Try the test using musketeers. It's far more decisive because of the very flat trajectory of their projectiles. They basically can't hit a unit standing behind another unit at all.
Printable View
Try the test using musketeers. It's far more decisive because of the very flat trajectory of their projectiles. They basically can't hit a unit standing behind another unit at all.
That doesn't necessarily mean that the projectiles are determining kills. The computer could show the results of a volley from muskets in a different manner than from, say, longbows because of different trajectories.
Let's take a look at this for a moment.
The Computer generates the kills and then the graphics represent such-
For- The drag on the computer would make the game unplayable without a high end system. I've personally played out battles with tons of archers, but never did I see the reduction in frame rates (I run a mod to ATI tray tools that allows me to see framerate) as significant as the opposite of the above would suggest.
I also have seen multiple hits on units that don't die, while the same hit on a different unit of the same type does result in a kill.
Against- ?
I think that we also need to consider which one is easier to program. I'm sure that the developers wouldn't make a bunch of unecessary extra work for themselves without any benefit.
Apologies in advance for cherry-picking text to reply to, but I didn't want this to get too long:Quote:
Originally Posted by the_foz_4
This would have been true several years ago, but much of the eye candy has been off-loaded to graphics co-processor cards. Calculating a ballistic trajectory is actually non-trivial, compared to shuffling soldiers around on the terrain, or melee combat:Quote:
1. That the battle calculations, not the rendering of the scene, is the limiting factor in your frame rate. I, on the other hand, would suggest that any bottleneck comes from something to do with the graphics processing as opposed to the battle calcs. It seems to me that some simple mathematical computations for moving an extra man and arrow around in a battle physics engine are completely miniscule compared to the thousands of computations that must be done to draw and dynamically light a single extra man and arrow into the scene.
http://en.wikipedia.org/wiki/Traject...f_trajectories
Yes, but try a frame rate test in simple terrain (see below).Quote:
2. That nothing else in the scene affects frame rate except the units. Look at the field of battle we're talking about here. The game draws a hugely complex scenery, including detailed blades of grass and numerous highly-detailed buildings often.
I would rather compare to a combat flight sim, some of which actually do calculate true ballistic trajectories, and none of which can show 2400 guys onscreen firing at each other, like M2TW does.Quote:
3. That thousands of object computations should horribly lag your system. Consider chess engines for a moment.
Okay, I just did a frame rate test. Custom battle, Spain vs. Portugal, each army using 10 Peasant Crossbows and 10 Peasant Archers. Setting is desert beach with clear weather, noon (no grass or buildings to interfere with framerate). Frame rate reported with FRAPS, at the start of the battle with missiles in the air, before there were too many losses.
-----
Small-size Armies: 960 men total, both sides
FPS looking away from army: 50 fps
FPS zoomed out to show all: 28 fps
FPS zoomed in on action: 20 fps
-----
Large-size Armies: 1800 men total, both sides
FPS looking away from army: 45 fps
FPS zoomed out to show all: 16 fps
FPS zoomed in on action: 20 fps
-----
Huge-size Armies: 2400 men total, both sides
FPS looking away from army: 35 fps
FPS zoomed out to show all: 10 fps
FPS zoomed in on action: 20 fps
Now, what this suggests to me is that we can easily see the graphics hit from the CPU and video card having to render 2400 individual soldier animations and 2400 missiles per volley, compared to 960 soldiers and missiles. The results from looking away from the action shows that the computer is crunching battle numbers that do slow it down a bit, but not that much. When zooming in on the action, whatever the computer has to do to calculate battle results, isn't taxing it enough to affect the frame rate.
Yes, it could be argued that CA's algorithms for "individual arrow trajectory" are just hyper-efficient, and it doesn't matter if there are 960 arrows in flight, or 2400 arrows. But that wouldn't match my experience with other games like combat flight sims, that actually do calculate real bullet stream trajectories. I can't begin to imagine a combat flight slim that could put 2400 planes in the air at once.
P.S. System specs, if it matters:
Athlon 64 3800+
2 gigs RAM
Nvidia GeForce 6800 Ultra
To be honest, I'm not sure why we're going off-topic arguing about how the game engine calculates arrow hits. It's pretty much anyone's best guess. This was all started with the following comment posted a while ago:
This is clearly not how the game works, since units standing in the flight path of archer fire will get hit and absorb fire. Whether the arrows are physical entities that are modelled by the engine and can hit other physical entities in the way, or are just a calculation kicking in as soon as a unit moves in between the flight path to divert a percentage of the damage to the unit standing in a way, is not too relevant.Quote:
Originally Posted by Reapz
What we really care about is the end result - that arrows behave like they're supposed to. i.e. if arrows are flying from point A to point B, and something gets in between, that something will get hit by most of the arrows. This is evident when your crossbowmen target a stack of units sitting behind another stack, and the non-targetted stack in front takes most of the damage as a result. It's even more evident with musket fire not hitting much farther than the first couple lines of troops, nevermind a unit standing behind another unit. Longbows from long range will have a considerable arc and hit more pikemen than crossbows, but the pavise in front of the pikemen still takes a considerable number of the hits. Try it yourself - make a proper spear wall with pavise crossbows right in front, against archers, and compare the casualties to a spear wall without anything getting in the way.
