Jump to content
Game-Labs Forum

Hidden mechanics and weapon damage degradation


pandakraut

Recommended Posts

My understanding is that the elevation lines feature was not retained from UGG because the contour lines were hand drawn and would have been a large amount of effort to replicate across a larger set of maps. Also elevation has far less impact in UGCW so that information is less relevant to the player.

Link to comment
Share on other sites

9 hours ago, pinsripes3333 said:

Thank you very much for the extra info. Do you know if the range for each type of ammunition is used is the same range or same proportion for every cannon? If it is the same range or proportion for every canon, do you (or anyone) happen to know what those numbers are? I could not find this info on a quick google search or search of this forum.

Canister shot is always out to .25 of a cannons range. Shell shot is from .25 to .5. If you haven't tried it yet check out the UI and AI Customizations mods. I updated the weapon arcs to show the ranges.

Link to comment
Share on other sites

On 8/10/2018 at 12:46 PM, pinsripes3333 said:

Edit: I also did not realize you were the author of the excellent artillery guide (although I probably should have given you name). I''m certainly not claiming to know more than you about artillery. Do you have a solid feel as to why the numbers I have do not match what your in-game and more accurate experience shows at least for the 20pdr? Perhaps playstyles as I did not look at mid to far damages much as I like to wheel my arty just a little behind my infantry and do not do much firing from over 700 range.

It's not all about stats.  The 20pdr Parrott Rifle, in my own experience, deals around 25 kills with even an unskilled crew at max range per volley.  This is enough to completely remove a gun from an enemy artillery battery in one shot, which alone is worth it's weight in gold.  But as you get closer, the gun just keeps on dealing more and more damage, and unusually for a rifled gun, it can absolutely murder enemy units in canister range.  Even the 3-Inch Ordnance Rifle can't keep up with the amount of kills at varying targets unless the targets are all at the same sweet spot range, let along actually do anything to an enemy unit in canister range (whatever your stats may say).  And artillery doesn't always have the leisure of shooting at targets less than 700m away, look at Malvern Hill for example.  Range is always a factor you have to take in, otherwise you may find yourself at the wrong end of a cannon, especially if you're Confederate.

As to why numbers don't translate well into the game, I couldn't tell you.  But I can tell you that experience trumps numbers any day.

Link to comment
Share on other sites

1 hour ago, TechnoSarge said:

Personally, I almost never see melee cavalry deliver fire. My thought is that their effectiveness is nearly limited to the saber, which has similar melee value. What are other people's perspective? Is there any good reason to up-arm melee cavalry past the 1842?

You can get them to fire if you click past the unit you want to target. Holding position or giving them a different move order once they are in range(just barely outside of melee) will trigger the pistols to fire. Getting this to work is inconsistent depending on unit angles. Otherwise they will usually fire after the unit tires out and can't keep up in melee anymore. Upgrading to Lematts can be nice, but usually it's not worth upgrading unless you have more money than you know what to do with.

Link to comment
Share on other sites

Thanks for the info on getting melee cav to fire their pistols; I've used it now! It certainly does require micro-management to get use of those guns.

BTW - I'm not seeing a degradation listing for Colt M1855 carbine, available only to Skirmishers.

 

Edited by TechnoSarge
Link to comment
Share on other sites

I wanted to rank all the weapons by relative performance; this has perhaps been done before, but I haven't seen it.

The spreadsheets Pandakraut already gave computed average accuracy and average damage.  These are affected by the Degradation of performance by increasing range, which is what many of the charts show.  Then I needed to factor in the Fire Rate (reload rate) which I assume is that listed in the Armory tool tips.

The final calculations are: AverageAccuracy x AverageDamage x AverageDegradation x FireRate. These are calculated in the new tabs for the WeaponStats spreadsheet.

These are then normalized by dividing by the result for a reference weapon I chose in each category. The reference weapons are early and commonly available.

A word here about range degradation... The word implies the drop-off in performance with range. That does not suggest what is true - that a high number is better than a low number. It may help to think of this number as the amount of lethality retained at the given range, not the amount lost.

