Home » MODDING HQ 1.13 » v1.13 Modding, Customising, Editing » v1.13 Weapon & Item Refinement » Presentation: Weapon Effectiveness Scales on 6 Dimensions
Presentation: Weapon Effectiveness Scales on 6 Dimensions[message #310233] Sun, 16 September 2012 21:42 Go to next message
sorca_2

 
Messages:206
Registered:September 2010
Location: California, USA
(Scroll to the third post if you just want to see some example data without my explanations.)

Background

I've tried several times to make an overall effectiveness index for the weapons in 1.13 because a) new players struggle to figure out what weapons are best, b) even regular players have never heard of some of the prototype weapons, and c) comparing weapons in game is difficult due to the sheer volume of data. My previous attempts were unsuccessful because, when it comes down to it, it's hard to put things like weapon speed and attachment compatibility in the same scale. It's also hard to say one weapon is "better" than another because range isn't very important inside of 10 squares, and a weapon with burst/auto capability is almost always better than a semi-automatic.

Inspired by recent games (notably the upcoming Mechwarrior Tactics), I decided to attempt a multidimensional approach instead. The following are the results of my efforts. Please let me know what you think, particularly if you have constructive suggestions.

Each of the six dimensions of analysis is rated on a scale of 1-7 (with one exception) where 1 is the worst and 7 is the best. This was done because 7 has a nice "average" of 4, it fits with the in-game implementation idea I have, and it just happened to work well in almost all of the weapon data I included.

Though the scales combine data from dozens of variables from the 4870 release, the following variables were not included in the current "proof of concept" stage: reliability, repair ease, no semi-auto, attack volume, selection of ammo types, magazine size, reload time, AP to reload manually. Moreover, I based this version on OCTH statistics because I understand them better and there are fewer numbers to work with. Now that I have created the scales, it would be relatively simple to create NCTH scales or scales for AIMNAS, UC/AFS/DL, etc. I excluded the Dart Gun and the Flamethrower because they were throwing off the scales.

Dimensions

1. Damage

This 1-7 scale is pretty straightforward. THe minimum damage from the 4870 guns is 20, max is 60. I converted these values to the 1-7 scale using percentiles--roughly the bottom 14% of damage scores got a "1" on the Damage scale, 15-28% got a "2" etc. up until the top 14% got a "7".
Limitations: This does not account for ammo types. It could if we think it's necessary.

2. Range
This 1-7 scale is done with the same logic as the Damage scale.
Limitations: The only thing I can think of is that, in my experience, the difference between range 12 and range 20 is much more important than the difference between range 50 and range 58. The percentile approach should address this to some extent, but it would be possible to do a geometric conversion so the values are skewed such that differences between shorter ranges are bigger than differences between longer ranges.
Example scores:
Name        Range    Range Scale
.38 Special  10.5        1
Rem. M870    15.0        2
MP5N         17.5        3
SIG SG550    38.0        6
M24          78.0        7



3. Accuracy
Again, this 1-7 scale was made using percentiles like the previous two scales. The min OCTH accuracy is -2 and the max is 12, so the distribution isn't very exciting. OCTH accuracy of -1 converts to 1 on the Accuracy Scale, OCTH accuracy of 6 converts to 5, and so on.
Limitations: This only addresses the accuracy value, not the Auto Penalty or Burst Penalty because they are factored in elsewhere. On a NCTH scale, it would probably make sense to factor in Handling and Aim Levels.


4. Speed
Here's where it starts to get interesting. The two variables here are ReadyTime (time it takes to raise the weapon) and ShotsPer4Turns (AP it takes to fire). First I created percentile scores for ReadyTime and ShotsPer4Turns. I then combined these scores weighting ReadyTime by .33 and ShotsPer4Turns by .66. This means that ReadyTime is essentially half as important as ShotsPer4Turns. I did this because you ready a weapon much less often than you fire it. Still, ReadyTime is important in some of the most dangerous situations.
Limitations: Does not currently factor in ManualReloadAPs. It would be fairly simple to give manual-reload weapons a blanket -1 hit on the scale, for example. We could potentially include Magazine Size and Reload Speed, but I thought these were much less important than ReadyTime and ShotsPer4Turns.
Examples:
Name        ReadyTime    ShotsPer4Turns   Speed Scale
Barret M82A2   19             6.57             1
AK-47          15             12.85            3
FAMAS G2       9              15.03            5
FN Five-seveN  0              23.5             7


(Continued below...)
Re: Presentation: Weapon Effectiveness Scales on 6 Dimensions[message #310234] Sun, 16 September 2012 21:43 Go to previous messageGo to next message
sorca_2

 
Messages:206
Registered:September 2010
Location: California, USA
5. Burst/Auto
This one is the most complicated. There are 5 variables under consideration: ShotsPerBurst, BurstPenalty, BurstAP, AutofireShotsperFiveAP, and AutoPenalty. First I created a simple Burst Shots per Burst AP calculation; we can agree that a weapon that can fire 3 shots with 10AP is more efficient than a weapon that can fire 2 shots with 9AP). Then I created percentile scores for each of the 4 variables (AutofireShotsperFiveAP, AutoPenalty, BurstPenalty, and the new one: BurstShotsperBurstAP).

To make matters more complicated, some weapons are single-shot only, some are single/auto only, some are single/burst only, and some are single/auto/burst (and yes, some are auto-only, but that does not impact the Burst/Auto Scale). I determined that weapons without any kind of burst or auto capability will receive a 0 on this scale because even the crappiest burst capability is better than no burst at all. Weapons with single/burst and single/auto would get scores between 1 and 6 while weapons with single/burst/auto would get 2-7. Thus, a single/burst/auto weapon essentially gets a "+1" on this scale compared to a clone of the same weapon that is single/auto only.

I created percentile scales for the single/burst weapons using a method similar to the Speed Scale above where BurstPenalty was weighted at 0.33 and BurstShotsperAP was weighted at 0.66. This is due to the importance of suppression fire and because it's easier to improve accuracy than rate of fire. I then did the same thing for single/auto weapons using the autofire variables.

For single/burst/auto weapons, I did the same thing but used a 4-variable score where BurstPenalty and AutoPenalty were weighted at 0.167 each while BurstShotsperAP were weighted at 0.33 each. This basically combines the two scales for single/burst and single/auto weapons above.
Limitations: None that I can think of if the variable weighting above is good.

Examples:
Name        Shots/Burst||BurstPenalty||BurstAP||AutoShots/5AP||AutoPenalty||Burst/Auto Scale
FN FAL         N/A           N/A         N/A         3               9          2
AK-47          N/A           N/A         N/A         3               8          3
Beretta 93R     3             8           8         N/A             N/A         5
FN P90         N/A           N/A         N/A         4               7          6
FAMAS G2        3             7           8          5               8          7    


I should mention here that the Beretta's score is higher than the AK-47's because the Beretta's burst ability is pretty awesome relative to the burst ability of other weapons with burst while the AK's auto ability is average relative to other weapons with auto fire.

6. Compatibility
Another hugely important factor in a weapon's effectiveness is its compatibility with attachments because they can hugely impact the other 5 dimensions. I created a simple count of how many attachments are compatible with each weapon using the weapons spreadsheet in my signature. I then used a percentile conversion to make a 1-7 scale. I combined the shotgun chokes into 1 attachment because they're only compatible with shotguns and therefore would have artificially inflated the shotguns' scores.

Examples: compatibility with 7 attachments yielded a score of 3, compatibility with 13 attachments yielded a score of 5, and compatibility with 18 attachments yielded a score of 7.
Limitations: This does not factor in the relative importance of particular attachments. For example, one could argue that grenade launcher compatibility is more important than C-Mag compatibility, etc. Thus, the scale measures how many attachments it can use, not how good those attachments are. I was hoping that naming it "compatibility" would convey this rather than "attachment quality" or somesuch.

Conclusion below...

