Question for Den

[individual]

I was a bit confused as to why the wiki said the SCAR's range is 1800m to 2200m, while the M4's range is 1900m to 2300m.

Because on your chart, the damage drop values for the M4 and SCAR seem to be identical.

I even made my own "ticks" every 100m and copied a screenshot of the M4 chart over the SCAR, and they line up almost perfectly.

It says:

Damage is near 100% correct with data from the PC version's games_mp.log.

The range at which damage drops though is only approximate.

I was confused.

But THEN I took a look at SnakeX's chart in the forum thread on it:
denkirson.proboards.com/index.cgi?board=general&action=display&thread=613&page=1

...And the M4 and SCAR are more clearly distinguished, and it became clear to me where the range differences were coming from.

Now, I'm asking: Is your chart correct, is it not to scale, is it a more "rough" estimate than Snake's, or am I just reading it incorrectly?

I just want to know which chart to go by, because I can't test damage at various distances, since I don't have the PC version and doing it in a private match with no health recharge, with stun grenades seems like it would take foreeeever. But I'm working on figuring out the TTK at various distances and need the info to continue.

[ssog]

Jan 5, 2010 20:59:55 GMT -5 individual said:

I was a bit confused as to why the wiki said the SCAR's range is 1800m to 2200m, while the M4's range is 1900m to 2300m.

Because on your chart, the damage drop values for the M4 and SCAR seem to be identical.

I even made my own "ticks" every 100m and copied a screenshot of the M4 chart over the SCAR, and they line up almost perfectly.

It says:

Damage is near 100% correct with data from the PC version's games_mp.log.

The range at which damage drops though is only approximate.

I was confused.

But THEN I took a look at SnakeX's chart in the forum thread on it:
denkirson.proboards.com/index.cgi?board=general&action=display&thread=613&page=1

...And the M4 and SCAR are more clearly distinguished, and it became clear to me where the range differences were coming from.

Now, I'm asking: Is your chart correct, is it not to scale, is it a more "rough" estimate than Snake's, or am I just reading it incorrectly?

I just want to know which chart to go by, because I can't test damage at various distances, since I don't have the PC version and doing it in a private match with no health recharge, with stun grenades seems like it would take foreeeever. But I'm working on figuring out the TTK at various distances and need the info to continue.

snakex's chart is a lot more detailed and rigorously tested. I'd go with that one.

[mannon]

I'm a little bothered by the graphs in snakex's chart since they stairstep in increments of 5 damage instead of 1. Den's charts use a steady slope from the min to max range which doesn't take into account any of the decimal truncation, but I find more appealing. I'd prefer if snakex's charts would link each tested point in the data from one to another instead of using the 5 damage steps and creating false plateaus in places where the damage is not actually flat but changing in increments of 1 damage over that range.

I do commend the testing and the effort that goes into the work, though. I would like a little independent verification, but still. Good job.

I'm not sure what Den's are based on. They are possibly approximations based on anecdotal experience and CoD4's range values. He may also be incorporating some of Snakex's information, but remaining more conservative and taking it with a grain of salt.

[Den]

Originally, the lines were based on how the range of weapon classes were set up in COD4.
The M9 and USP were the M9 and USP. (and they still seem to be so)
The Magnum and Eagle were both the Eagle.
The Machine Pistols were all M9-ish.
The rest are the rest.

I updated the lines to more or less reflect snakex's, as his appear to be the most accurate.
Since damage is not incremented in decimals, it drops by a whole point and can stick to each point for a few feet.

[mannon]

Yeah but it isn't rounded down to the nearest number divisible by 5 which snakex's graphs seem to suggest it just drops the decimals and sticks to integer damage values.

Snakex's data seems to be quite good, but the charts are somewhat misleading in that the lines neither match the truncation of decimals to integers nor show the approximation of damage before truncation with strait lines.

I guess you could say I'm glad he's getting the data, but I prefer your charts. ;3

[snakex]

i know what u mean.
but i goes from 200 health and count every shot.
and find the different dmg and writed it.

i can edit the values to 5-0dmg.

[mannon]

I'm just saying I'd prefer that the line be drawn from peak to peak rather than dropping arbitrarily by 5 damage then showing a plateau to the next data point. We know there are plateaus due to the truncation, but the chart does not place them accurately, at every integer, thus it would be more accurate to simply draw the lines from peak to peak and leave the interpretation of truncated decimals to the user, which is how Den's charts generally do it, though he only has the two data points at min and max range.

[mannon]

I'm not sure what you're using to make the charts though. Might be a limitation of your software.

[snakex]

Jan 6, 2010 3:16:26 GMT -5 mannon said:

I'm just saying I'd prefer that the line be drawn from peak to peak rather than dropping arbitrarily by 5 damage then showing a plateau to the next data point. We know there are plateaus due to the truncation, but the chart does not place them accurately, at every integer, thus it would be more accurate to simply draw the lines from peak to peak and leave the interpretation of truncated decimals to the user, which is how Den's charts generally do it, though he only has the two data points at min and max range.

i know now what u mean, lol.
it´s not a limitation from software it´s dmg decrease of every metre
example:
AA-12 on my chart:
the dmg goes from 20 to 17 straight downward. but now the 17 goes not from 5m-6m to 16dmg
it have the dmg by 5m-6m and by 7m it goes down

[mannon]

My suggestion is merely that all the graphs should look more like the MP5K graph instead of dropping 5 damage over 1 meter then proceeding in a flat line to the next data point. Is testversion the same as a free trial?