By the way, I also think that graphics cause the biggest hit to the CPU. In my case (3.2GHz, 1gb RAM, GeForce 6600GT), if I turn graphics down to bare minimum, the game practically flies compared to when I have them on medium/some on high, where large battles get very sluggish. I'm assuming all the calculations are still taking place regardless of graphics setting.
I think it is relevant, to know how this works in the game. If each arrow is being individually modeled for trajectory, and then the game is doing a collision detection to see if a soldier is located where it lands (which I don't believe at all).... then you're going to get a different result than if the game is simply operating on a "unit level," comparing unit stats and then throwing up randomized animations to show the results.Quote:
Originally Posted by Spark
Take the case of mixed units, for example. Let's say you have two units placed in the same physical location, which the game allows. The enemy AI 's archers can only target individual units, not a mixed physical space. So when the arrows fall out of the sky, are they falling as individually modeled physical objects that will damage both units in the mix indiscriminately, as would happen in real life? Or do they affect the "targeted" unit more than the unit mixed in with them?
BTW, I'm not suggesting that line-of-sight blocking isn't active.... I do think that's part of the picture, but that too can be done on a per-unit basis, not individual soldiers and arrows.
I think Foz has got the right idea. I saw arrows hitting two units even though only one was targeted. The overal losses were spread between the two units, with a higher rate of casualties on the Xbows who were up front. I'm wondering if the Xbows have to be turned around and reloading in order to get their shield defense bonus against the arrows. I also wonder if the AI archers were targeting the xbows or the pikes. I am assuming they targeted the Xbows, in which case I wonder what effect targeting the pikes instead would have.
As far as how I feel about mixing the pikes/Xbows:
They do to an extent aid and protect each other. Its pretty hard to deny that the Xbows are abosrbing hits that would have otherwise killed pikes. WIthout Xbows, pikes are annihilated completey by one archer unit. Mix Xbows and they survive with 2/3-ish men remaining. THats pretty big. Its safe to say the two are better off together then apart when facing a intelligent opponent who knows the pikes stand between him and victory. Also, its not like the archers are going to help defend the flanks or anything. THey were going to be out front anyway, may as well have them help protect the mission critical pikes Who MUST stay alive.
Against the AI it is probably better to keep seperate because the AI is too stupid to target the unarmoured pikes as opposed to the armoured Xbows. Against human players I'd definately mix them so I don't lose my pikes entirely. If you are facing a cavalry/CavArcher army and you lose your pikes, well, you lose the battle. Pikes are the only unit I've seen that can -stop- a cavalry charge. Every other infantry unit, including spears, are decimated by a formed cav charge.
As far as this whole argument regarding whether missiles have presence or not, I'm not concerned. I'm already 120% convinced that arrows do have presence in some form or another or otherwise I would not have got the results I did. I don't care about the exact game mechanics behind how the devs made it work this way. It just does, and thats good enough for me.
I was talking about calculating ballistics versus the rather lengthy calculations needed to render a 3-D scene at all, not just shuffle some guys around. If you've taken any physics dealing with simple parabolic flight paths, you'll recognize your ability to fairly readily calculate the flight path of something based on its fired velocity and angle of firing, by hand. Rendering an object, however, requires some rather unwieldy math to accomplish, and I'd never want to try to do it all by hand, I'd be at it for a long time just for a single simple arrow. I'm not trying to trivialize the trajectory computation, but drawing the arrow on screen is many orders of magnitude worse. I don't really expect anyone who hasn't had a computer graphics class to understand why, so just trust me when I say rendering a 3-D scene on the scales we're talking about is more numerically intensive than you can imagine.Quote:
Originally Posted by Zenicetus
Also, Z, as you point out, graphics are primarily not in the processor's hands anymore. This begs the question: when you get 20 or less fps, does the actual speed of your battle slow to a crawl, or does it happen at regular speed, just choppy as all hell? If it becomes choppy but with regular play speed still, then the problem is all graphics. I know this because if the problem is battle calculations, the game play has to slow down, as the game cannot go forward at all until it knows which men are being hit by arrows or slain in combat. In short, the game must wait for battle calculations since they determine outcomes and therefore ARE the battle - if they are slow, the battle is slow. Graphics are just your occasional snapshot of the battle the computer is carrying out in computations, so the corollary is that if the game speed itself slows down, then either the calculations are the cause of the bottleneck, or the game speed is forced down in low fps situations to alleviate the chopiness - in other words, devs may decide to slow the battle speed down to facilitate users being able to see the battle and make decisions reasonably (since you probably couldn't control a 5 fps battle for instance happening at regular speed). Either way it goes though, that knowledge from you will tell us something potentially useful.