[Updated on: Sun, 16 September 2012 23:06] by Moderator

Re: Presentation: Weapon Effectiveness Scales on 6 Dimensions[message #310235] Sun, 16 September 2012 21:43 Go to previous messageGo to next message
sorca_2

 
Messages:206
Registered:September 2010
Location: California, USA
With the six scales created, it is possible to evaluate weapons in a quick and easy way. It is also possible to combine these scores to evaluate the overall "effectiveness" of each weapon. Effectiveness is simply a combined total of the scores on each of the 6 dimensions. Because of the percentile approach, there will be a bias towards well-rounded weapons: a weapon with an extreme in two categories will probably be rated as less effective as a weapon that is above-average in 4 categories.

This is NOT meant to give you the "final word" on which weapons are absolutely the best because we're aggregating and generalizing. Weapons with the same effectiveness could vary drastically but would be equivalent in the sense that they make up for a deficiency in one area with a strength in another.

In summary: a weapon with higher effectiveness means that it has more strengths and/or fewer weaknesses in the six dimensions than a weapon with lower effectiveness. You could also think of it as a judge's scorecard with a theoretical maximum possible score of 42 and a "practical" maximum possible score of about 35 (a 30+ effectiveness weapon with attachments).

Examples:
Name         Damage+ Range+ Accuracy+ Speed+ Burst/Auto+ Compatibility = Effectiveness
Ruger S Rdhwk   3      1       1        3         0            1               9
SIG P229R       1      1       1        7         0            2               12
FN Five-seveN   2      2       2        7         0            1               14
Ingram M10      2      1       3        4         4            3               17
Steyr TMP       1      2       2        7         4            3               19
Dragunov SVD    7      7       6        1         0            2               23
AK-47           5      4       3        3         3            7               25
HK G3A3         6      6       4        2         2            7               27  
FAMAS G2        4      5       4        5         7            5               30
Steyr AUG-A3    6      6       4        6         3            6               31


If someone can justify a rationale like "damage is twice as important to effectiveness than speed, but half as important as range", we could make an "Adjusted Effectiveness" score that reflects weights.

I think this will be useful for lots of people trying to figure out what weapon to use or buy. I do have a pretty ambitious idea for how this could be implemented in-game graphically, and I will present that idea in another thread that I will link here. Please let me know what you think, especially if you have constructive advice.

[Updated on: Mon, 17 September 2012 00:20] by Moderator

Re: Presentation: Weapon Effectiveness Scales on 6 Dimensions[message #310237] Sun, 16 September 2012 23:15 Go to previous messageGo to next message
Headhunter

 
Messages:268
Registered:November 2009
Location: Sweden
http://images.ientrymail.com/titanquest/smilies/rockon.gif

sorca_2
1. Damage
Limitations: This does not account for ammo types. It could if we think it's necessary.


This would only be necessary if the MAM-mod got implemented in stock 1.13, as right now there isn't any difference between calibers beside availability.


Quote:
6. Compatibility
Limitations: This does not factor in the relative importance of particular attachments. For example, one could argue that grenade launcher compatibility is more important than C-Mag compatibility, etc. Thus, the scale measures how many attachments it can use, not how good those attachments are. I was hoping that naming it "compatibility" would convey this rather than "attachment quality" or somesuch.


I disagree with this, as not all attachments are equal, and not all weapon-types need all attachments. For instance, whether or not sniper rifles can mount a Laser Sight isn't important as the Laser Sight isn't useful at the ranges you would use such a rifle. A Rifle-LAM however is a far more useful attachment. For pistols it's the complete opposite.


And what about Reliability?


As a whole however, this is excellent stuff!

Also, http://images.ientrymail.com/titanquest/smilies/rockon.gif
Re: Presentation: Weapon Effectiveness Scales on 6 Dimensions[message #310243] Mon, 17 September 2012 00:16 Go to previous messageGo to next message
sorca_2

 
Messages:206
Registered:September 2010
Location: California, USA
Headhunter
This would only be necessary if the MAM-mod got implemented in stock 1.13, as right now there isn't any difference between calibers beside availability.

Alright, good.

Headhunter
I disagree with this, as not all attachments are equal, and not all weapon-types need all attachments. For instance, whether or not sniper rifles can mount a Laser Sight isn't important as the Laser Sight isn't useful at the ranges you would use such a rifle. A Rifle-LAM however is a far more useful attachment. For pistols it's the complete opposite.


If we think it's that big of a factor, then the only solution is to assign a "quality" value to each attachment. Since you say it varies by the type of weapon, that would be pretty intense to do. Maybe I'll make some proposals about that later on (and obviously I'm open to suggestions).

Headhunter
And what about Reliability?

Right, I stated early on that reliability and repair ease were not included because they were of minor importance relative to the other six dimensions and I had to draw the line somewhere in terms of what to include.

Think of it this way: how many points of reliability would it take to be equivalent to 10 squares of range or 5 damage or 5 attachments? If the answer is something like "at least 3", then we don't have enough range in Reliability values to be meaningful because the minimum is -2 (there is one -3 and one -4, not enough to factor in) and the max is 4. If we converted that to a 1-7 scale, the distance between 1 and 7 in terms of effectiveness is tiny. The same goes for other stuff like attack volume and repair ease, I propose. Does that make sense?

Headhunter

As a whole however, this is excellent stuff!

Also, http://images.ientrymail.com/titanquest/smilies/rockon.gif

Thank you!
Re: Presentation: Weapon Effectiveness Scales on 6 Dimensions[message #310244] Mon, 17 September 2012 00:20 Go to previous messageGo to next message
Sam Hotte

 
Messages:2033
Registered:March 2009
Location: Middle of Germany
Headhunter
Quote:
6. Compatibility
Limitations: This does not factor in the relative importance of particular attachments. For example, one could argue that grenade launcher compatibility is more important than C-Mag compatibility, etc. Thus, the scale measures how many attachments it can use, not how good those attachments are. I was hoping that naming it "compatibility" would convey this rather than "attachment quality" or somesuch.


I disagree with this, as not all attachments are equal, and not all weapon-types need all attachments. For instance, whether or not sniper rifles can mount a Laser Sight isn't important as the Laser Sight isn't useful at the ranges you would use such a rifle. A Rifle-LAM however is a far more useful attachment. For pistols it's the complete opposite.

It furthermore disregards integrated attachments: Presence of an e.g. integrated retractable stock results in a lower compatibility number resulting in lower effectiveness result. But effectivenes is not influenced at all by the impossibility to attach a stock to this particular gun because it's already attached.
Re: Presentation: Weapon Effectiveness Scales on 6 Dimensions[message #310245] Mon, 17 September 2012 00:26 Go to previous messageGo to next message
sorca_2

 
Messages:206
Registered:September 2010
Location: California, USA
Sam_Hotte
It furthermore disregards integrated attachments: Presence of an e.g. integrated retractable stock results in a lower compatibility number resulting in lower effectiveness result. But effectivenes is not influenced at all by the impossibility to attach a stock to this particular gun because it's already attached.

This is a good observation.

You're talking about attachments that are "built in" to the stats on the weapon, not the "default attachments" that come with the weapon, right?

Is there a flag or field anywhere for "integrated attachments"? If not, then, unless someone can produce a list of weapons with integrated attachments, it might just be a limitation we have to live with.

I might be able to spot them with circumstantial evidence in the bonus values associated with the weapons, like the ReadyTime decrease from one of the stocks. Does anyone at least know what the "integrated attachments" are? Bipod, foregrip, folding stock... ?

[Updated on: Mon, 17 September 2012 00:31] by Moderator

Re: Presentation: Weapon Effectiveness Scales on 6 Dimensions[message #310247] Mon, 17 September 2012 00:43 Go to previous messageGo to next message
Sam Hotte

 
Messages:2033
Registered:March 2009
Location: Middle of Germany
Yes, i meant "built in".
Don't know if there is a flag or field in XMLs to show waht is built in (besides deriving it from the boni).

Almost anything can be built in: Bipod, foregrip, folding/retractable stock, LAM (e.g. OTs 39), Reflexsight (P90), Scope (G36), Launcher, silencer, suppressor, Triggergroup, c-mag (well, there are some LMGs that already hold mags with 100 rds which makes them indifferent to c-mag incompatibility).

