🛈 Announcement
Greetings! The Division forums are now archived and accessible in read-only mode, please go to the new platform to discuss the game
  1. #211
    CategoryTheory's Avatar Senior Member
    Join Date
    Feb 2015
    Location
    Tokyo, Japan
    Posts
    675
    Originally Posted by RichardOshea Go to original post
    Sure, it's just not that relevant imho as the damage done is just that and the number of rounds required is just that.
    I don't understand what you're disagreeing with here. since "damage done" and "number of rounds" is precisely what I'm looking at here. I'm calculating those for different builds in the situations I envision as most common and/or important, and choosing the one that does the most damage. Are you proposing something different?

    I don't agree here as it's not about a given situation or the build at that point but about how a player performs from moment to moment and the math can only guess at that too.
    Right, but what doesn't guess at that? Do your experiments tell you how you're going to perform moment to moment in the future, so you know that even though you're using build A that does 10% less damage than build B, at 17:50 tomorrow build A will perform better than buld B, but at 15:32 the day after, build B will perform better than A?

    Saying that there are some things the math can't predict accurately makes no difference if nothing else predicts them any more accurately. On the other hand, the math is distinctly better at predicting how much damage a bullet from a specific build is going to do when it hits an enemy. ("Better" in that it's both faster and more accurate than experimentation, once the model has been confirmed experimentally to match the game's model.)

    Confining it to the same player doesn't help either as somoene else can copy that build and perform much worse simply because of experience and skill.
    Right. Neither the modelling I'm doing nor the testing you're doing help with that, so I'm not seeing the relevance of this at all.

    I think you may be operating under the mistaken assumption that I'm claiming that my model is the be-all and end-all of everything, or perhaps that if the model is not absolutely perfect, it's useless. Neither is true. The question is: under what circumstances is it better not to use the model to compare two builds? For any case where you're asking, "how much damage does this build do when a bullet hits an enemy?" the answer is "never." (Assuming that the relevant parts of the model have been verified. For those where you're unsure what the game is doing, the thing to do is figure out what the game is doing, not throw some experiments against the wall and ponder the resulting mess.)

    It matters a lot imho and even though I accept that you can describe much with the math its value to a build is only made real once built....
    I don't see what you mean here. There is nothing but math, either in my model or in the game. How is a computer adding and multiplying some numbers somehow more "real" than me adding and multiplying some numbers. (And in fact doing that with the very same computer, as it happens.)

    ...and its effectiveness determined by the skill of the player. There is a nuanced perspective there and I think it's an important distinction to remember.
    I would not call this "nuanced"; I would call this "completely obvious." I think this is the source of our disagreement; you think that I'm trying to do some sort of magic that I'm not. All I'm saying is that I can determine the damage of a bullet hitting an enemy more quickly, more easily and more accurately than you.
    Share this post

  2. #212
    Sircowdog1's Avatar Senior Member
    Join Date
    Mar 2016
    Posts
    3,953
    Originally Posted by CategoryTheory Go to original post
    It sounds like we're in agreement since, as you say above in your "other than" part, there is some form of "special attention" (or whatever you care to call it) they need. I'd put it more precisely as, "bad things happen if you don't kill them before they reach you." (Maybe not all that bad, but at the least an unnecessary distraction from the rest of the battlefield.) I don't know if that's the same thing you're saying or not, but I think we can all agree that it's more often worse to take a couple of extra seconds to kill a rusher than a even a few extra seconds to kill distant enemy in cover.





    I don't think that defensive builds are an issue in this conversation; we're talking about high- or all-red builds here, after all. It's simply the case that this is not an on or off issue; there is no build where without DOOC (or with DOOC, for that matter) you're guaranteed never to have a rusher reach you. You're simply trying to set that probability of that as low as reasonable when balanced against the cost of "I feel it's taking me too long to take out guys in cover," and increasing your DPS significantly against out-of-cover enemies at the cost of slightly lower DPS against in-cover enemies tends to be a good exchange for this, for many people at least..
    All I'm really saying is that I'm not altering my build in any way to take rushers into consideration. Which means that my overall strategy and tactics surrounding that build also doesn't take rushers into account.

    I actually have been waffling back and forth on using DTOOC on my weapons instead of CHD. Since CHD works vs everything in every situation. Even if it mathematically does less damage.

    This way I do optimal damage in more situations. Such as when I have to switch targets away from something more dangerous to stop a rusher.

