[Math/balance] Are Ppcs Balanced Now? A Mathematical Model For Weaposn Balance
#21
Posted 08 February 2013 - 01:45 PM
#22
Posted 08 February 2013 - 01:48 PM
Quote
Depends. If you have ECM yourself then Streaks are one of the best weapons.
Also SRM6s are one of the best, if not THE best weapon, and are unaffected by ECM.
#23
Posted 08 February 2013 - 02:00 PM
StalaggtIKE, on 08 February 2013 - 01:45 PM, said:
Tonnage for ammo is figured into the equation for the DPS/TTon index. I have included the ammo tonnage required for 2.5 minutes of continuous fire. This is somewhat arbitrary, but in my experience, its enough to last through any normal engagement that ends with you surviving and doing considerable damage.
Also, energy weapons have a penalty modifier applied to their expected DPS/TTon just for being energy weapons. This reflects the inherent advantages of energy weapons (no ammo, heat isn't a hard limit, etc.). Lasers have an additionally modifier applied because they're hitscan as well.
Khobai, on 08 February 2013 - 01:48 PM, said:
Depends. If you have ECM yourself then Streaks are one of the best weapons.
Also SRM6s are one of the best, if not THE best weapon, and are unaffected by ECM.
The SRM6, in the current implementation of the weapons, has the highest DPS/TTon of any weapon except LRMs. GRs and UAC5's compete because of their longer range and because of the SRM6's firing mechanics (missile spread).
#24
Posted 08 February 2013 - 02:14 PM
Sadly I lack the motivation and I'm too tired to bring sensible arguments, so I'll just commend you for trying.
#25
Posted 08 February 2013 - 02:33 PM
Redshift2k5, on 08 February 2013 - 10:35 AM, said:
Gut feeling? The LL always was spot-on. Making the PPC and ERPPC more like the LL is perfect.
baddies use gut feeling, because the numbers never lie
#26
Posted 08 February 2013 - 03:23 PM
Abrahms, on 08 February 2013 - 02:33 PM, said:
baddies use gut feeling, because the numbers never lie
Are you referring to the Author of the OP? Not very discussionly-like...
In any case, the OP's numbers are relatively similar to what other Mech titles already figured out. MW3 and MW:LL came to the same type of balancing for certain weapons which are good starting points. I.E., LB-X are far more damaging compared to 'TT' rule numbers, AC/10 - AC/20 are slightly better in damage compared to TT rule numbers, Ultra's are usually given twice as good firing rate (not double shell shot), most missile types have the same or similar cool down, etc.
I agree with those points. The MG was left out of these number changes for some reason OP. MW3 has probably the best balance for the MG doing actual damage but they also gave it a cool down/recycle time for balance.
#27
Posted 08 February 2013 - 04:28 PM
General Taskeen, on 08 February 2013 - 03:23 PM, said:
Are you referring to the Author of the OP? Not very discussionly-like...
In any case, the OP's numbers are relatively similar to what other Mech titles already figured out. MW3 and MW:LL came to the same type of balancing for certain weapons which are good starting points. I.E., LB-X are far more damaging compared to 'TT' rule numbers, AC/10 - AC/20 are slightly better in damage compared to TT rule numbers, Ultra's are usually given twice as good firing rate (not double shell shot), most missile types have the same or similar cool down, etc.
I agree with those points. The MG was left out of these number changes for some reason OP. MW3 has probably the best balance for the MG doing actual damage but they also gave it a cool down/recycle time for balance.
Based on Abrahm's posting history, I don't think he was adressing that to me, but rather to those who trust their gut over numbers.
As for Mguns ... well, I left them out because I haven't the slightest idea what to do with them. Same with flamers.
#28
Posted 08 February 2013 - 04:43 PM
zorak ramone, on 08 February 2013 - 04:28 PM, said:
Based on Abrahm's posting history, I don't think he was adressing that to me, but rather to those who trust their gut over numbers.