To add to difficulty, there are attachments that are incompatible with each other and some that incorporate different things in one attachment.

[Updated on: Mon, 17 September 2012 00:49] by Moderator

Re: Presentation: Weapon Effectiveness Scales on 6 Dimensions[message #310248] Mon, 17 September 2012 00:44 Go to previous messageGo to next message
DepressivesBrot

 
Messages:3736
Registered:July 2009
Not launchers, but other than that, what we call built in is just a bonus tag or two.


Re: Presentation: Weapon Effectiveness Scales on 6 Dimensions[message #310252] Mon, 17 September 2012 01:37 Go to previous messageGo to next message
Headhunter

 
Messages:268
Registered:November 2009
Location: Sweden
sorca_2
If we think it's that big of a factor, then the only solution is to assign a "quality" value to each attachment. Since you say it varies by the type of weapon, that would be pretty intense to do. Maybe I'll make some proposals about that later on (and obviously I'm open to suggestions).


Yeah, generally the fewer the attachments a weapon can take the lower the quality of the attachments, so compatibility is a reasonable way to measure this. I'm just saying it may be an issue.


Quote:
Right, I stated early on that reliability and repair ease were not included because they were of minor importance relative to the other six dimensions and I had to draw the line somewhere in terms of what to include.


Missed that, sorry... :headbanger:


Quote:
Think of it this way: how many points of reliability would it take to be equivalent to 10 squares of range or 5 damage or 5 attachments? If the answer is something like "at least 3", then we don't have enough range in Reliability values to be meaningful because the minimum is -2 (there is one -3 and one -4, not enough to factor in) and the max is 4. If we converted that to a 1-7 scale, the distance between 1 and 7 in terms of effectiveness is tiny. The same goes for other stuff like attack volume and repair ease, I propose. Does that make sense?


If you put it that way then sure, it's not as important, but since both Suppressors (-1) and AET ammo (-3, -1 for 5.7x28) reduce Reliability, it becomes something of an issue for stealth-oriented squads.


Take the MP5SD5, an ultra-quiet weapon (5) with a built-in suppressor and a Reliability value of 0. Right now it looks crazy good, as 0 Reliability isn't much of an issue, but with AET it's -3. Now it is an issue. I've had it jam within one (1!) magazine and that #

[Updated on: Mon, 17 September 2012 01:40] by Moderator

Re: Presentation: Weapon Effectiveness Scales on 6 Dimensions[message #310258] Mon, 17 September 2012 03:22 Go to previous messageGo to next message
sorca_2

 
Messages:206
Registered:September 2010
Location: California, USA
Headhunter
Yeah, generally the fewer the attachments a weapon can take the lower the quality of the attachments, so compatibility is a reasonable way to measure this. I'm just saying it may be an issue.

Ok. If anyone else has any further arguments for this, speak up.

Headhunter
All of the sudden, the Calico M-960A comes up from nowhere.

Damn! I hate when that happens. Uh Oh

Headhunter
If you put it that way then sure, it's not as important, but since both Suppressors (-1) and AET ammo (-3, -1 for 5.7x28) reduce Reliability, it becomes something of an issue for stealth-oriented squads.

For assault rifles, the M4A1 and the HK 416 14" seem very similar but the M4 has a very low Draw speed (5,6) while the 416 is slower at 9,1. They both have a built-in retractable stock and the same shot speed, damage, accuracy, range, attachments. The 416 is a bit more quiet (70, 31,5 w/ suppressor) than the M4 (77, 34,65 w/ suppressor) but it's a tiny difference and frankly not worth it over the draw speed advantage the M4 has.
However, the M4 has a reliability of -1 (-2 w/ suppressor) while the 416 has a reliability of 3 (2 w/ suppressor). This is a big advantage for the 416, especially if you find yourself in a drawn out firefight.


Okay, so it seems like there might be more of a threshold effect: anything over 1 is gravy, 1 is okay, and anything below 1 starts to really suck when going silent. Still, since this mostly applies to stealth ops and/or AET use, I'm not sure if it's worth making the "at a glance" level of analysis more complicated (my original goal was 4 dimensions but I begrudgingly expanded to 6).

Headhunter
Also Grippod for the Ultimax and M60E3.

Argh. It sounds like it's going to be a pain to find all of these.

But let's say that I track them all down--I assume you guys are saying that the solution is to give +1 effective possible attachments per integral attachment. Also, note that this may or may not actually increase the Compatibility score, depending on how many attachments the weapon can already take.

------

UPDATE:
OK, so that wasn't as difficult as I thought it was going to be. Aside from the launchers, I found 167 integral attachments spread across 139 weapons. They included stocks, suppressors, flash suppressors, foregrips, grippods, bipods, reflex sights, and laser sights. Grenade launchers were more difficult to find, so I just did a visual scan of all the weapons and came up with the following list:
AICW
FN F2000
Groza OC-14
XM-29 OICW

Please let me know if I missed any. I will give bonuses to all of those weapons for each integrated attachment and recalculate the Compatibility scale accordingly.

[Updated on: Mon, 17 September 2012 06:27] by Moderator

Re: Presentation: Weapon Effectiveness Scales on 6 Dimensions[message #310259] Mon, 17 September 2012 07:10 Go to previous messageGo to next message
sorca_2

 
Messages:206
Registered:September 2010
Location: California, USA
I updated the compatibility scale by factoring in the integrated attachments and launchers above.
13 weapons lost a Compatibility point in the recalculation as they were "passed by" in the rankings. 37 weapons received a Compatibility boost, accounting for about 27% of all weapons with integral attachments--which more or less fits my earlier prediction, so that's good.

The most notable gains were by:
FAMAE SAF Silenciada
Street Sweeper
XM-29 OICW

...which all jumped 2 points with their integral attachments.

For our amusement, here are the bottom 10 weapons in Effectiveness (score 10 or lower):

Name	        Damage	Range	Accur.	Speed	Bst/Aut	Compat.	Effectiveness
Ruger S. Rdhk	3	1	1	3	0	1	9
Barracuda	2	1	2	4	0	1	10
Grizzly .50	3	2	1	3	0	1	10
Man. MR-73	2	1	2	4	0	1	10
RugerRHawk	3	2	1	3	0	1	10
Automag4	3	2	1	3	0	1	10
Automag5	3	2	1	3	0	1	10
D Eagle .50AE	3	2	1	3	0	1	10
FN Forty9	1	1	1	6	0	1	10
Grizzly .45	3	2	1	3	0	1	10


And here are the top 17 (score 30 or higher):
Name	        Damage	Range	Accur.	Speed	Bst/Aut	Compat.	Effectiveness
AEK-971  	4	5	4	4	7	6	30
AEK-973	        5	5	4	3	6	7	30
AK-108	        4	5	4	4	6	7	30
Colt M16A1	5	5	4	4	5	7	30
FAMAS G2	4	5	4	5	7	5	30
FN Minimi	4	5	5	4	6	6	30
HK 416 14"	4	5	4	5	5	7	30
HK G11	        4	6	3	5	7	5	30
Steyr ACR	5	6	5	5	6	3	30
Tavor TAR 21	4	5	3	6	5	7	30
AN-94 "Abakan"	4	4	6	4	7	6	31
Barrett M468	6	6	5	4	3	7	31
Colt M4A1	4	5	4	6	5	7	31
FN F2000	4	5	5	6	6	5	31
FN MAG	        6	7	5	2	6	5	31
FN SCAR-68 SV	6	6	5	4	3	7	31
Steyr AUG-A3	6	6	4	6	3	6	31


We can think of each of the 6 dimensions 1-7 scale like this:
0: Not applicable--no burst/auto
1: Worst in the game
2: Poor
3: Below Average
4: Average
5: Above Average
6: Excellent
7: Best in the game

Please let me know if you have any other suggestions.