    But that's just me. I like builds and equipment and tools that are useful in many different situations rather than specialized tools with limitations in other areas.
    Share this post

  3. #213
    Originally Posted by CategoryTheory Go to original post
    I don't understand what you're disagreeing with here. since "damage done" and "number of rounds" is precisely what I'm looking at here. I'm calculating those for different builds in the situations I envision as most common and/or important, and choosing the one that does the most damage. Are you proposing something different?
    Trusting the math is precisely what I'm doing, and it produced that damage for each test. I'm not disagreeing with you at all; yep you can do the math for that stuff; yep it's useful info; yep it can help you improve a build etc etc but the results are what they are, so trusting the math means I should trust the result. The CHC and CHD are specified in the build anyway with only the latter altering between tests. Again, I'm not really disagreeing with how, or why, you're doing it. What's necessary to answer the simple question of which method kills an out of cover armoured target faster is given an answer by the tests shown in that video and that's described accurately by how many rounds it took and how long it took to do it.

    Originally Posted by CategoryTheory Go to original post
    Right, but what doesn't guess at that? Do your experiments tell you how you're going to perform moment to moment in the future, so you know that even though you're using build A that does 10% less damage than build B, at 17:50 tomorrow build A will perform better than buld B, but at 15:32 the day after, build B will perform better than A?

    Saying that there are some things the math can't predict accurately makes no difference if nothing else predicts them any more accurately. On the other hand, the math is distinctly better at predicting how much damage a bullet from a specific build is going to do when it hits an enemy. ("Better" in that it's both faster and more accurate than experimentation, once the model has been confirmed experimentally to match the game's model.)

    Right. Neither the modelling I'm doing nor the testing you're doing help with that, so I'm not seeing the relevance of this at all.

    I think you may be operating under the mistaken assumption that I'm claiming that my model is the be-all and end-all of everything, or perhaps that if the model is not absolutely perfect, it's useless. Neither is true. The question is: under what circumstances is it better not to use the model to compare two builds? For any case where you're asking, "how much damage does this build do when a bullet hits an enemy?" the answer is "never." (Assuming that the relevant parts of the model have been verified. For those where you're unsure what the game is doing, the thing to do is figure out what the game is doing, not throw some experiments against the wall and ponder the resulting mess.)

    I don't see what you mean here. There is nothing but math, either in my model or in the game. How is a computer adding and multiplying some numbers somehow more "real" than me adding and multiplying some numbers. (And in fact doing that with the very same computer, as it happens.)

    I would not call this "nuanced"; I would call this "completely obvious." I think this is the source of our disagreement; you think that I'm trying to do some sort of magic that I'm not. All I'm saying is that I can determine the damage of a bullet hitting an enemy more quickly, more easily and more accurately than you.

    So we both agree that paper performance doesn't necessarily translate to in-game performance. And we both seem to agree that it's tautological to say that a build with potential X run by a player with potential Y produces damage Z. We agree it's obvious. So what I'm trying to get at is that because of that we can't make overgeneralisations about a build e.g. "Best AR build ever! Melts NPCs.". Because the theory has a varying degree of value until you mirror it precisely at which point everything is in agreement. Doing that requires building it, and sure you can exploit some stuff and go from gear zero to mathematical hero in a few hours, but you're still bound by your own Z in the above so it would be an error to take the theory as anything more than a description of an ideal that you may or may not achieve.

    I don't think we disagree on as much as you feel we do -- my god I've asked for it now Our approaches are different but I have no problems with your methods or with what your math is telling you. I just think it's way more complicated than the math alone. Way more than I could tackle on my own or for what my skill set would allow.
    Share this post

  4. #214
    Oatiecrumble's Avatar Senior Member
    Join Date
    Jan 2016
    Location
    London
    Posts
    4,850
    Originally Posted by CategoryTheory Go to original post
    I suspect, though, that if you have two incoming rushers, you don't kill one and then turn to killing a distant controller before coming back to the other rusher. If you do both rushers first, that seems to me to be setting them at higher priority than controllers.

    All rushers are top priority, basically anything that pushes towards me at a fast speed must go down.

    Are you sure that RC vehicles don't receive the DOOC bonus? At any rate, I'm not sure how much it matters for them, since they seem explode with one hit. Drones, however, are another matter; on Heroic they often take several shots to kill, so it would be nice to know if they receive DOOC damage or not.

    Not sure, they die too quick to take notice....BUT they are not my main concern, my main concern is the RC controller.
    See above
    Share this post

  5. #215
    CategoryTheory's Avatar Senior Member
    Join Date
    Feb 2015
    Location
    Tokyo, Japan
    Posts
    675
    Originally Posted by RichardOshea Go to original post
    So we both agree that paper performance doesn't necessarily translate to in-game performance.
    No, we disagree there. I say that paper performance translates exactly to in-game performance. If my calculation says that a bullet will do X damage when it hits an enemy, I'm very convinced that the game will produce exactly that result. If the paper calculation says that you will get Y damage overall if Z percent of your targets are out of cover when the bullet hits them, then the game will do exactly that.