I might be able to help with Excel stuff. It can do some decent graphs with some finagling.

[snakex]

it´s trial version sry testversion is in german.

[mannon]

You're sure about that huge step down in damage over 1 meter over and over again? I didn't think you would be able to get that good of resolution by merely measuring hits to kill. hmm...

[snakex]

yes.
1 box on the chart is 2metre
1/2 box on the chart is 1metre
i use my method and use Dens trick to find out metre from the train tracks in Derail.
and it´s confirm my method as correct

[mannon]

I just checked on the resolution you can achieve with measuring hits to kill across all four health levels (30, 50, 100, 200).

The values you can ascertain with certainty are the following:
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
19
29

All over values can only be determined with some level of ambiguity, meaning that you cannot tell the difference between some values in a given range. For example all damage values from 34 to 39 will produce the same results.

The ranges of values which cannot be determined individually are as follows:
17-18
20-22
23-24
25-28
30-33
34-39
40-49
50-66
67-99
100-199

I can see if using Painkiller might reduce the ambiguity, but there is a slight issue in that we don't know if damage values get truncated, have painkiller adjust the damage and then get truncated again or if there is only a final truncation at the end. I would hope that all modifiers are applied with the damage kept as a float and then only converted to integer at the end, but I really don't know. Thus every multiplication is an opportunity for truncation where you could lose up to 0.999... damage. Headshot multiplier, Stopping Power, Range, and Painkiller are all such opportunities that can produce decimal values which will be lost.

hmm... Actually I need to double check which values you can check with SP as well, and I suppose you could always use headshots as well, though you need to know the actual headshot multipliers to be accurate, rather than just using an approximation.

The more variables you use to determine the damage values the greater the margin of error due to possibly not modeling the game's damage formula properly, though.

[mannon]

Stopping Power and Painkiller definitely add some more resolution, but there are still gaps. I'm a little reluctant to add Stopping Power+Painkiller or headshots without having a better idea how the values will get truncated when SP and Painkiller interact or having precise headshot multipliers. You would also have to insure every single hit was a headshot and not some other body part which gets more difficult at range, and range testing is the whole point.

Okay with a hits to kill methodology that uses all four health values, (30, 50, 100, & 200) and tests normal damage, stopping power, and normal damage against painkiller, but does not combine stopping power and painkiller or use headshot multipliers the following is true.

Damage Values that can be determined with certainty:
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
29
33
39
50
200

Ambiguous ranges of damage values:
25-26
27-28
30-32
34-35
36-38
40-41
42-44
45-47
48-49
51-56
57-59
60-66
67-68
69-71
72-74
75-86
87-89
90-99
100-101
102-142
143-149
150-199

[snakex]

it can but it´s no needed.
my chart based on the shots to kill of 200 health
when i made ALL different dmg on the chart, it will be overfill.
i made a chart who anyone can see when the dmg drops and when a shot more needed. it´s easier to overview

but also i CAN made so a chart.
i have tested few days ago.
i tested on double health w/ and w/ out stopping power + on 100 health w/ and w/ out stopping power.

[mannon]

I'm not saying I want to see all the values as lines on the chart, just for the diagonal lines on the chart to match the actual damage at that range as closely as possible, though it's not necessary for it to take all the truncation into account.

[snakex]

to adjust the lines several tests needed.
but i think it´s so easier to see when u need more shots.

[mannon]

Hmm... I see. I'd rather see that in a hits to kill chart myself rather than fudging the damage lines to display it.

[snakex]

yes, u must think, the range can only be showed,
first by hits to kill
or second u must find out ALL different dmg values to write it on a 5/0 charts (35,30,25dmg)
but 2nd works only on pc
#
and my chart only give u a clue of dmg
but will update every week
and my methods will be better every time i test.
later u see then the final chart with all correct dmg and dmg lines

[mannon]

BTW it's been mentioned that double health mode does not actually change health to 200, but cuts damage in half, which also means more truncated decimals and will change some of the damage values gathered by testing.

[mannon]

Okay I've recalculated. Double health mode is now counted as 1/2 damage, which means it's yet another chance for truncation. So I went ahead and added in SP+Painkiller observations.

In other words this is for testing all four health modes (30, 50, 100, and "Double") with normal damage, stopping power, normal vs painkiller, and stopping power vs painkiller. I did not include headshots. I assumed in my modeling of the damage formula that all multiplication and division is done first in floating point and then truncated to an integer only in the final step. (This could be incorrect, resulting in even greater truncations.) Even this testing methodology still leaves some gaps.

Only the following damage values can be determined with certainty under this methodology:
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
24
25
26
27
28
29
33
34
35
36
39
42
48
49
50
51
72

The following inclusive ranges can not be reduced to greater precision with this methodology:
22-23
30-32
37-38
40-41
43-44
45-47
52-53
54-55
56-59
60-64
65-67
68-71
73-74
75-77
78-85
86-89
90-99
100+

As far as truncation uncertainties go you're pretty safe with the 30, 50, and 100 damage modes with regular damage and stopping power since the SP multiplier will always result in a whole number due to the base damage values all being divisible by 5. But when painkiller and "double health" get involved you wind up with 1/3 and 1/2 division and truncation on top of any range truncations.

[snakex]

i say u already, that my new project begins.
u see then accurater values.

i made a chart with excact dmg drops
and u know, it´s not my topic
it is for Den