[Updated on: Mon, 17 September 2012 15:18] by Moderator

Re: Presentation: Weapon Effectiveness Scales on 6 Dimensions[message #310268] Mon, 17 September 2012 15:18 Go to previous messageGo to next message
Sam Hotte

 
Messages:2033
Registered:March 2009
Location: Middle of Germany
Going from your description, you missed "triggergroup" as built-in attachment to any gun that is semi, burst and auto capable out of the box.

Just a very subjective point of view at the results:
FN MAG is the bronze medalist in the list whereas HK 21E is not in the top 17. But with OCTH i would always prefer the 21E over the MAG for its (the 21E's) 3rd burst, ability to take 7xscope and less weight.

IMHO in terms of decision guidance "which gun to pick" for a certain role or situation, it's not very helpful to have all guns from smallest pistol to largest AMR in one list.
Wouldn't it be better to compare within the same caliber and/ar "class"; Hence compairing effectiveness between e.g. Sniper Rifles or ARs with 7.62WP or something?
Re: Presentation: Weapon Effectiveness Scales on 6 Dimensions[message #310269] Mon, 17 September 2012 15:49 Go to previous messageGo to next message
DepressivesBrot

 
Messages:3736
Registered:July 2009
So while it does a good job at filtering the junk to the bottom (all ten are handcannons, who'd'a thunk it), the top ranks are heavily biased towards jack of all trades, master of none ARs (the MAG is an interesting outlier here, the Minimi is pretty close to an AR)
You should either break it down by classes or, another interesting idea, give each one 3 or 4 category ratings based on fewer/different attributes and scaled to be comparable. E.g. a P90 might get a 4.5 overall, but a kickass 6.3 for close quarters and a lousy 1.9 as a sniper.


Re: Presentation: Weapon Effectiveness Scales on 6 Dimensions[message #310272] Mon, 17 September 2012 17:08 Go to previous messageGo to next message
sorca_2

 
Messages:206
Registered:September 2010
Location: California, USA
Sam_Hotte
Going from your description, you missed "triggergroup" as built-in attachment to any gun that is semi, burst and auto capable out of the box.

This is interesting because the triggergroup has a 3-round burst and a -3 to-hit penalty. Some weapons have a 2-round burst capability, and I don't think any of the semi+burst+auto weapons have the -3 penalty that comes with the triggergroup. Hmmm...

Sam_Hotte
Just a very subjective point of view at the results:
FN MAG is the bronze medalist in the list whereas HK 21E is not in the top 17. But with OCTH i would always prefer the 21E over the MAG for its (the 21E's) 3rd burst, ability to take 7xscope and less weight.


The FN MAG would be tied for 1st place with 6 other weapons with a 31 total, if you think of it that way. Here are the stats on those two weapons:
http://imageshack.us/a/img832/8718/devmagvshk21.png

The HK21E has 1 more square of range, and they have the same damage, so those aren't important. The MAG can take one more attachment, but that's also not important (it used to take 2 more, but now we're counting the HK21's bipod).

The HK21E has 3 better ReadyTime and .38 better ShotsPer4Turns (12.8 vs 12.42. Unfortunately, this is not a big enough difference to put it into the next Speed category. The average 3 Speed weapon has about 11 ReadyTime--in fact, the HK21 is still on the "slow" end of a 2 Speed rating for ReadyTime.

That leaves accuracy and burst/auto. The MAG has 2 more accuracy than the HK21, so that's where it's getting its biggest advantage.

In terms of auto fire, they fire at the same rate, but the MAG has 1 less auto penalty, which puts it among the top 25% of all weapons in the game in terms of autofire capability. The HK21E has burst fire, but its burst fire is average in terms of shots/AP and burst penalty relative to other burst fire weapons. It gets overshadowed by the MAG's autofire advantage, essentially.

So the 2 key questions are:
1) Is it ok that the MAG is rated higher because of its accuracy advantage?
2) How big of a bonus to the Auto/Burst score should burst capability give relative to single shot/auto weapons? In other words, how good would the autofire capability of a weapon have to be in order for you to pick a single/auto weapon over a single/burst/auto weapon?

Maybe some hypothetical examples using data from MGs in the game will help us:
Name  Shots/Brst BrstPen Auto/5AP AutoPen
MG1	  3	     8	     4      8	
MG2	  3	     7       4	    8	
MG3		             4	    7	
MG4		             4      6	
MG5		             5	    8	
MG6		             5	    7	


This question is for everyone:
Would you take MG6 over MG1? It doesn't have burst, but its autofire penalty is better and it fires more autofire shots per AP. How about MG5 over MG1? MG4?

And what about MG2? Would you pick MG6 over MG2?

Sam_Hotte
IMHO in terms of decision guidance "which gun to pick" for a certain role or situation, it's not very helpful to have all guns from smallest pistol to largest AMR in one list.
Wouldn't it be better to compare within the same caliber and/ar "class"; Hence compairing effectiveness between e.g. Sniper Rifles or ARs with 7.62WP or something?

Yes, I pretty much agree with you. Once we've got this refined, I fully intend to make lists by weapon type and by coolness. I wasn't going to recalculate the scales for those sub-selections of weapons, but now you've convinced me. On the other hand, I still find it meaningful to compare the weapons in one list because you can still easily tell which weapons are the fastest, have the best burst/auto, and the best compatibility. This should be make more sense when I describe my in-game implementation idea after we've got these triggergroup and auto/burst issues ironed out. When we're done, you should still be able to look at the Effectiveness score and say that a weapon with something like a 7-point advantage over another is probably better overall.
Re: Presentation: Weapon Effectiveness Scales on 6 Dimensions[message #310273] Mon, 17 September 2012 17:15 Go to previous messageGo to next message
sorca_2

 
Messages:206
Registered:September 2010
Location: California, USA
DepressivesBrot
So while it does a good job at filtering the junk to the bottom (all ten are handcannons, who'd'a thunk it), the top ranks are heavily biased towards jack of all trades, master of none ARs (the MAG is an interesting outlier here, the Minimi is pretty close to an AR)
You should either break it down by classes or, another interesting idea, give each one 3 or 4 category ratings based on fewer/different attributes and scaled to be comparable. E.g. a P90 might get a 4.5 overall, but a kickass 6.3 for close quarters and a lousy 1.9 as a sniper.

:drool: This is good stuff. Once we've got this Burst/Auto thing resolved (see previous post) and the triggergroup issue, if we want to do that, we'll need to come up with those "roles" (sniper, CQB) and what stats are important to those roles.

Also: what about no semi-auto weapons? Should they just get a Speed penalty since they always have to pay the extra AP? Should they also get a Accuracy Penalty because they always have to pay the AutoFirePenalty?

[Updated on: Mon, 17 September 2012 17:18] by Moderator

Re: Presentation: Weapon Effectiveness Scales on 6 Dimensions[message #310275] Mon, 17 September 2012 17:27 Go to previous messageGo to next message
DepressivesBrot

 
Messages:3736
Registered:July 2009
Some random notes:
Accuracy: It's meaningless under OCTH. It's just a flat bonus and definitely doesn't deserve the same weight as range.
2S/9AP vs. 3S/10AP: I don't see this issue as obvious as you. If I'm going for a backup where every AP counts, 2 rounds to the head for fewer AP rule. If I can't bring 3rds on target reliably, a faster double will be preferable.
MG2 vs. MG6: Depends on the intended role. For a 'real' MG, 5 or 6 would work equally well and beat the others easily. For some kind of ersatz DMR, probably MG2.
EDIT:
Auto Only: The first round has no auto penalty. So if at all, there's a general insecurity about how many AP you'll end up paying. Which may tip the balance for some applications and is meaningless for others.