    The problem here seems to be that you think there are things on that paper that are simply not there, such as, "when you next play mission X, what percentage of your shots will miss?" That's neither in my calculations nor yours (though in mine that can be put in as an estimate, if you like, which you can't so easily do experimentally). So I don't see how it's relevant. Neither of us predict the exact future, nor, as far as I know, are either of us claiming to do so. (I certainly am not.)

    So what I'm trying to get at is that because of that we can't make overgeneralisations about a build e.g. "Best AR build ever! Melts NPCs.".
    No, of course not. But you can still get more useful information out of running the numbers, such as "X will do better than Y if more than Z% of your hits are to armour." If Z is 10%, or 90%, that's very useful information that won't be shown by just shooting health and armour and comparing the two values.

    ...so it would be an error to take the theory as anything more than a description of an ideal that you may or may not achieve.
    This is another source of confusion, I think. There are no ideals here (except what you keep trying to bring to this). I just say "this happens under these conditions." You're absolutely correct that, for this information to be applicable to all players you must separate that from the player's choice about what conditions to meet and the value judgements about whether "this happens" is desirable or not. So don't assign things like "player X."

    Let's sum that up again: I am never saying things like an unqualified, "a player will do this much damage." You should assume all the reasonable and obvious qualifications such as, "the damage will be this much from this bullet only if the player hits the target with it." You'll note that the game makes the exact same assumptions on the inventory screen when it shows you things such as the damage of a weapon.

    I just think it's way more complicated than the math alone.
    It is when you're trying to answer questions that I never asked, yes. But from the very beginning I've left things like "what's your ratio of out-of-cover to in-cover hits" as blanks to be filled in by the user, not something my model supplies.
    Share this post

  6. #216
    Sircowdog1's Avatar Senior Member
    Join Date
    Mar 2016
    Posts
    3,953
    .

    Originally Posted by CategoryTheory Go to original post
    I don't think that defensive builds are an issue in this conversation...
    I mainly bring it up because I see a LOT of people extolling the virtues of all blue builds, or bulwark/liberty builds over on reddit. And of course people in game misunderstand or misuse it. Usually resulting in doing garbage damage and getting killed slowly, but repeatedly in groups.

    This is off topic slightly, but by way of contrast makes sense: Division 2 does not reward defensive play or builds in an equitable manner. You CAN do it if you think it's fun. But players would be better served by using more damage-focused builds and getting better at cover and positioning to keep them alive.
    Share this post

  7. #217
    N3mB0t's Avatar Senior Member
    Join Date
    May 2016
    Posts
    2,911
    what i take from this thread is that some people overthink stuff too much , and in trade play too litlle , and sometimes experience beats math , just saying.
     2 people found this helpful
    Share this post

  8. #218
    Sircowdog1's Avatar Senior Member
    Join Date
    Mar 2016
    Posts
    3,953
    Originally Posted by N3mB0t Go to original post
    what i take from this thread is that some people overthink stuff too much , and in trade play too litlle , and sometimes experience beats math , just saying.
    It's true. I just did a 4 man heroic run of bank HQ. Literally did more damage and got more kills than the other 3 players combined with their silly Scorpio/Firewall shield builds. And they all had over 1000 SHD levels to my 200.

    Using optimal builds and math is good. But knowing how to play those builds is better.
    Share this post

  9. #219
    Oatiecrumble's Avatar Senior Member
    Join Date
    Jan 2016
    Location
    London
    Posts
    4,850
    All players need to play this game is:

    Decent build that synergizes well
    Decent playstyle
    Know the enemies
    Know their spawn spots

    Don't think maths plays a major part of the game due to most of it being quite obvious......over thinking might gain you a little bit more dps but that extra dps is hardly a game breaker.
    Share this post

  10. #220
    CategoryTheory's Avatar Senior Member
    Join Date
    Feb 2015
    Location
    Tokyo, Japan
    Posts
    675
    Originally Posted by Sircowdog1 Go to original post
    Using optimal builds and math is good. But knowing how to play those builds is better.
    And who has ever argued otherwise on this thread?

    Originally Posted by Riflemania Go to original post
    Don't think maths plays a major part of the game due to most of it being quite obvious......over thinking might gain you a little bit more dps but that extra dps is hardly a game breaker.
    That DPS can easily be 30% or so, if you have a look at an Ongoing Directive build that was recently posted to this forum.
    Share this post