Where you may wish to quibble with my analysis is judging an average Degradation. I chose a profile for each segment that has to do with how I use the troops in battle. For instance, my long-range skirmishers I always try to shoot from out of sight of their enemies, to avoid return fire (sniper rifles are expensive!). So, for long-range skirmish weapons, I chose 300 yds (in range of return fire from most infantry), 450 yds (generally out of sight), 500 yds ('cause the Whitworth reaches that far) and 600 yds (where the sniper rifles can reach). I did not use closer ranges than 300 because I am getting too many shot down except in pursuit or my target is facing another of my units. These profiles I chose are in the Infantry, Skirmisher, Cavalry, and Artillery spreadsheets under the "Analysis" tabs.

But the ranking outcome is under the "Summary" tab of the WeaponStats sheet. Each weapons segment is sorted by Relative Effectiveness. For myself, I've printed the Summary to keep by me as I play. These all pertain to vanilla stats, not modded ones. (For instance, the Whitworth artillery looks underwhelming, but it only delivers 5K kills in the mods! ... Artillery in general should be acquired by the role you will give it, not its effectiveness across the entire artillery line.)

Price Efficiency is about how much it costs to improve the effectiveness of troops by replacing their weapons with later, better ones that are purchased. The numbers are usually <1.00, indicating there is a cost penalty for the higher performance. The calculation is WeaponCost x ReferenceCost / WeaponRelativeEffectiveness. (Well, actually, you divide that by ReferenceRelativeEffectiveness, but since that is always 1.00, calculationally it makes no difference!)  By this standard, the Spencer Carbine is a bargain for cavalry, buying extra capability vs Sharps 1855 for less-than-proportional extra cost!  And by all means, replacing with captured weapons is optimal!

BEFORE YOU OBJECT that the Spencer Carbine, for instance, has different values for cavalry than for skirmishers, let me hasten to remind you that the AverageDegradation depends on the range profile for the use.  I expect cavalry to often deliver close-range fire, since it is often used to break up unwary artillery. Skirmishers with the same weapon, I want to use at longer range.

 

Edited by TechnoSarge
Files replaced
  • Like 2
Link to comment
Share on other sites

Thanks for putting this together. Still looking through it but some initial comments:

- I'm not sure if it affects your results but the tooltip displayed fire rate is actually 3000 / weapon.baseReload.
- Checking the pistols at a shorter range could be considered relevant, they tend to fire off at the end of melees in effectively melee range.
- You also may want to consider that for non artillery, at point blank range the degrade of every weapon is basically the same.

Edit: It looks like you are applying accuracy twice. The damage low and damage high are from base damage * AccLow and base damage * accuracy high. Then you are taking the average of those two values and multiplying it by the average accuracy value again. That is going to make the lower accuracy weapons look a bit worse.

Edited by pandakraut
Link to comment
Share on other sites

17 hours ago, pandakraut said:

Thanks for putting this together. Still looking through it but some initial comments:

- I'm not sure if it affects your results but the tooltip displayed fire rate is actually 3000 / weapon.baseReload.
- Checking the pistols at a shorter range could be considered relevant, they tend to fire off at the end of melees in effectively melee range.
- You also may want to consider that for non artillery, at point blank range the degrade of every weapon is basically the same.

Edit: It looks like you are applying accuracy twice. The damage low and damage high are from base damage * AccLow and base damage * accuracy high. Then you are taking the average of those two values and multiplying it by the average accuracy value again. That is going to make the lower accuracy weapons look a bit worse.

Thanks! I wasn't aware of those distinctions.

I think I'm OK about the reload... weapon.BaseReload as a divisor gives a larger number for "Fire Rate" if it is smaller, so if it were entered as a time measurement, a shorter time gives a higher Fire Rate, ie, faster reload, and this makes sense. The value is only adjusted by constants, so in comparing them between weapons, the unique actual time of reload is preserved in proper proportion.

I'll add a shorter pistol range in the profile, as I have not yet had enough experience to see when they fire. The advice in a post above finally got me seeing cavalry use their pistols at all!

Regarding point-blank fire -sure, there are times it is delivered, but I can't say how often, as there are no range markers. As I mentioned, I based profiles on the way I use troops. I will pluck a closer range and add it to rifle/carbine calculation though. (I COULD integrate the whole area under each curve, but that is a stupidly pedantic exercise!)

I agree accuracy is entering my computation twice because I missed how DamLow and DamHigh came about. I'll redo the numbers. So, folks, ignore the little man behind the curtain; the Great And Powerful OZ will be with you shortly. Lol!