[Updated on: Mon, 17 September 2012 17:32] by Moderator



Re: Presentation: Weapon Effectiveness Scales on 6 Dimensions[message #310276] Mon, 17 September 2012 17:50 Go to previous messageGo to next message
Sam Hotte

 
Messages:2033
Registered:March 2009
Location: Middle of Germany
IMO the problem is not to be solved by just looking at plain gun's stats. Because especially scopes do somehow overrule a lot:
I don't care about basic accuracy if i can put a better scope on the gun than onto a competitor. Because (the example of HK 21E vs. FN MAG) the Battlescope allows for well aimed burst that brings mostly all bullets to the target over nearly gun's full range when being prone on a bipod. And for that sort of long range combat, ready time is of less importance; it's more important that once i am in position to fire, the scope increased vision range allows for early interrupts and sure shots at a distance.
Naturally, in Close Quarters, it's very different ...
Hence it's preferable (IMHO) to bring 2 specialist weapons along rather than a jack-of-all-trades: use 21E at a distance and switch to CQC gun (PDW, compact AR, SMG, ...) when closing in and entering buildings and similar. It's bottom line more efficient than having the more versatile FN MAG that loses comparison at longe range to 21E and at close quarters to faster and eaier to handle ARs/SMGs/PDWs.
So, yes, what DB proposed should be able to take this "specialist for certain roles" into account.

About Burst/Triggergroup: What i meant is the same thing as with stocks: If the gun already is semi+burst+auto the lack of triggergroup-attachment is no drawback to this particular gun. But its incompatibility with triggergroup attachment gives it a -1 in compatibility, thus decreasing effectiveness, which is wrong.
Re: Presentation: Weapon Effectiveness Scales on 6 Dimensions[message #310280] Mon, 17 September 2012 18:43 Go to previous messageGo to next message
sorca_2

 
Messages:206
Registered:September 2010
Location: California, USA
DepressivesBrot
Accuracy: It's meaningless under OCTH. It's just a flat bonus and definitely doesn't deserve the same weight as range.

Hmm, ok. If we're sure about that, I can remove it from the OCTH scale. How about for NCTH?

DepressivesBrot
Auto Only: The first round has no auto penalty. So if at all, there's a general insecurity about how many AP you'll end up paying. Which may tip the balance for some applications and is meaningless for others.

Ah, ok. So we'll leave Accuracy alone (looks like we're leaving it out anyway for OCTH), but we'll keep the Speed penalty idea unless someone argues otherwise.

Sam_Hotte
About Burst/Triggergroup: What i meant is the same thing as with stocks: If the gun already is semi+burst+auto the lack of triggergroup-attachment is no drawback to this particular gun. But its incompatibility with triggergroup attachment gives it a -1 in compatibility, thus decreasing effectiveness, which is wrong.

Alright--and what about single/burst weapons like the M16A4 or Beretta 93R? We'll basically give a +1 integral attachment to all single/burst/auto and all single/burst weapons, right?

Sam_Hotte
I don't care about basic accuracy if i can put a better scope on the gun than onto a competitor. Because (the example of HK 21E vs. FN MAG) the Battlescope allows for well aimed burst that brings mostly all bullets to the target over nearly gun's full range when being prone on a bipod.

Okay, so we're back to the "attachment quality" factor, which we previously acknowledged is a potential issue--and one that may not be easy to address. I'll make a separate post below about this to get your ideas.

DepressivesBrot
MG2 vs. MG6: Depends on the intended role. For a 'real' MG, 5 or 6 would work equally well and beat the others easily. For some kind of ersatz DMR, probably MG2.

Sam_Hotte
IMO the problem is not to be solved by just looking at plain gun's stats. Because especially scopes do somehow overrule a lot:
Hence it's preferable (IMHO) to bring 2 specialist weapons along rather than a jack-of-all-trades: use 21E at a distance and switch to CQC gun (PDW, compact AR, SMG, ...) when closing in and entering buildings and similar. It's bottom line more efficient than having the more versatile FN MAG that loses comparison at longe range to 21E and at close quarters to faster and eaier to handle ARs/SMGs/PDWs.
So, yes, what DB proposed should be able to take this "specialist for certain roles" into account.

So we're back to the drawing board, somewhat. But it's a good thing... I think we're getting somewhere, and our progress on this will heavily carry over. Will continue in a new post below.
Re: Presentation: Weapon Effectiveness Scales on 6 Dimensions[message #310283] Mon, 17 September 2012 18:57 Go to previous messageGo to next message
sorca_2

 
Messages:206
Registered:September 2010
Location: California, USA
Alright, so we're scrapping the general scales for now in favor of role-specific scales because certain stats are only important to certain roles.

So, from the top: first we need to decide what our "roles" are, then what constitutes a great/awful weapon in that role. We want to have as few roles as possible for the sake of data-reduction, n00b-friendliness, and in-game integration.

At first glance, roles seem to be based on range and number of targets.
1. CQB - Close range, 1 or 2 targets
2. General Purpose (needs a better name?) - Medium range, 1 or 2 targets
3. LMG/SAW/MG - Medium to long range, many targets
4. DMR/Sniper - Long range, 1 or 2 targets
5. Stealth. The only problem is that it isn't really a "role" as much as it is a subtype for the real roles, like Stealth CQB or Stealth DMR. Hopefully our system will allow you to easily see that Good CQB + Good Stealth = Good Stealth CQB.

Can you think of any other roles?
Can any roles be combined?
Must any roles be split? (I sincerely hope not--we don't want to go over 6 and we've got 5 already)

[Updated on: Mon, 17 September 2012 18:59] by Moderator

Re: Presentation: Weapon Effectiveness Scales on 6 Dimensions[message #310294] Mon, 17 September 2012 23:43 Go to previous messageGo to next message
sorca_2

 
Messages:206
Registered:September 2010
Location: California, USA
So keeping in mind our 5 roles, I want to start by getting rid of stats/attachments we can ignore. I propose:
- Damage (not that it isn't important, but it's equally important to all roles)
- ShotsPerBurst (will convert to BurstShotsPerBurstAP)
- Repair Ease (same as damage)
- Grippod (redundant with bipod and foregrip)

So let's start with CQB. There are 30 variables to consider (and that's after removing the above and combining lots of attachments into single variables) and we need to decide which ones are the most important to determine the weights for the scale. Once we can get some kind of agreement on this, I'll calculate it and give the top and bottom results.

Importance 5/5 (vital to the role):
ubReadyTime
ubShotsPer4Turns
APsToReload
APsToReloadManually
Stock - Folding/Retractable
Reflex/Kobra

Importance 4/5 (important to the role):
ubBurstPenalty
BurstShotsPerBurstAP
bAutofireShotsPerFiveAP
AutoPenalty
Foregrip
Trigger Group

Importance 3/5 (helpful to the role):
bReliability

Importance 2/5 (mildly helpful to the role):
bAccuracy
Match Sights
ubMagSize
Magazine Adapter
Rifle LAM, Aimpoint, ISM

Importance 1/5 (barely helpful to the role):
usRange
Shotgun Chokes, Duckbill
Scope 2x

Importance 0/5 (irrelevant to the role):
NoSemiAuto
AttackVolume
Flash Suppressor
Suppressor
Grenade Launcher
Scope 10x, PSO-3
Scope 7x
Scope 4x, PSO-1, ZF-42, No 32, PEM

[Updated on: Mon, 17 September 2012 23:51] by Moderator

Re: Presentation: Weapon Effectiveness Scales on 6 Dimensions[message #310295] Tue, 18 September 2012 00:07 Go to previous messageGo to next message
JAsmine

 
Messages:321
Registered:May 2011
It's obvious that you put a lot of work & a lot of brain into this. :professor: good job.

My 2c-proposal on the variables for CQB are:

- reliability from 3/5 to 2/5 (reliability is good of course, but other variables are just more important)
- Scope 2x from 1/5 to 2/5 or 3/5 (even at shortest ranges, cth can drastically improve with a 2x scope. More cth = more hits = more damage dealt)
- range from 1/5 to 0/5 (in ncth it's unimportant. in octh almost)
- match sights from 2/5 to 3/5 (these are just a very good add-on)


Proposal for additional roles:
- sidearm (they need to be small, light, fast, maybe capable of burst or auto fire and maybe need silencers(?) )

Re: Presentation: Weapon Effectiveness Scales on 6 Dimensions[message #310297] Tue, 18 September 2012 00:46 Go to previous messageGo to next message
Sam Hotte

 
Messages:2033
Registered:March 2009
Location: Middle of Germany
I'd swap 'APsToReload' with 'ubMagSize + Magazine Adapter' as not needing to reload at all is more important than speed of relaoding.