Another useful thing to note here is that in order for the game to draw arrows in the dynamic fashion from each archer and through the air as it seems to, it has to be keeping track of them already in the coordinate system. If units were cookie-cutter (aligned in nice neat blocks of fixed dimensions always) then a "group firing code" could be used, but as I've seen groups of archers in disarray let loose volleys that seem to originate one from each archer's particular position, this strongly points to them being distinct in the code. Individual men certainly must be distinct from the group in code, or there'd be no way to draw them in unique positions doing their own things. If these are both the case as seems likely from a scene-rendering perspective, then the extension to arrow and weapon hit detection just requires collision detection, something that causes very little overhead.
One last note: I didn't mean to suggest the graphics were slowing down the processor necessarily... just that they IMO are the likely source of any bottleneck that occurrs. This could be due to the video card not being able to keep up with its heavy load, or the memory bandwidth being in too short supply, or any given data channel on your motherboard being backlogged with data - I just wanted to be clear that I think the graphics, for whatever reason are the source, since I'm so certain they require much more power to crunch out than the battle calculations do, even if you assume the individual entity tracking we've been so hotly debating is in place.
OT we are talking about the effectiveness of Crossbow Militia vs. longbowmen.
The OP said
I stated that in my testing (even basic) longbowmen are far better; the difference in results came from subjecting longbowmen to unreturned missile fire - allowing slow firing crossbows to catch up.Quote:
I have beend experimenting a bit with various missile units and I came to a conclusion that the best unit is - Genoese Crossbow Militia.
This came as a suprise since I would rather expect English to lead the way
Blademun then said
I then said that even against mixed unit targets the longbowmen still do well.Quote:
While many people stick up for longbows, I would have to say that Genonese Xbows do get a major advantage: Numbers. 8 Xbows vs 4-5 Retinues price for price, Genonese Xbows deal more damage per minute.
I think though that Xbows really shine when used in conjunction with pikes. Make a line of Pikemen milita, and then a line of Genonese Milita. Make the pikes 3 ranks deep and the archer 2 ranks deep. Place the Archers partialy inside the pikes, toward the front. Seperate, these two units are not so strong. TOgether they cover each others weaknesses perfectly. The Xbows act like human sheilds, their huge pavises intercepting enemy arrows for the vulnerable pikes.
Blademun responded:
Notice the change here? He's testing PEASANT archers against his mixed unit cocktail and it is still doing poorly.Quote:
Well, I went on and did those Impirical tests that were requested.
Setup: Enemy was one unit of Peasant archers, to rule out AP effects and other 'special abilities some archers get. I was HRE with a standard pike militia and Pavise Xbow. I let the archers fire their arrows till they ran out, then recorded the number of men left BEFORE the archers charged.
Pike/Xbow Formation: 49/33 respectively. Total Losses: 53
Nothing that anybody posted has argued against the power of longbows as they are used in the campaign, unless you use them in a non-sensible fashion.
Nothing posted shows the mixing of units allows armored units to "protect" weaker units. The mixing of units does dilute casualties - they are spread over several units. Whatever viewpoint you have on the physics of combat, longbowmen are very powerful missile troops and pavise crossbow militia can't stand up to them. Adding pikemen doesn't help.
Even so, I think some people are not yet convinced, as actual evidence of longbows vs XBows is a bit lacking. I haven't run them 1v1 in custom, or seen a reasonable confrontation between them that didn't get ruined by other units. Maybe I'll get a chance to do it soon, and we'll know more.Quote:
Originally Posted by Reapz
"Diluting casualties" is protecting the weaker unit, is it not? If the aim was to prevent pikes from being slaughtered before they enter combat, which it sounded like it was... then it has been done.Quote:
Originally Posted by Reapz
I suspect a few points about the 2-unit-in-one-space theory, though:
1. It has a more pronounced effect when the shooters are closer to your 2 units. They will fire on a lower trajectory, and the XBows should be right in the way and almost totally protect the pikes. From further away, arrows may fall in at a 45 degree angle or even steeper, making it much less likely a XBow body is in the way of an arrow that would otherwise nail a pike.