  • Like 1
Link to comment
Share on other sites

When I went back into the numbers, I saw I picked off rifle degradation at 100 yds. I think that is an adequate stand-in for close-range fire and don't feel that working the profiles upward by over-emphasizing the better lethality as the lines close further would give a better sense of their performance overall. I have in the new spreadsheets, kept 100, 240, and 300 yards for infantry rifles.

Cavalry pistols have had their degradation at 10 yds averaged in, to include their behavior of firing on breaking contact.

The BIG change is fixing the issue with factoring in accuracy twice, which Pandakraut pointed out.  This does change the rankings! Suddenly the Sharps Model 1855 carbine becomes a Big Man on Campus! If you could arrange all the fire-fights by your infantry brigades to fall in 230 yards or less - that would be the weapon to give 'em!  (... This explains to me why I often get Sharps 1855's after a win when the enemy had no cavalry - the AI knows which weapons give good dps!)

I've edited my previous post to remove the links to the old spreadsheets.  I'm adding the new ones here. Again, the study summary is the tab that opens in "WeaponStats analyzed." That page prints on one sheet, in landscape format.

PS: It's curious that the game gives the 10pdr Parrott and the 12pdr Whitworth degradations that rise at the longer ranges.  I don't think anything else does. That means the Parrott actually gets more accurate beyond 1195 yards and the Whitworth beyond 1640 yds.  That's weird and unsupported by any physics I know of.

SkirmisherWeaponCurves analyzed.xlsx

WeaponStats analyzed.xlsx

ArtilleryWeaponCurves analyzed.xlsx

CavalryWeaponCurves analyzed.xlsx

InfantryWeaponCurves analyzed.xlsx

  • Like 1
Link to comment
Share on other sites

Errata: I just found Maynard's reload rate was copied wrong; it should be 67. That elevates its Relative Effectiveness to 1.22

It also has its range swapped with Burnside. It should be 260, and Burnside should be 275, in  the Cavalry section.

Edited by TechnoSarge
Mistake reported
Link to comment
Share on other sites

On the skirmisher spreadsheet analysis tab, it looks like the colt m1855 values might be wrong?

The 1855 showing so well is why I am hesitant to weigh this kind of analysis to heavily in my own weapon choices. The lack of range and it's impact on the ability to deliver damage is fairly obvious so I won't comment further on that. My main issue with valuing rapid fire dps is that it relies on being able to deliver multiple volleys. While a sniper unit can somewhat reliably stand and dps, most other unit types rely on very optimal conditions to be able to do so with any consistency. Flanking units and artillery are probably the next two most likely situations where you can somewhat reliably hit those conditions. In most other cases by the time a unit is delivering it's 2nd or 3rd volley it will potentially be performing far worse than on the 1st volley.

As a unit takes damage its damage per volley falls and its morale decreases. The morale decrease also negatively impacts unit efficiency further reducing damage. If a unit drops to wavering its ability to deal damage is even further reduced by the lack of volley fire. Because of this the unit that hits first and hardest tends to be in a better position for all future exchanges. Unless a weapon fires fast enough that it can get off two consecutive shots before the enemy can reload and fire, it will take multiple volleys before the faster firing weapon has a chance to catch up.

Because of the above, in nearly all cases I prefer to deal more damage per volley rather than relying on damage over multiple volleys. The harder I can hit in a single attack the more likely it will be to lower enemy morale to the point where they are ineffective. I do think that this is at least somewhat play style dependent. I've had very similar discussions with other players and they value fire rate far more than I do and are equally successful in the game. 

7 hours ago, TechnoSarge said:

PS: It's curious that the game gives the 10pdr Parrott and the 12pdr Whitworth degradations that rise at the longer ranges.  I don't think anything else does. That means the Parrott actually gets more accurate beyond 1195 yards and the Whitworth beyond 1640 yds.  That's weird and unsupported by any physics I know of.

I would guess this is more in the game balance over realism category. It gives those guns a very specific niche that is difficult to represent since the ability to be 'accurate' doesn't exist in a game where all shots hit their target.

Link to comment
Share on other sites

15 hours ago, pandakraut said:

On the skirmisher spreadsheet analysis tab, it looks like the colt m1855 values might be wrong?

No, look again! Both the Colt Model 1855 rifle and the Colt M1855 carbine are Skirmisher weapons. The carbine's tab is under the Skirmisher sheet, but my analysis includes the rifle from the Infantry sheet, because it is usable by Skirmish brigades. (The carbine is there, too, in the analysis.)

