 100% perfect ranks - happens to strike two models sometimes. I've already decreased its size now, but still there are many shots that take out 4, or even 5! models.

Sorry I had a typo in there. 7 (6 too short, 1 too high) out of 15 shots missed. After this test, I’ve increased artillery accuracy a little bit. The ground was perfectly flat (grassy flatlands).

Here is a short video that gives you an impression of the effect of two batteries of 6pdrs (plus a prototype howitzer) at long range. I left in the wwI-impact-effects for better visibility of the grazes. Unfortunately the AI went crazy as soon as it came within grape range, so the rest of the encounter didn’t qualify to be shown. You can also see how the first two shot in the video goes high for they were still aimed at a spot farther away, at maximum effective reach (so that the gun had to be elevated).

[Didz]

So, in the video then there are 2 gun models plus a howitzer model, representing 8 x real guns and 4 x real howitzers firing on an infantry battalion.

The ground there is pretty flat and open so it seems that you would expect better than average ricochet results unless the ground was particularly soft. I suppose the question is whether you would expect a real battalion to withstand the constant attention of eight guns as it advanced that distance.

I checked the rate of fire and it seems reasonable at just over two shots per minute. If you wanted to reduce the casualties you could argue that at long range gunners would take longer to reload and relay their piece and thus have a lower rate of fire. But then the counter arguement would be that at sort range when canister was in use they would reload much more urgently and not bother to relay the gun producing a much faster rate of fire.

I was trying to judge whether the lethal width of the hit was artificially wide and if that was inflicting an undue number of double file hits, but I couldn't really see any evidence of that.

You're orginal figures suggest an average accuracy of 50%. whereas Chart 4 in Imperial Bayonets suggests that ranges beyond about 550 metres the hit percentage ought to be less than that. At 1,200 metres for example a 6pdr was only managing a 15% hit percentage. That was against a 31 x 1.9 meter target screen. The only logical reason I can think of for the percentage to drop is unpredictable bounces at each graze.

The chart shows a quite steady drop in accuracy from 100% to about 45% over the first 550 metres, which would be 'point blank' range (before the ball touched the ground). Then a different drop in accuracy rate from 550-850m which would be after the first bounce, and yet another change from 850m onwards, which would probably be the second and third bounce and the final rolling along the ground.

This would suggest to me that something is eliminating some of the shot at each graze with the earth. Either burying them so they don't travel any further or causing them to skip higher than usual so that they clear the top of the targets. If this wasn't the case then the accuracy curve ought to be constant as the windage is determined by the barrel of the gun and so would not vary with range.

So basically the accuracy of each shot fired is constant, the only variable is the effect upon it when it makes contact with the ground.

The fact, that you are getting a higher accuracy rate would therefore have to be down to the ground impact results being too predictable, even for a range test. But I don't know if its possible to influence the predictability of a bounce in the TW engine.

[Kaunitz]

Thank you very much for the feedback, Didz.

The rate of fire can be set differently for each ammunition type, so I have attributed a much higher reloading speed to canister (3 shots per minute) than to roundshot. What we can not have is the crew to reload faster with decreasing distance to the target.

Indeed I think that it's still a matter of too high accuracy at longer ranges. I need to find out how to make guns less accurate at longer ranges, which is not as easy a task as it might sound, especially when accuarcy is at least equally a matter of the grazes as it is a matter of the crews' aim. I have no means to influence "bouncing behaviour", I can only adjust the guns/crews accuracy.

According to your data (first graze at 550/137m with 0 elevation) I'm going to reduce the projectiles' velocity (=range) a little bit. Right now, the first graze of a 0° shot happens at a much longer distance.* This means that the crew will have to elevate the gun on closer distances and/or has to rely on riccochet fire much more on the longer distances. There is however one problem with riccochet firing in ETW: you have to target the right spot on the ground, ideally so that the barrel is at 0 elevation. If you aim too "short", then the ETW-crews react stupidly by actually "lowering" the gun which makes ricochet fire less likely (the ball looses a lot of momentum with the first graze, or it doesn't bounce at all).

I think that - even though the accuracy was too high at extreme range - the video showed the power of the "one-model=4 guns-batteries". Imagine how a single gun model takes out 3x28 models per minute at 80 ingame yards (320 real yards) with grape...I can see if I can put up a video of canister and grape tests for you to comment on.

