Post by mannon on Dec 6, 2009 7:47:03 GMT -5
When you use a weapon the game hitscans your target for the range and then uses a very simple formula to convert range into damage.
We know this formula. For MW2 we are merely missing some of the variables.
But it occurred to me that it will be consistent even if we only know half the variables.
In other words it's not too difficult to test weapons at their min and max ranges even if we don't know exactly what those ranges are, and now that we can do that and get the actual damage value rather than merely shots to kill we should have all of the damage values clearly defined quite soon. (Might be best to go ahead and have some level 70 reset(delete) their log and then test fire each weapon at min and max ranges with no penetration shots or any other tomwhackery. And probably best to do so exclusively w/o SP since it's effects are known. Such a MP log would be pretty definitive, but that's not the point.
If we know the accurate damage statistics on the weapons then we can in a way turn the formula around and use the damage values in the MP log as something of a range indicator almost.
It won't tell us accurately in inches the range values of the weapons, but what it could do is let us compare their relative range values.
In other words it becomes trivially simple to use the MP log to test two weapons to see if their min and max range variables are the same. The key is simply staying within those ranges, since it tells you nothing if you've maximized or minimized damage. But if you fire a clean shot and get a value between the min and max damage then you know that what just happened was the range affecting damage. You then switch to the second weapon you wish to test and fire it at exactly the same range. You then just need to fill in some more data points at a few more ranges to get good data. Obviously if the two weapons don't do exactly the same damage at every range they they are not identical.
If you get exactly the same damage values testing the guns against each other at first one range, and then a second, AND neither weapon hit it's max or min damage, and the weapons have the same damage stats, then you have verified that the weapons relative ranges are in fact identical.
But you could go further. For weapons that do not have identical ranges as long as you can get a few data points at a few ranges where they overlap you could figure out their relative range statistics.
This would work best with weapons that have identical damage statistics, but fortunately there are plenty of those. It would also be most accurate with weapons that have a greater range between their min and max damage stats and thus would probably work best with SP since the multiplicative effect also multiplies that gap.
It would also be possible to normalize the damage values for weapons that do not have identical damage statistics to a value between say 0% for min damage done and 100% for maximum damage. Thus you could compare 40-30 and 30-20 or whatever weapons so long as you recognize that you will lose some fidelity rather than having perfectly identical percentages for weapons with identical range statistics.
A lot of the testing could simply be done by finding easily repeatable locations such as backing into corners. You could switch classes and kill yourself, as well as repeatedly kill your enemy as he switches classes to drop a variety of weapons in one convenient spot, then assume your positions and test a single range against them all.
I suggest keeping a manual log of what you fired at what approximated distances and especially noting all the hits fired from exactly the same range.
Then when switching range don't start with the last weapon used in case you did not accurately record (or see) the number of actual hits. The weapon change should provide a good place to double check the MP log against your manual one. So then find a new spot at a different range and do it all again. (In fact doing the same weapons in the same order might be the most convenient.)
If any of your shots registered 100% or 0% then those lines won't provide any data and can be discarded, but everything within min and max range for a weapon should provide a data point.
This wouldn't get us the empirical weapons ranges, but if we have enough overlapping weapons ranges we can determine all the relative values.
This might be a bit of a problem for AR's which appear to have their min range with silencer set to the max range for SMG's without, but perhaps not. If we don't get enough overlaps between all groups we may have to simply have a few groups.
Still we could easily assign arbitrary range numbers simply for relative comparisons. If we could possibly link all the weapons via some amount of overlap then we could easily define a maximum value for our arbitrary range such as 100% or 1000 and convert all the min max relative range information into our arbitrary system if we chose. Honestly I think this would be almost as good as having the empirical numbers. What we really want with this data is mainly to compare the weapons to each other anyway.
On the other hand if we have all that nice, neat, relative weapon to weapon data linking the whole thing then we would only need to know one of the empirical range values to convert the entire thing over to the actual numbers, give or take some accuracy due to data fidelity being a bit low. But then again we also know that IW tends to make the range values fall on some pretty big landmark numbers so we may be able to make some assumptions based on that to make corrections for our low data fidelity.
Given the similarities in MW2 and MW and their tendency to use identical values for so many weapons range stats, and the fact that many weapons almost certainly have the same values as before, we could probably go ahead and make some assumptions.
The idea basically is that all we would be doing is plotting points on a graph where we don't know exactly what the horizontal scale is, but we can easily plot the same point on the horizontal for two different weapons and get the vertical (damage) for both of them. Additionally since we know the formula is linear once we have two data points for one weapon we can place any additional point for it somewhere on the graph, along with any other weapon tested at the same exact range. That's why it only takes two ranges. Plot two points on a weapon's graph and you know the slope. Since we know the high and low points we already know the ends of the slopes in Y (damage) and could thus calculate X (range) albeit in some arbitrary (or assumed) unit of measure.
Anyway, it's a thought...
Mainly I'd just like to use this method to verify some of our assumptions about weapons ranges. I think we can already make some pretty decent assumptions based on CoD4 and anecdotal experience coupled with the damage stats, but it would be nice to know for sure at least if the weapons we think are identical in range stats really are.