I'm with you in wanting to hit first and hit hard. That's not possible with the Sharps 1855 carbine, because anyone armed with it is going to have to close in from whatever range their opponent's weapon has to the meager 230 yards the carbine boasts. Going up against a Springfield 1861 means absorbing a couple volleys before reaching a point where fire can be returned. - You did notice, didn't you, that I said "if you could arrange all the fire-fights...to...230 yards or less"?? Fat chance!

15 hours ago, pandakraut said:

I would guess this is more in the game balance over realism category. It gives those guns a very specific niche that is difficult to represent since the ability to be 'accurate' doesn't exist in a game where all shots hit their target.

Boy, the graphics sure don't support that! When artillery fires, it's so common to see impacts a long way away from the target, it's as if they have a different target in mind!

Link to comment
Share on other sites

4 minutes ago, TechnoSarge said:

The carbine's tab is under the Skirmisher sheet, but my analysis includes the rifle from the Infantry sheet, because it is usable by Skirmish brigades. (The carbine is there, too, in the analysis.)

The value that looks off is in the skirmisher spreadsheet, analysis tab, column J. Every other weapon has values in the double digits except that one which is <1. Is that one really that much worse when compared to the others?

6 minutes ago, TechnoSarge said:

Boy, the graphics sure don't support that! When artillery fires, it's so common to see impacts a long way away from the target, it's as if they have a different target in mind!

Anytime a unit shoots damage against the target is calculated. Damage dealt can be <1 so it can appear that no damage was dealt. Maybe the artillery animations that seem to miss only happen when the damage is low?

8 minutes ago, TechnoSarge said:

I'm with you in wanting to hit first and hit hard. That's not possible with the Sharps 1855 carbine, because anyone armed with it is going to have to close in from whatever range their opponent's weapon has to the meager 230 yards the carbine boasts. Going up against a Springfield 1861 means absorbing a couple volleys before reaching a point where fire can be returned. - You did notice, didn't you, that I said "if you could arrange all the fire-fights...to...230 yards or less"?? Fat chance!

Agreed, that's why I only briefly mentioned the range and didn't bring it up further. When range is added to the equation the ranking of all of the carbines go down considerably. My main point was that, at least the way I play, fire rate just isn't something that I value very much.

Either in a new comment or in one of your previous ones could you combine your updated explanation of your data, conclusions, and spreadsheets? I'd like to link them in the main post.

Link to comment
Share on other sites

Pandakraut, you are correct on the entry on column J, Skirmisher - I didn't multiply by 100, as I did in the other instances. However, I see I did do so in the place that mattered, cell M41 on the Infantry tab of the WeaponStats sheet. So the carbine's performance is properly represented in the rankings of the Summary tab.

  • Like 1
Link to comment
Share on other sites

To summarize the work I did on weapon efficiency rankings - 

I define "Weapon effectiveness" as how much damage it does in a unit of time. This calculation is independent of the troops' quality who use it; it is what the weapon is capable of.

Damage per unit time = Average accuracy  x average damage per shot  x  average degradation with range  x  shots per unit time

Some of the numbers are hidden from players of the game, but Pandakraut has winkled them out and published them in his spreadsheets.

So, the formula above becomes DPS = ((AccLow x DamLow + AccHigh x DamHigh) / 2) x Degradation x FireRate in my calculations. This is not an absolute number, because FireRate is itself a scaled number, but when you compare one weapon with another by division, the constants of the scaling "drop out" . I have done that, using certain common, available, & early weapons as standards for their class. Thus my Summary tab on the "WeaponStats analyzed" spreadsheet gives rankings where the "standard" weapons are rated 1.00 and the effectiveness of others is higher or lower.

A comment is necessary here. My numbers for Degradation are pulled from Pandakraut's charts at specific ranges, based on the "normal" use of the weapon. For instance, in "InfantryWeaponCurves analyzed" I chose close range of 100 yards, mid-range 240 yards, and long-range 300 yards, picked off the numbers, and averaged them for those weapons. So my numbers for Degradation are based on profiles that make sense to me. If you use weapons differently, it's easy to substitute your own profile and plug it into the "Weaponstats analyzed" spreadsheet.  Also note that "degradation" is a bit of a misnomer - it is the lethality retained by the shot at increasing range, so that a higher number is desired rather than "less degradation".