2. You are probably taking considerably more total casualties by putting 2 units in the space of 1. This follows from simple logic. A loose formation alleviates losses from archers because arrows fall between the men more often, so clearly less men per "unit area" (think square foot or square meter) is tougher to kill for enemy archers. Conversely, putting more men per unit area (tight formation) gets them killed more often. By stacking 2 units on top of each other, you make this even worse by cramming even MORE men in the same amount of space, thus making it much more difficult for arrows to make it to the ground safely. Arrows that would fall harmlessly around the pikes if they were alone are now likely to slam into the XBows who are taking up most of the non-pikeman space, so it's easy to see why you should in fact be suffering greater losses than would a single unit from an equal number of enemy salvos. Even if you are putting the 2 units into loose formation, a single unit in loose formation itself should suffer less total casualties, so this strategy seems like a poor idea if my assessment is correct.
Nobody wants to argue against your Longbows anymore Reapz. That was dropped a long time ago. I'm glad you love your Longbows so much you have to pick peoples posts apart. IMO, I hope Longbows get nerfed, since you've proven so irrefutably that they are uber powered. I hope your happy.
My impirical tests were made to deal with the argument that projectiles can't be intercepted before reaching a soft target. The two statements you pulled out are completely unrelated. THe topic of the post shifted between those two posts.
And as for the last poster. I agree partially, but it depends on which unit we're talking about. Pikes by themselfs get totally slaughtered by a single unit of Peasant archers. Pikes/Xbows suffered less overall losses. Infact, at 53 losses, the combined unit suffered less overall then the single pikemen.
I think because pikemen are so vulnerable to arrows, they are one of the only units that actually benefit from having another, heavily armoured unit, mixed in their front to absorb arrows. I'm not sure how it would do against an AP archer like Xbows or Lonbows, since i haven't tested that. I imagine, it would be alot worse with 80+ casualties.
I have to say that i dont have the feel of missile superiority when i field longbowmen because they are far from unique enough, alot of factions have units that are just as good are even better. This takes away alot of the english firepower. Its also historically inccorect that so many other factions could field ranged units similar in power to the english longbowmen who were prob the best ranged units in the world until the invention of rifled guns. I just dont get the feel of this when i play the game, so i would really like the game to be balanced somewhat more in this regard and make the english more powerfull as they should be. And limit or remove the mass recruitable elite units that ruin the game balance.
still....
the best non gunpowder missile unit is retinue longbowmen,....
they have extremly god defensive stats, extremly good missile attack, long range, the ability to fire flaming missiles ( which devastates all peavise crossbowmen, cause their shilds on teh back when they reload catch fire extremly quick), but their best feature is sharpened stakes, .....
when used right these stakes have won much a victory for me!!!! :)
so: ENGLISH RETINUE LONGBOWMEN takes them all :smash:
True but that shouldnt be why the longbowmen are better then the others, it should be that they have longer range and a faster rate of fire, and you cant use the stakes for anything when your attacking an opponent unless your opponent is stupid enough to charge you first.Quote:
Originally Posted by zstajerski
My problem is that i feel that the english longbowmen are just like any other elite archer unit excepr for the stakes ofc, but thats not enough for me. I want to see the longbowmen as they should be, and i think its stupid that CA gave so many other factions longbowmen as the only country to ever really use the longbow in a large scale was england and not france or some such.
At Crecy the French sent forward mercenary Genoese crossbowmen to engage the English longbows. The outnumbered longbows let them have it full bore and the crossbows hastily withdrew, only to be trampled by the french knights.Quote:
Originally Posted by katank
From http://en.wikipedia.org/wiki/Crecy:
But the crossbowmen would prove completely useless. With a firing rate of three to five volleys a minute, they were no match for the longbowmen, who could fire ten to twelve arrows in the same amount of time.....
On the other hand, because of their arched trajectory bows are infinitely more useful in sieges and in rough terrain, since buildings and hills are no obstacle for bows while making xbows totally useless.
I think this test is a little flawed, because it only tests on the missile grounds.
When all is said and done, I cannot get enough of dismounted Dvor as the best foot missile troops without gunpowder.
They have a Missile attack of 10, good defense (good armour) and a Melee attack of 11 right out of the gate.
If you fire a volley of arrows, it's usually those you see getting hit that dies. I've never seen a man die without getting struck by an arrow.
Ofcourse, sometimes they may take more than one hit.
If the enemy fires a volley of arrows and you quickly start moving the targeted unit, the arrows won't change direction to hit a particular man.
If you fire a cannon through moving units only the units getting struck by the cannonball dies, try placing a serpentine across a brigde when the enemy charges across and see the carnage. two long rips in the enemy lines, 100 casualties in one volley compromising of soldiers in different units, but nevertheless in a line.
I really don't think this is calculated before the arrows/cannons are fired, if it was the effects would be odd if you start moving targeted units around after volleys are fired, like arrows/cannonballs changing direction or people dying when not struck by arrows/cannonballs.
Just a thought?
rgds
adventuriers or genoese crossbowmen or also the unskinned war wagon crossbowmen that dont have a war wagon.
i use ottoman inf. all the time cause im an inf.person and these can shoot arrows until i tell them to charge. :yes: