3 replies to this topic

[Vlad Ward]

Introduction

I was doing a bit of tinkering on smurfy-net the other day and I noticed something interesting. For some reason, taking a Mech with, say, 3 MLs and 3 SRM6s and adding 3 DHS and 3MLs to it decreased its Maximum Sustainable DPS (hereby abbreviated MSDPS). Now, this seemed a bit silly to me, since in a worst case scenario I ought to be able to choose to only fire 3 of my MLs with my SRM6s and have more heat dissipation overall, leading to a higher MSDPS than the previous model.

After looking into the calculations that most of us, myself included, have been using for MSDPS, I found the problem. In general, we assume that a player is trying to fire all of their weapons as often as heat will permit. What this does not take into account, however, is the fact that some weapons do more damage per heat than others, and a player may choose not to fire some of their weapons. This method thus leads to a less than accurate representation of a Mech's potential DPS in the hands of a good pilot.

Proposed Changes

Rather than assuming that a pilot will try to fire all of his weapons as often as possible, I propose a methodology which identifies the best sustainable firing pattern concurrently with the true maximum sustainable DPS of a given machine.

To start, the total heat dissipation of a Mech must be identified. I'll use one of my builds as a template, seen here.

Heat dissipation with 22 DHS and Master-level perks is [ (0.2 * 10) + (0.14 * 12) ] * 1.15 = 4.23 heat/second

The next step is to arrange each weapons system on the Mech in order of Damage per Heat ratio and identify their generated heat per second. This would look like:

SRM6 x4 -- 4.5 Damage/Heat each -- 1 Heat/Second each
ML x6 -- 1.25 Damage/Heat each -- 1 Heat/Second each

Then, weapons systems should be pulled from the top of the list until the HPS of the total firing pattern equals the total heat dissipation of the Mech. For example,

SRM6 x4 -- 4 Heat/second, with a remainder of 0.23 heat/second
ML x1 fired at (4 / 0.23 = ~17) 17 second intervals, or approximately every 4 cycles of SRMs.

This gives the example STK-5S a MSDPS of 15.23, as opposed to the originally proposed 8.32.

Conclusions

1) Changes in methodology would have no effect on the theoretical MSDPS of Mechs who only use a single weapon type (Laser boats, Missile boats, Gauss/AC20 cats, etc) as the difference in results stems from the varying Damage/Heat ratios of different weapons systems.

2) There are probably only about 5 people on the face of the planet who care to read this. That said, I don't really mind posts in the vein of "TL;DR" and "DPS is a worthless stat in MWO anyways", as they serve to keep the thread bumped until one of those 5~ish people gets a chance to read it.

Tally ho, good morrow, and so on and so forth.

[Vapor Trail]

Is this in service of the "firepower" and "heat efficiency" stats in the Mech Lab?

I'd agree with this. Trick will be building the algorithm that calculates this based on any weapon combination on any chassis. Not that's particularly difficult, just a bit of a tricky code.

It would give a better picture of exactly what the maximum heat and damage outputs per unit time (rather than instantaneous maximums) for a given mech loadout... which is ultimately a better picture of the mech's performance that currently. And might help a bit with decisions like: do I put 4 MPL on my 6SRM6 Stalker... 4 ML and 4 DHS... or 6 ML and 2 DHS.

[peve]

Yes, you have a point. Now I would like to take it even further.

There are two forms of DPS: theoretical maximum DPS and actual DPS. Latter depends on the weapon, distance and pilots accuracy.

For example, I do an average of 3.5 points of damage per UAC5 round fired. I have not tested any other weapons yet, but I plan to.

I would imagine that my true DPS with missiles other than streaks would be even lower in relation to UAC5.

So, if you want to tweak your mech even more, find out whatyou can do with different weapons and consider that when calculating your average sustainable DPS

[TheFlyingScotsman]

Thanks for bringing up this excellent math based point. As a theorycrafter, I love to get all the equations out into perspective.
This all supports my long held position on how a mech (or any player in any game) is capable of affecting the fight through what I call "presence." AKA how much you contribute to your team.