Observations:

The hidden numbers affect rankings considerably. Take as example the Palmetto 1842 compared to the Springfield 1842.  In the in-game tool-tips, the two are equal in damage, range, and fire rate, but the Palmetto has accuracy of 13.5 vs. 12.5 for the Springfield. That sounds like an 8% improvement, but the rankings show only 1.02 (2% improvement) for the Palmetto. Most of the difference is in the fact that AccHigh is the same for both weapons, so the average accuracy is really almost the same. (The Palmetto also has a slight advantage in Degradation by my profile.)

I have included weapon ranges in the summaries, because they matter. A lot. Look at the Sharps 1855 single-shot carbine as an Infantry brigade weapon. Its Relative Effectiveness rating of 2.64 vs the Springfield 1842 is superlative, but your troops would have to walk a long time under fire to get within its limited range of 230 yds. Your morale might crack before ever getting in range if you're facing a long-range enemy. And... curiously, nearly every rifle has degradation ratings out to at least 300 yds, regardless of the range shown in the tool-tips. Here's where you notice for the first time that the tool-tip says "effective range,." not maximum range!

I have added "Price Efficiency" too. I picked off the cost numbers in my current campaign (CSA BG, after Antietam) with the career discounts and bonuses I have acquired. Because they are uniform, even if the numbers I use don't match what you are seeing now, comparing any two should be valid, since those discounts "divide out."  Price Efficiency I define as the value of one weapon in game dollars vs the standard, after applying the Relative Effectiveness. Nearly all numbers are less than 1.00. That indicates that better tech comes at a higher price. Look at Springfields 1842 and 1855 as an example. The 1842 is the standard, so it is 1.00 in Effectiveness and Price. The 1855 gives a hefty increase in effectiveness to 1.74, but price efficiency of 0.62 indicates that the improved performance is "worth" only $0.62 for every $1.00 you would have spent on the 1842.  (That may be so - but if I can afford the 1855, I'm buying it, thank you very much! I use it for its performance, not its price!) Price Efficiency is totally moot when using captured weapons!  And conversely, an Efficiency > 1 means you're buying better performance at a discount, relative to the standard.  Look at the Spencer carbine.

I declined to give price efficiency for cannons. You only buy them by small numbers of dozens, rather than by thousands, and you buy them by their role, not their cost.

 

Edit: Now that this post is linked to the main post, I'll point out you can get the spreadsheets mentioned in my post higher up in this page.

Edited by TechnoSarge
Noting where to get spreadsheets
  • Like 2
Link to comment
Share on other sites

Linked summary in main post. 

So one of the pieces of advice I remember hearing early on was that as the union, it is better to buy Lorenz's than Springfield 1855s. But that once you could buy Harpers Ferries it was worth upgrading. The reason for this is that the stat increase between the two is small enough that the increased price is not worth it. From looking at your data, it looks like the biggest differences between the two are the higher melee of the Lorenz and the faster fire rate of the Springfield. The HF is only a minor upgrade over the 1855. So with the new data it seems like if you are melee focused the only reason to move on from Lorenzs is when you run out of stock. But otherwise there is little difference in deciding to upgrade to the 1855 or the HF. The cost and increase in fire performance is about the same. Thoughts?

Link to comment
Share on other sites

I quite agree. The Lorenz has always been a favorite of mine because of the tool-tip number of Accuracy =75, but I never suspected this was AccLow and that there was such a thing as AccMax.  With the same AccMax, the difference in average is only half the difference indicated by the tool-tip. I figured the better accuracy for Lorenz was offset by the higher fire rate of the Enfield and the two were roughly equivalent. For gunnery, that's not true. Enfield, Spring 55 and Harper's are all nice upgrades from Lorenz.  But Lorenz is a better melee weapon.

In my own play, brigades that work up to a Melee stat of 45 or above, I designate as "assault" brigades and don't hesitate to charge with them, if they have sufficient morale and condition. That stat is independent of the weapon, so such brigades can be enhanced by equipping Lorenz.

Now that I have these numbers available, I won't purchase weapons unless they are, say, 0.25 better in effectiveness than what I have. If equipping with captured weapons, I'll work from the best down.

Link to comment
Share on other sites

  • 2 weeks later...

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!

Register a new account

Sign in

Already have an account? Sign in here.

Sign In Now
×
×
  • Create New...