* Sidenote: positioning your gun on a slope (in the video my guns had a rather big advantage in height) increases the reach of the guns enormously. I will see if the negative effects are correct as well (less bounces, difficulties at close range for not being able to lower the gun too much)

PS: I've added the intro video as youtube video to the first post. I've dropped the idea of creating a new one. The old one isn't that bad and a new one would require so much time and work.

[Didz]

The rate of fire can be set differently for each ammunition type, so I have attributed a much higher reloading speed to canister (3 shots per minute) than to roundshot. What we can not have is the crew to reload faster with decreasing distance to the target.

That seems a reasonable compromise, most experienced players will remember to switch to canister when the range is short enough, and the AI is usually pretty good at it. There were occassions in real life when gunners stuck with ball at short ranges but I imagine it was pretty rare.

Indeed I think that it's still a matter of too high accuracy at longer ranges. I need to find out how to make guns less accurate at longer ranges, which is not as easy a task as it might sound, especially when accuarcy is at least equally a matter of the grazes as it is a matter of the crews' aim. I have no means to influence "bouncing behaviour", I can only adjust the guns/crews accuracy.

Thats a nuisance. At point blank range (less than 550 meters) one wants artillery to be pretty lethal so I'd be reluctant to suggest a general reduction in accuracy just to compensate for the fact that they are too accurate after the first bounce.

According to your data (first graze at 550/137m with 0 elevation) I'm going to reduce the projectiles' velocity (=range) a little bit. Right now, the first graze of a 0° shot happens at a much longer distance.*

Ok! it will be interesting to see if that helps. To be strictly accurate the range test data I have from earlier research suggests that a 9pdr fired at 0 degree's elevation over a relatively flat range will make its first graze at 400 yards, its 2nd at 800 yards and its third at 900 yards, whilst a 6pdr under the same conditions achieved 360 yards; 720 yards and 800 yards.

What I suspect will be harder to model is that at elevation the figures were:
9pdr (1 Degree elevation)
1st graze 700 yards; 2nd graze 1,000 yards (with the shot non-lethal between 0-650 yards)
9pdr (2 degree elevation)
1st graze 900 yards; 2nd graze 950 yards; 3rd graze 1,000 yards (non lethal from 0-850 yards)

As you can see elevation rapidly reduces the lethal distance of ricochet fire, whether its a consequence of deliberately elevating the barrel, or simply firing from an elevated position on a target on lower ground. It also reduces the travelling distance of the ball between each bounce as the angle of impact is greater and the ball gets kicked higher into the air by each impact.

This means that the crew will have to elevate the gun on closer distances and/or has to rely on riccochet fire much more on the longer distances. There is however one problem with riccochet firing in ETW: you have to target the right spot on the ground, ideally so that the barrel is at 0 elevation. If you aim too "short", then the ETW-crews react stupidly by actually "lowering" the gun which makes ricochet fire less likely (the ball looses a lot of momentum with the first graze, or it doesn't bounce at all).

Yeah! I noticed that you were manually targeting the ground. I tend not to do that unless the AI is being particularly stupid.

* Sidenote: positioning your gun on a slope (in the video my guns had a rather big advantage in height) increases the reach of the guns enormously. I will see if the negative effects are correct as well (less bounces, difficulties at close range for not being able to lower the gun too much)

Thats not as much of a problem as long as the shot is not lethal for its entire travel. In practice firing from higher ground gave gunners a problem in hitting troops close under the lee of the slope, as at 0' elevation the shot simply flew over their heads. Hence at Waterloo for example, many batteries had to be moved forward until they were actually on the forward slope of the ridge, so that the barrel was naturally angled down the slope by the carriage.

[Kaunitz]

Thats a nuisance. At point blank range (less than 550 meters) one wants artillery to be pretty lethal so I'd be reluctant to suggest a general reduction in accuracy just to compensate for the fact that they are too accurate after the first bounce.

Hm. I don't know. It's almost impossible to find out how ETWs accuracy parameters work in detail. Ideally, I would make accuracy drop at ca. 550 meters, so that we would have a kind of "abstract" representation of the reduced accuracy due to the deflections/bounces. Apart from the fact that I'll never find it out, I fear that ETWs accuracy gradually decreases over distance.

Ok! it will be interesting to see if that helps. To be strictly accurate the range test data I have from earlier research suggests that a 9pdr fired at 0 degree's elevation over a relatively flat range will make its first graze at 400 yards, its 2nd at 800 yards and its third at 900 yards, whilst a 6pdr under the same conditions achieved 360 yards; 720 yards and 800 yards.