Kick out 'bReliability' as it is equally important to all roles like repair ease.
As there is probably no gun that allows for ISM but would not accept laser and Reflexsight, the IM could be omitted, methinks.

Downgrade 'ubReadyTime' and stocks as CQC not necessarily implies to change stance/heading very often (in game).

(Rod&Spring would be importance 4 methinks as it increases burst/autofire speed IF there were any burst/auto capable guns not accepting R&S at all, that is. Are there any? Dunno.)

Not sure about omitting damage.
Within certain limits damge and speed are interchangeable: If a slow but hard hitting gun is able to inflict the same damage with one shot as a quick 3rd burst from a lighter gun while consuming the same amount of APs - effectiveness would be the same here.
Maybe we could have something like "average damage per X AP" or something?
Re: Presentation: Weapon Effectiveness Scales on 6 Dimensions[message #310299] Tue, 18 September 2012 00:54 Go to previous messageGo to next message
DepressivesBrot

 
Messages:3736
Registered:July 2009
Sam_Hotte
(Rod&Spring would be importance 4 methinks as it increases burst/autofire speed IF there were any burst/auto capable guns not accepting R&S at all, that is. Are there any? Dunno.)
Yes, by definition all guns with weird actions don't accept R&S ... prime example would be the G11


Re: Presentation: Weapon Effectiveness Scales on 6 Dimensions[message #310300] Tue, 18 September 2012 01:06 Go to previous messageGo to next message
sorca_2

 
Messages:206
Registered:September 2010
Location: California, USA
DepressivesBrot
Yes, by definition all guns with weird actions don't accept R&S ... prime example would be the G11

Before I address all the suggestions so far, this brings up one quick question--what do you guys think about the Sci-Fi guns? Should they be included in the scales? I'm not sure how many people play with them (I personally don't, but maybe I should). We could either include them like everything else, or we could formulate the scales without the Sci-Fi weapons and then give them scores afterward. Their crazy stats wouldn't throw off the scales in that case, but you'd still be able to see how well they fit each role.
Re: Presentation: Weapon Effectiveness Scales on 6 Dimensions[message #310331] Tue, 18 September 2012 15:48 Go to previous messageGo to next message
sorca_2

 
Messages:206
Registered:September 2010
Location: California, USA
JAsmine
It's obvious that you put a lot of work & a lot of brain into this. :professor: good job.

Thanks, I appreciate that. My wife has started calling it my "gun thesis" because I should be spending time working on my actual thesis.

JAsmine
- reliability from 3/5 to 2/5 (reliability is good of course, but other variables are just more important)
- Scope 2x from 1/5 to 2/5 or 3/5 (even at shortest ranges, cth can drastically improve with a 2x scope. More cth = more hits = more damage dealt)
- range from 1/5 to 0/5 (in ncth it's unimportant. in octh almost)
- match sights from 2/5 to 3/5 (these are just a very good add-on)

Sam_Hotte
I'd swap 'APsToReload' with 'ubMagSize + Magazine Adapter' as not needing to reload at all is more important than speed of relaoding.


Downgrade 'ubReadyTime' and stocks as CQC not necessarily implies to change stance/heading very often (in game).

(Rod&Spring would be importance 4 methinks...

Changed Scope 2x to 3/5. Kept range at 1/5 because we're only doing OCTH right now and a range 8 weapon would be more limited than a range 13 weapon. Changed Match Sights to 3/5. Switched APsToReload to 2/5 and MagSize and Magazine Adapter to 4/5. Downgraded ReadyTime, Stocks, Reflex/Kobra to 4/5.

Sam_Hotte
As there is probably no gun that allows for ISM but would not accept laser and Reflexsight, the IM could be omitted, methinks.

I combined the Rifle LAM, ISM, and Aimpoint into one variable because they all do the same thing. A weapon would get a bonus (weighted by role importance) here if they could accept at least one of those attachments.

Sam_Hotte
Kick out 'bReliability' as it is equally important to all roles like repair ease.
Not sure about omitting damage.

I rethought this. Damage and reliability should be included and weighted equally in every scale because of the following example: imagine 2 identical weapons, but one has higher damage and/or reliability. That weapon is better at every role than the other one, and that should be reflected rather than ignored. We'll put damage in at 3/5 for each role so it's somewhat important but still overshadowed by the most important aspects of the role. We can put reliability in at 1/5. One exception could be that reliability should matter more for stealth because of the reliability loss with silencers. I'm still keeping Repair Ease out because it's outside of combat.

JAsmine
Proposal for additional roles:
- sidearm (they need to be small, light, fast, maybe capable of burst or auto fire and maybe need silencers(?) )

Duly noted. It seems like it's CQB with a size restriction, essentially.

So here's the revised CQB:
Importance 5/5 (vital to the role):
ubShotsPer4Turns
APsToReloadManually (let's hope I remember to turn this into a penalty...)

Importance 4/5 (important to the role):
ubReadyTime
ubMagSize
ubBurstPenalty
BurstShotsPerBurstAP
bAutofireShotsPerFiveAP
AutoPenalty
Stock - Folding/Retractable
Trigger Group
Foregrip
Reflex/Kobra
Magazine Adapter

Importance 3/5 (helpful to the role):
Damage
Match Sights
Scope 2x

Importance 2/5 (mildly helpful to the role):
bAccuracy
APsToReload
Rifle LAM, Aimpoint, ISM

Importance 1/5 (barely helpful to the role):
usRange
bReliability
Shotgun Chokes, Duckbill

Importance 0/5 (irrelevant to the role):
NoSemiAuto
AttackVolume
Flash Suppressor
Suppressor
Grenade Launcher
Scope 10x, PSO-3
Scope 7x
Scope 4x, PSO-1, ZF-42, No 32, PEM

So, if there are any other suggested changes, speak up and I'll try to get some calculations later today.
Re: Presentation: Weapon Effectiveness Scales on 6 Dimensions[message #310332] Tue, 18 September 2012 15:58 Go to previous messageGo to next message
elenhil

 
Messages:59
Registered:June 2008
Is there a chance of seeing this project result in a new weapon progress xml for NCTH?
Re: Presentation: Weapon Effectiveness Scales on 6 Dimensions[message #310333] Tue, 18 September 2012 16:23 Go to previous messageGo to next message
Sam Hotte

 
Messages:2033
Registered:March 2009
Location: Middle of Germany
sorca_2
Sam_Hotte
As there is probably no gun that allows for ISM but would not accept laser and Reflexsight, the IM could be omitted, methinks.

I combined the Rifle LAM, ISM, and Aimpoint into one variable because they all do the same thing. A weapon would get a bonus (weighted by role importance) here if they could accept at least one of those attachments.

I do not get this reasoning. You omitted the "normal" laser and LAM-200 (because every gun accepts one or has one built-in, i assumed). The ISM replaces this normal laser and adds a reflex-sight, where as Rifle-LAM brings a long range laser (that is not important for this role) and Aimpoint has an even larger laser and an AP bonus inferior to reflex sight.
Hence the ISM would be as important as the reflexsight, IMHO.

[Updated on: Tue, 18 September 2012 16:24] by Moderator

Re: Presentation: Weapon Effectiveness Scales on 6 Dimensions[message #310386] Wed, 19 September 2012 02:26 Go to previous messageGo to next message
sorca_2

 
Messages:206
Registered:September 2010
Location: California, USA
elenhil
Is there a chance of seeing this project result in a new weapon progress xml for NCTH?

Sure. There are plenty of implications for knowing what weapons are best besides general convenience and helping n00bs. NCTH seems to be in flux at the moment (thinking of Cyborg's current work), but maybe this would help speed up the balancing process. We would be able to accurately price weapons, optimize coolness levels, and balance enemy gun drops for starters. It would also let us take a look at modifying current weapons if they are over- or under-performing, and it would help us when adding new weapons. In any event, I'm just trying to get through plain 1.13 OCTH right now because it's the easiest.