I'm sorry I don't have the PC version (or have a PC that could handle it) or I'd have done a lot of my own testing already.
We know this formula. For MW2 we are merely missing some of the variables.
But it occurred to me that it will be consistent even if we only know half the variables.
In other words it's not too difficult to test weapons at their min and max ranges even if we don't know exactly what those ranges are, and now that we can do that and get the actual damage value rather than merely shots to kill we should have all of the damage values clearly defined quite soon. (Might be best to go ahead and have some level 70 reset(delete) their log and then test fire each weapon at min and max ranges with no penetration shots or any other tomwhackery. And probably best to do so exclusively w/o SP since it's effects are known. Such a MP log would be pretty definitive, but that's not the point.
If we know the accurate damage statistics on the weapons then we can in a way turn the formula around and use the damage values in the MP log as something of a range indicator almost.
It won't tell us accurately in inches the range values of the weapons, but what it could do is let us compare their relative range values.
In other words it becomes trivially simple to use the MP log to test two weapons to see if their min and max range variables are the same. The key is simply staying within those ranges, since it tells you nothing if you've maximized or minimized damage. But if you fire a clean shot and get a value between the min and max damage then you know that what just happened was the range affecting damage. You then switch to the second weapon you wish to test and fire it at exactly the same range. You then just need to fill in some more data points at a few more ranges to get good data. Obviously if the two weapons don't do exactly the same damage at every range they they are not identical.
If you get exactly the same damage values testing the guns against each other at first one range, and then a second, AND neither weapon hit it's max or min damage, and the weapons have the same damage stats, then you have verified that the weapons relative ranges are in fact identical.
But you could go further. For weapons that do not have identical ranges as long as you can get a few data points at a few ranges where they overlap you could figure out their relative range statistics.
This would work best with weapons that have identical damage statistics, but fortunately there are plenty of those. It would also be most accurate with weapons that have a greater range between their min and max damage stats and thus would probably work best with SP since the multiplicative effect also multiplies that gap.
It would also be possible to normalize the damage values for weapons that do not have identical damage statistics to a value between say 0% for min damage done and 100% for maximum damage. Thus you could compare 40-30 and 30-20 or whatever weapons so long as you recognize that you will lose some fidelity rather than having perfectly identical percentages for weapons with identical range statistics.
A lot of the testing could simply be done by finding easily repeatable locations such as backing into corners. You could switch classes and kill yourself, as well as repeatedly kill your enemy as he switches classes to drop a variety of weapons in one convenient spot, then assume your positions and test a single range against them all.
I suggest keeping a manual log of what you fired at what approximated distances and especially noting all the hits fired from exactly the same range.
Then when switching range don't start with the last weapon used in case you did not accurately record (or see) the number of actual hits. The weapon change should provide a good place to double check the MP log against your manual one. So then find a new spot at a different range and do it all again. (In fact doing the same weapons in the same order might be the most convenient.)
If any of your shots registered 100% or 0% then those lines won't provide any data and can be discarded, but everything within min and max range for a weapon should provide a data point.
This wouldn't get us the empirical weapons ranges, but if we have enough overlapping weapons ranges we can determine all the relative values.
This might be a bit of a problem for AR's which appear to have their min range with silencer set to the max range for SMG's without, but perhaps not. If we don't get enough overlaps between all groups we may have to simply have a few groups.
Still we could easily assign arbitrary range numbers simply for relative comparisons. If we could possibly link all the weapons via some amount of overlap then we could easily define a maximum value for our arbitrary range such as 100% or 1000 and convert all the min max relative range information into our arbitrary system if we chose. Honestly I think this would be almost as good as having the empirical numbers. What we really want with this data is mainly to compare the weapons to each other anyway.
On the other hand if we have all that nice, neat, relative weapon to weapon data linking the whole thing then we would only need to know one of the empirical range values to convert the entire thing over to the actual numbers, give or take some accuracy due to data fidelity being a bit low. But then again we also know that IW tends to make the range values fall on some pretty big landmark numbers so we may be able to make some assumptions based on that to make corrections for our low data fidelity.
Given the similarities in MW2 and MW and their tendency to use identical values for so many weapons range stats, and the fact that many weapons almost certainly have the same values as before, we could probably go ahead and make some assumptions.
The idea basically is that all we would be doing is plotting points on a graph where we don't know exactly what the horizontal scale is, but we can easily plot the same point on the horizontal for two different weapons and get the vertical (damage) for both of them. Additionally since we know the formula is linear once we have two data points for one weapon we can place any additional point for it somewhere on the graph, along with any other weapon tested at the same exact range. That's why it only takes two ranges. Plot two points on a weapon's graph and you know the slope. Since we know the high and low points we already know the ends of the slopes in Y (damage) and could thus calculate X (range) albeit in some arbitrary (or assumed) unit of measure.
Anyway, it's a thought...
Mainly I'd just like to use this method to verify some of our assumptions about weapons ranges. I think we can already make some pretty decent assumptions based on CoD4 and anecdotal experience coupled with the damage stats, but it would be nice to know for sure at least if the weapons we think are identical in range stats really are.
I'm sorry I don't have the PC version (or have a PC that could handle it) or I'd have done a lot of my own testing already.