The following is a little TLDR, but should give players some insight in how your actions contribute to you winning. At least beyond what is obvious. In a very loose and easy to understand paraphrasing, presence is determined by how much effectiveness a mech can have towards actual results of the game within any period of time in which others are "present". Presence is also heavily affected by communication. The ability to read and react appropriately to information can lead to dramatic increases to battlefield presence.

Your presence is equal to your ability to inflict damage or reduce the amount of damage being inflicted on you or your team over the period of time necessary for positive results such as victory or a kill to occur. Typically, the time it takes for any team to run out of players, the amount of time necessary for a cap win, or the time limit will be this time period. Periods of ineffective action are directly proportionate to presence based on the result of those actions, even if you spend only a few seconds not doing anything useful, such as dissipating heat.

Most importantly, presence is multiplicative. Two 'mechs fighting together have a greater presence than the sum of their solo presence, often in dramatic ways. A lone assault is never as dangerous as an assault with a companion. LRM boats and Lights are easy enough to deal with alone, but throw a tag on the light and get caught by surprise, and suddenly you may as as well have a 140kmh stalker behind you.

Certain weapons or equipment add more presence per ton simply by how much range they have, or how easily they can hit. This is because mechs whose presence is intented to be limited by their low speed are only balanced in such a way when their speed prevents them from continuing to be useful. (Very much rarely the case in MWO, or any game with ranged weapons.) Thusly, longer ranged and easier to strike with weapons produce more presence than equally damaging weapons with less range or more difficult targeting. The presence of a weapon is proportionately increased by the power of other weapons capable of hitting at the same time as it and their ability to fire continuously. Thusly the multiplicative presence of a mech increases with size more than with speed, so long as the firing mech can strike its target. (Lights, prepare for a rude awakening when the netcode is fixed.)

For example, an assault with maxed firepower and armor that is constantly firing and hitting will have more presence than the same assault firing constantly and missing half the time, or a mech that is the same size and fp, hits just as often, but dies faster or is forced to hide and stop firing. Likewise, being entirely away from the fight can turn your presence (Or effectiveness over time) down to zero, unless you are capping the base or caps faster than the other team can outcap your own. In addition, unless you can guarantee that the cap will result in a win more reliably than fighting instead, your presence is still zero. If you fail to win by cap or contribute to an actual win, your presence is zero. If you fire on an enemy the entire game but never destroy them or any armed/useful components, your effectiveness is still zero unless and until your contribution leads to destruction or a win. No amount of damage is actually useful unless it results in a kill or component destruction by either you or your team. (Or FF among enemies, for that matter.)

Beyond this is just a TLDR example of a presence net. If you want a more specific example of how presence percolates throughout a battlefield, read on. Otherwise, prepare to be bored.

You fire on an opponent who is partly in cover, but can still fire at you or teammates. One of many results can stem from this.
If you miss, your opponent may continue firing or notice your attack. If you fire and miss them, and they keep firing and hit your team, you have contributed no measurable presence.If you hit them and they keep firing, and miss your team, you have added presence. If you fire, hit them and they continue to fire, hitting your team, you only have added presence if you deal more damage to them than they inflicted, unless your target is eventually destroyed. If you fire and hit OR miss, and they move into cover, you have added presence, but moreso if you strike. If they hide but can and do continue firing from their cover and either make hits or cause reduced presence among your team, you have not gained presence. If they simply return fire by lobbing LRMs at you, their presence remains unaffected beyond any loss of damaging weapons fired by your target that were in range and had line of sight to you or your team. Assume you take full damage from the LRMs, and sent down into red. If you are then able to go an destroy several opponents without dieing all game, the damage from the LRMs may well have had a negative presence for the LRM boat, as it may have reduced the >effective< damage they could have otherwise dealt with the lost missles or by firing at one of your teammates and actually killing them or reducing their presence. If you take the damage and are forced to run off and hide or risk instant death but still lose, they may as well have had enough presence to equate to having killed you.


I'm sure I've missed a point or two about presence, but this more or less sums it up.

Edited by TheFlyingScotsman, 19 January 2013 - 12:31 PM.