Sam_Hotte
Hence the ISM would be as important as the reflexsight, IMHO.

This is a good catch. I moved the ISM from the Rifle LAM/Aimpoint variable to the Reflex Sight/Kobra variable (which is now Reflex/Kobra/ISM).
I'm going to work on calculations and I'll try to post something tonight.
Re: Presentation: Weapon Effectiveness Scales on 6 Dimensions[message #310389] Wed, 19 September 2012 08:54 Go to previous messageGo to next message
sorca_2

 
Messages:206
Registered:September 2010
Location: California, USA
Wow, that took a lot more work than I thought. I had to flip a bunch of the variables around (some of them are penalties where high = bad), and I had to recalibrate for the integral attachments again. I converted each of the variables into a 1-20 scale with zeroes for not applicable. I chose 1-20 because it gives more room for variation than the 1-7 scale I used earlier. For the attachments I simply gave a 0 if the weapon couldn't accept it and a 10 if the weapon could accept it. For auto and burst, I gave a 0 if the weapon had no burst/auto and then scaled the burst and auto stats on a 1-10 scale, added 10 to the result and entered it into the variable. For example, the CAWS has the worst auto penalty, but it still gets an 11/20 for having auto capability at all.

Then I weighted all of that by the importance factors that we determined above and divided by 11.5 to give us a scale with a mathematical maximum possible weighted CQB score of 100. For example, the ReadyTime scale score was multiplied by 4/11.5 and ShotsPer4Turns was multiplied by 5/11.5 . We should be able to use this method to make all of the roles have a maximum of 100 which enables us to discuss cross-role comparisons like "this weapon is 20 points better for CQB than sniping."

The good news is that, with all of that done, we should be able to change our minds about importance of variables without driving me to insanity.

The tables below have 3 numbers: CQB Score (raw unweighted number of "points" the weapon earned in the categories that were not 0/5 importance), CQB Weighted Score (CQB Score with values weighted by importance), and CQB Rank (1-307 rank in order of Weighted CQB Score).

Here are the top 20 weapons:
Name	      CQB Score	Wt CQB	CQB Rank
KAC PDW	        549.6	75.74	1
HK G36C	        568	75.57	2
AEK-973	        582.84	74.43	3
AK-107	        596.5	73.83	4
SIG SG552 Com.	565	72.83	5
AK-108	        592.42	72.76	6
AEK-971	        580.42	72.5	7
SIG SG551	577	72.35	8
AN-94 "Aba	575.08	72.34	9
HK MP5A4	511	71.96	10
HK 53A3	        524.76	71.6	11
SIG SG550	577	71.09	12
HK MG36 RA	569.5	70.96	13
FAMAE SAF	481.5	69.48	14
HK G11	        536.6	69.33	15
HK 23E	        532.5	69.22	16
FAMAE Silen.	471.82	68.97	17
HK MP5SD5	465.5	68.7	18
FAMAS G2	516.66	67.96	19
HK MP7A1	490.82	67.32	20


You can see that the weights made a pretty significant impact here. I'll talk about specific weapon comparisons in a separate post below so we can discuss whether we need to adjust the CQB importance weights.

Here are the bottom 10:
Name	      CQB Score	Wt CQB	CQB Rank
Blaser R 93 L	225.1	17.03	298
Mosin-Nagant	210.41	16.88	299
SV-98	        222.412	16.75	300
M24	        230.6	16.37	301
Steyr SSG-P1	223.82	16.33	302
Sako TRG 42	209.93	15.84	303
Beretta M501	219.8	15.8	304
Erma SR-100	208.96	15.72	305
M40A1	        227.6	15.68	306
KSVK	        205.16	15.01	307


The fact that 25 of the bottom 30 weapons are sniper rifles seems to suggest that our importance weights did a good job of weeding out weapons that would be terrible for CQB.

[Updated on: Wed, 19 September 2012 08:57] by Moderator

Re: Presentation: Weapon Effectiveness Scales on 6 Dimensions[message #310391] Wed, 19 September 2012 09:40 Go to previous messageGo to next message
sorca_2

 
Messages:206
Registered:September 2010
Location: California, USA
To determine if our importance weights are at least in the vicinity of usefulness, let's first compare the #1 weapon with the median weapon, #154. I'm going to talk in terms of the actual game stats instead of the 1-20 scales because I think they'll be easier for everyone to relate to.

Name	        CQB Score	Wt CQB	CQB Rank	RdyTime	Shts/4Trns Damage	Accu	MagSize	APsReload	Range	Brst/AP	BrstPen	Auto/AP	AutoPen
KAC PDW	        549.6	        75.74	1	        2	19.16      30	        5	30	20	        185	0.23	5	3	7
HEZI SM-1	394.7	        46.01	154	        6	14.89      29	        5	30	28	        200				


The KAC PDW readies faster, fires much faster, does more damage, is faster to reload, and has burst and auto capability. The HEZI SM-1 has 1.5 squares longer range. In terms of attachments, the KAC PDW can have (or integrally has) a stock while the HEZI SM-1 can take a bipod that the PDW can't--but bipods aren't important for CQB. I think this demonstrates that our weights are pretty good so far.

For a finer assessment, let's compare #1 against #20:
Name	   CQB Score	Wt CQB	CQB Rank	RdyTime Shts/4Trns Damage	Accu	MagSize	APsReload	Range	Brst/AP	BrstPen	Auto/AP	AutoPen
KAC PDW	   549.6	75.74	1	        2	19.16      30	        5	30	20	        185	0.23	5	3	7
HK MP7A1   490.82	67.32	20	        1	21.67      26      	5	40	16	        160			4	8


The KAC deals 4 more damage, has 2.5 squares greater range, has an obscenely good burst capability and a better auto penalty than the MP7. The MP7 readies 1AP faster, fires slightly faster, holds 10 more rounds, reloads 4AP faster, and has a more rapid autofire than the KAC. The weapons are identical in terms of CQB attachments--the KAC can take a 4x scope, but that isn't reflected in its CQB scores.

This is obviously a much more difficult choice than the one above. All in all, I think the range, damage, and ridiculous burst advantage of the KAC make it the best CQB choice. What do you guys think?
Re: Presentation: Weapon Effectiveness Scales on 6 Dimensions[message #310393] Wed, 19 September 2012 11:43 Go to previous messageGo to next message
DepressivesBrot

 
Messages:3736
Registered:July 2009
Overall, pretty good and predictable result. However, I'd really like an indepth analysis how an LMG managed to trump both the Magpul PDR-D and the FN P90?
However, it's likely that the overall system works and I's just my refusal to let some numbers tell me to use a machinegun in close quarters battles. Or that one's a candidate for 'readjusting outliers' after we got all the use cases covered.


Re: Presentation: Weapon Effectiveness Scales on 6 Dimensions[message #310396] Wed, 19 September 2012 15:10 Go to previous messageGo to next message
Sam Hotte

 
Messages:2033
Registered:March 2009
Location: Middle of Germany
The fact that LMG like MG36 are top ranked might be to the importance factor of large mag size combined with otherwise "normal" AR stats.

The FN P90 not being top rank CQC gun might be due to its integrated reflex sight; perhaps this was erroneously not considered?

Another possible systematical error i could think of is the ISM: Any gun that accepts ISM also accepts Reflexsight (and pistol LAM) - but not vice versa (BTW: Why is this? Is the ISM bound to some specific rail system?). So perhaps guns allowing for ISM got a double "reflex sight possible" bonus (or the other way round, e.g. P90 was double "punished" for its integrated reflex sight)?

But this is just speculation, i have not yet been able to really follow the calculations.
Re: Presentation: Weapon Effectiveness Scales on 6 Dimensions[message #310397] Wed, 19 September 2012 15:51 Go to previous messageGo to next message
DepressivesBrot

 
Messages:3736
Registered:July 2009
I'm not that concerned about the MG36, it's a G36 with a slightly heavier barrel, that's all. It doesn't fall out of line that much compared to the Sig SG550 or the AK-108.
No, I was wondering about the heavy and unwieldy MoFo that is the HK 23E(!)