What I suspect will be harder to model is that at elevation the figures were:
9pdr (1 Degree elevation)
1st graze 700 yards; 2nd graze 1,000 yards (with the shot non-lethal between 0-650 yards)
9pdr (2 degree elevation)
1st graze 900 yards; 2nd graze 950 yards; 3rd graze 1,000 yards (non lethal from 0-850 yards)

As you can see elevation rapidly reduces the lethal distance of ricochet fire, whether its a consequence of deliberately elevating the barrel, or simply firing from an elevated position on a target on lower ground. It also reduces the travelling distance of the ball between each bounce as the angle of impact is greater and the ball gets kicked higher into the air by each impact.

Obviously I have set all the spots of the first grazes (=velocity of the projectiles) far too long then. I guess it should be possible to produce believable results since DeeJay has pointed out that there is a "maximum elevation" variable available. So, basically, I have two parameters at my disposal:

1) velocity of the projectile (determines how far the projectile travels)
2) maximum elevation of the barrel (if you aim at too high/low a spot, the gun simply won't fire)

Right now, I allow for a maximum elevation of 2 (whatever unit it is) which results in adequatly flat trajectories. If you aim at a spot very far away, the gun will still be elevated so much that the ball realistically travels over a mans' height for most of its journey. And in this case the higher angle at which the ball hits the ground leads to shorter distances between the following grazes, just like you've said.

The spot of the first graze can be determined by tweaking the velocity. I will reduce velocity quite a lot in order to approach the distances that you've mentioned, even though cannon balls will then travel very "slowly" through the air.

The "effective range" entry in the projectiles table doesn't really matter (the AI uses it, I guess). However, if you set velocity too low and don't allow for enough elevation, the gun might be unable to fire, even if the target is within the "effective range". This will very likely be the case when I decrease velocity. E.g. I've set the effective range of 12pdrs to 300 (=1200) yards. At the same time, I need to decrease velocity a lot because I want to have the first graze occur at shorter distances. This means that the gun is very likely to refuse to fire if you target a spot too far away but still within effective range (=the fire-arc). The gun cannot "directly" target the spot because of the velocity/max. elevation settings. Rather, you'd have to target a much closer spot and rely on riccochet.

I see if I can put up a video of this in the course of the week. It's better explained with a video than with words.

Thats not as much of a problem as long as the shot is not lethal for its entire travel. In practice firing from higher ground gave gunners a problem in hitting troops close under the lee of the slope, as at 0' elevation the shot simply flew over their heads. Hence at Waterloo for example, many batteries had to be moved forward until they were actually on the forward slope of the ridge, so that the barrel was naturally angled down the slope by the carriage.

I think we might be able to produce these results. I will try to capture it in a video.

PS: A small success: the rifle armed chasseurs will get "rifle"-fire sounds! (I didn't manage to create a new sound-assignment, I simply took an existing one and edited it.)

[ACMilan88]

Hi Kaunitz, how is your mod project progressing? I'm uncertain exactly what it does but it looks very interesting If you have anny questions about the sponge mod and how the units is catigorized don't hesitate to ask

The unit icons looks very nice btw.

Greetings

[Didz]

Hm. I don't know. It's almost impossible to find out how ETWs accuracy parameters work in detail. Ideally, I would make accuracy drop at ca. 550 meters, so that we would have a kind of "abstract" representation of the reduced accuracy due to the deflections/bounces. Apart from the fact that I'll never find it out, I fear that ETWs accuracy gradually decreases over distance.

Simply by observation I would have to agree with you. I doesn’t look to me as though the shot hits the ground early enough (e.g. after about the first third of its travel 400-500 yards) or that it’s further travel is significantly affected by its impact.

In your test video it looked as though the shot were impacting the ground for the first time almost in front of the target where you had told the battery to fire, even though the target was at extreme range. That would suggest a high elevation was being used by the gunners but this was not evidenced by the shots trajectory.

Of course once again the results are also affected by the oversized models. In truth the muzzles of the model cannon are probably about three times further above ground level than those of the real cannon they represent would be, and so there is an artificial degree of elevation to every shot anyway simply due to vertical scale.

This is further affected by the fact that the target models are about three times taller than their real life targets would be, and so few shot are going to bounce over their heads.

It’s a real quandary.

If you aim at a spot very far away, the gun will still be elevated so much that the ball realistically travels over a mans' height for most of its journey. And in this case the higher angle at which the ball hits the ground leads to shorter distances between the following grazes, just like you've said.