As for the ISM, yes the better attachments are usually restricted to MIL-STD-1913 ... although with the advent of NAS, there's nothing* superior about the ISM anymore, a weapon can easily carry all it's components at no penalty.



*Might be wrong, haven't looked at the stats lately.


Re: Presentation: Weapon Effectiveness Scales on 6 Dimensions[message #310398] Wed, 19 September 2012 16:04 Go to previous messageGo to next message
Sam Hotte

 
Messages:2033
Registered:March 2009
Location: Middle of Germany
You're right about ISM: with NAS it's even inferior to the separate components (same stats IIRC) as it prevents adding a scope.
Re: Presentation: Weapon Effectiveness Scales on 6 Dimensions[message #310399] Wed, 19 September 2012 16:41 Go to previous messageGo to next message
sorca_2

 
Messages:206
Registered:September 2010
Location: California, USA
Sam_Hotte
The FN P90 not being top rank CQC gun might be due to its integrated reflex sight; perhaps this was erroneously not considered?

Nope, the P90 is receiving a Reflex Sight bonus. I went in and put all of the integral attachment bonuses into the variables.

Sam_Hotte
Another possible systematical error i could think of is the ISM: Any gun that accepts ISM also accepts Reflexsight (and pistol LAM) - but not vice versa (BTW: Why is this? Is the ISM bound to some specific rail system?). So perhaps guns allowing for ISM got a double "reflex sight possible" bonus (or the other way round, e.g. P90 was double "punished" for its integrated reflex sight)?
Based on your observation in an earlier post, I combined the ISM with the Reflex sight variable. But since, as you point out, every gun that accepts an ISM accepts a Reflex Sight, the ISM is basically redundant in terms of our scales. In that variable, guns either got 0/10 for not accepting a reflex sight, Kobra, or ISM and they got 10/20 if they could accept at least one of those 3 (and accepting more than 1 did not improve their score).

In any event, let's check out the more in-depth analysis below to see where the P90 is losing points.

DepressivesBrot
Overall, pretty good and predictable result. However, I'd really like an indepth analysis how an LMG managed to trump both the Magpul PDR-D and the FN P90?
However, it's likely that the overall system works and I's just my refusal to let some numbers tell me to use a machinegun in close quarters battles. Or that one's a candidate for 'readjusting outliers' after we got all the use cases covered.


It's true that some weapons may surprise us in terms of their cross-role effectiveness, but aberrations like this give us a chance to see if our weights are having the impact we want.

Let's look at some mysteriously high-ranked MGs compared to some mysteriously low-ranked PDWs.

Name	CQB Scr	WtCQB	CQBRank	RdyTime	Shts/4T	Damage	Accu	MagSize	RelodAP	Range	Brts/AP	BrstPen	Auto/AP	AutoPen	Reliab	Flash S	Suppres	Stock	TriggrG	Foregrp	Bipod	MatchS	Launchr	Chokes	Scpe10x	Scope7x	Scope4x	Scope2x	Reflex	MagAdpt	RiflLAM
MG36RAS	569.5	70.96	13	17	14.03	30	5	100	24	365	0.25	6	3	7	3	10	10	10	10	10	10	0	10	0	0	10	10	10	10	0	10
HK 23E	532.5	69.22	16	19	14.74	30	5	100	32	360	0.27	6	4	6	1	10	10	0	10	10	10	0	0	0	0	10	10	10	10	0	10
HK 21E	498.14	64.4	35	22	12.8	37	4	100	32	645	0.27	7	4	8	1	10	0	0	10	10	10	0	0	0	0	10	10	10	10	0	10
FN P90	456.78	62.47	44	2	19.44	27	6	50	24	235		6	4	7	2	10	10	0	10	10	0	0	0	0	0	0	0	10	10	0	0
Magpul	450.7	59.1	65	1	22.12	31	5	30	28	275		6	3	7	1	10	10	0	0	10	0	0	0	0	0	0	10	10	10	0	10


Here is the same table showing the difference in weighted points for each variable for the P90 vs. the HK 23E. A positive number means the P90 is gaining points on the 23E, and a negative number means it is losing points to the 23E. A 0 means that the weapons are equal on that variable. Since it's a 0-100 scale, you could think of each difference as "percent better or worse at CQB". If you sum the total difference, it accounts for the entire deficit of the P90 against the 23E.

Name	RdyTime	Shts/4T	Damage	Accu	MagSize	RelodAP	Range	Brts/AP	BrstPen	Auto/AP	AutoPen	NoSemiA	APRelod	AtkVol	Reliab	Flash S	Suppres	Stock	TriggrG	Foregrp	Bipod	MatchS	Launchr	Chokes	Scpe10x	Scope7x	Scope4x	Scope2x	Reflex	MagAdpt	RiflLAM
FN P90	4.2	2.6	-1.2	0.8	-0.3	0.6	-0.3	-5	-6.3	0	-0.3	0	0	0	0.3	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	-1.7


The P90 makes big gains in ReadyTime and ShotsPer4Turns and small gains in accuracy, reload AP, and reliability. The HK 23E gets huge points for being burst-capable (a total of a 11.3 points), moderate bonuses for damage and accepting a Rifle LAM, and small bonuses for having a bigger magazine and a better autofire penalty.

In summary, it's pretty clear that the 23E's biggest advantage is burst fire. Without points from burst, the 23E would fall from being the #16 CQB gun to being the #70 CQB gun. So, the question is: how important is burst and auto? 25 of the top 30 CQB guns have burst, so it's obviously having a huge impact.

EDIT: One solution could be to change the burst and auto scales to 6-20 instead of 11-20 so the gap between having and not having burst is smaller.

[Updated on: Wed, 19 September 2012 16:50] by Moderator

Re: Presentation: Weapon Effectiveness Scales on 6 Dimensions[message #310400] Wed, 19 September 2012 16:51 Go to previous messageGo to next message
Gambigobilla

 
Messages:719
Registered:July 2008
For a minute forget about maths and numbers. Just look me in the eye and say "FN P90 is not one of 20 best CQC weapons."

Re: Presentation: Weapon Effectiveness Scales on 6 Dimensions[message #310401] Wed, 19 September 2012 17:02 Go to previous messageGo to next message
sorca_2

 
Messages:206
Registered:September 2010
Location: California, USA
Gambigobilla
For a minute forget about maths and numbers. Just look me in the eye and say "FN P90 is not one of 20 best CQC weapons."

That's why we're improving the maths and numbers.
Re: Presentation: Weapon Effectiveness Scales on 6 Dimensions[message #310402] Wed, 19 September 2012 17:07 Go to previous messageGo to next message
Sam Hotte

 
Messages:2033
Registered:March 2009
Location: Middle of Germany
That's diffficult to tell. On the one hand it shouldn't make a difference if i bring 2 or 3 bullets to a target by spending X APs with fixed burst or full auto. OTOH non-fixed-burst might result in firing more bullets as intended thus wasting ammo and APs.

So me personally tend to prefer burst, yes (I'd think in CQC suppression is not that important).

Another detail about damage: How about XPB vs. AP ammo? Can make quite a difference if facing spectra clad elites. OTOH if you are able to place almost certain headshots in CQC, XPB's better penetration loses its value, IMHO.

[Updated on: Wed, 19 September 2012 17:09] by Moderator

Re: Presentation: Weapon Effectiveness Scales on 6 Dimensions[message #310403] Wed, 19 September 2012 17:58 Go to previous messageGo to previous message
Headhunter

 
Messages:268
Registered:November 2009
Location: Sweden
How about upping ReadyTime to 5/5 and lowering the Burst-related stuff to 3/5?

[Updated on: Wed, 19 September 2012 17:59] by Moderator

Previous Topic: Underbarrel missile pod
Next Topic: OCTH and NCTH Damage Calculations and Caliber Damage.
Goto Forum:
  


Current Time: Wed Apr 26 12:50:12 EEST 2017

Total time taken to generate the page: 0.02493 seconds