That sounds good it just remains to try and get the distances right. The furthest distance a shot fired from a standard 9pdr carriage is recorded as travelling before bouncing at 2 degree elevation is about 900 yards. So, if you’re currently getting distances well in excess of that before the first bounce it would suggest that possibly the effective range is too long overall.

The spot of the first graze can be determined by tweaking the velocity. I will reduce velocity quite a lot in order to approach the distances that you've mentioned, even though cannon balls will then travel very "slowly" through the air.

Yeah! It needs some experimentation to get the bounce pattern right. My only concern was that if I recall rightly you were forced to increase the velocity to give the shot a flat trajectory in the first place.

What we don’t want is to go back to the catapult rock lobbing behaviour that we saw in the early version of the game.

E.g. I've set the effective range of 12pdrs to 300 (=1200) yards. At the same time, I need to decrease velocity a lot because I want to have the first graze occur at shorter distances. This means that the gun is very likely to refuse to fire if you target a spot too far away but still within effective range (=the fire-arc). The gun cannot "directly" target the spot because of the velocity/max elevation settings. Rather, you'd have to target a much closer spot and rely on ricochet.

I don’t have accurate figures for the 1st graze of a 12pdr shot, but based on the difference between a 6pdr and 9pdr, logic would suggest it should have been somewhere around 600 yards.

The situation you describe where artillery cannot hit a target because the gun cannot be elevated enough to pitch a ball far enough, actually sound quite realistic to me. Whilst the theoretical maximum effective range of a 12pdr was around 1,200 yards, we have both seen examples where gunners were unable to target effective fire at much less than this theoretical range due to the profile and nature of the intervening terrain.

At Waterloo for example the initial French gun line was about 1,200m from the Allied line, and only 800m from Bjlandt’s Brigade and La Haie Sainte, and only 500m from Hougoumont. But primary evidence suggests that Bjlandt and the centre of the Allied right was the main focus of the gunners attention.

Baring’s men in La Haie Sainte hardly suffered at all during the opening bombardment despite being in the middle of the French gun line, and nor did the defenders of Hougoumont.

The reason becomes plain when looking at the topography of the battlefield in more detail. In that there is a second rise of approximately the same height as the French gun line about 400-500M in front of its position which extends to the east of the Charleroi Road about 700m (across almost half the French right flank).

Spoiler Alert, click show to read: 

Therefore, not only were La Haie Sainte and Barings men in the lee of this rise, and so partially hidden from the gunners view, but shot from their guns would have hit the top of this rise and bounced over Barings position.

On the left Hougoumont had even better protection as dense woodland that surrounded the southern face of the enclosures not only prevented the French gunners seeing a target, but these mature tree’s (30’ tall) would have absorbed most of the shot fired into them, even at higher elevations, and shot fired at 2 degree elevation would not have become lethal until they hit the front face of the Allied ridge behind the chateau.

In practical terms therefore, only Bjlandts Brigade and the centre of the Allied right was a target at the start of the battle and most of the French bombardment hit those area’s. Bjlandts was particularly unlucky as the low rise peters out almost directly opposite his position and left his brigade exposed to 0 elevation ricochet fire about 800-900m metres from the guns of Marcognet’s and Duruttes Division. These guns are shown as deployed at a height of 120m along the French ridge and the first graze of their shot would have hit the valley floor, which thanks to the low rise was only about 10m lower than their position at its deepest point, before bouncing to make their 2nd graze virtually amongst the Dutch lines. So, the French gunners would not even have pitched their shot high to clear the valley in order to hit Bjlandts Brigade, and his men suffered badly because of it.

Soon after the advance of Donzelots Division on La Haie Sainte the French tried to move their gun line forward onto the low rise between their original position and the farm, and it was these guns and their still attendant limbers that were overrun by the British heavy Cavalry during their charge. However, by about 3:30pm this low rise was occupied by several 12pdr reserve batteries from the Imperial Guard who were now able to alternate their fire, either firing down the slope directly into the southern walls of La Haie Saint at virtually point blank range, or more often using elevation to lob their shot over the farm and onto the forward slope of the Allied ridge behind it. The casualties amongst Ompteda’s and Kielmansegges Brigades at this time are said to have been severe and the only respite was when the French cavalry or infantry attacks masked the gunners fire.

[bloody bill]

Kaunitz buddy, the vid looks great. The unit cards are pretty cool. If you need any help testing with a person let me know. Iam always up for a battle and i got some buddies very interested to help in any way.