As for Mguns ... well, I left them out because I haven't the slightest idea what to do with them. Same with flamers.
I would look at these numbers here:
http://ppc.warhawken...echbay/mw3.html
Its a good starting point.
#29
Posted 08 February 2013 - 04:48 PM
Have you ever tried ac20pult, Zorak? That babe rocks . Don't offer 25 damage to ac20 as that will certainly lead to new op and bad balance. No math is good enough to justify that.
Edited by Alexander Malthus, 08 February 2013 - 04:52 PM.
#30
Posted 08 February 2013 - 04:50 PM
#31
Posted 09 February 2013 - 04:30 AM
zorak ramone, on 08 February 2013 - 12:29 PM, said:
Hmm. If you had some fancy charts like I did, and put all the current weapon numbers on a chart, we could see if our systems share outliers.*
But I noticed one thing - I believe you suggest basically no changes - at worst a heat reduction - to the medium laser. This is pretty much opposite of what I would do - the medium laser is too powerful. But you started from the premise of "these weapons are deemed good, let's take them as baseline". I don't necessarily think this is a good idea - if you raise all underperformers to this baseline, you will probably increase the overall damage output of all mechs, which will make combats faster. This may not be desirable. SO it may be necessary to actually to take something else as a baseline.
*Fancy Charts:
Underlying Calculations as Google Docs: https://docs.google....cXc&usp=sharing
Thread: http://mwomercs.com/...-5th-2013-patch
If you have the Excel Tables for your charts, you could upload and import them to GoogleDocs.
Edited by MustrumRidcully, 09 February 2013 - 04:35 AM.
#32
Posted 11 February 2013 - 11:42 AM
Alexander Malthus, on 08 February 2013 - 04:48 PM, said:
Have you ever tried ac20pult, Zorak? That babe rocks . Don't offer 25 damage to ac20 as that will certainly lead to new op and bad balance. No math is good enough to justify that.
RE the devs using models: If they have one, I'd love for them to share it with us. Some of us, including myself, figured out that the PPCs and LLs needed heat reduction way back in july (back when PPCs were 10 damage/10 heat, and LLs were 8 damage/8 heat)using only very simple math. Meanwhile, the devs have been dropping heat by one point every few months on these weapons. I suspect they'll do it again when they see that PPCs still aren't there. I also suspect that eventually they'll drop it down to the 5 or 6 heat that I'm suggesting here when they see that PPCs still aren't up to speed, baring any drastic changes in game mechanics.
I suspect that either they do have a model, but it has some issues, they have a good model but they're very risk adverse, or they don't have a model, but they do have statistical data on weapons useage, and are buffing when they see underperformance.
RE the AC20 pult: Nope, but I have run dual-gauss pults and dual-gauss CTFs, and I've played against many dual gauss pults. For a mech that packs two huge weapons that preclude the use of XL engines, goes really slow, and requires short range engagement, I think it under performs compared to GR and UAC5/AC5 builds. For example, the 2xGR 4X can pack a pair of MLs, giving it the same alpha, runs cold as ice, and can hit accurately out at 600-700m with 75% of its alpha. Oh and it doesn't need an XL engine to do it either.
shadevarr, on 08 February 2013 - 04:50 PM, said:
You don't even know how clan tech works, huh? Please, tell me how this is clan tech.
MustrumRidcully, on 09 February 2013 - 04:30 AM, said:
Hmm. If you had some fancy charts like I did, and put all the current weapon numbers on a chart, we could see if our systems share outliers.*
But I noticed one thing - I believe you suggest basically no changes - at worst a heat reduction - to the medium laser. This is pretty much opposite of what I would do - the medium laser is too powerful. But you started from the premise of "these weapons are deemed good, let's take them as baseline". I don't necessarily think this is a good idea - if you raise all underperformers to this baseline, you will probably increase the overall damage output of all mechs, which will make combats faster. This may not be desirable. SO it may be necessary to actually to take something else as a baseline.
Way back in July, in closed Beta when we had 4 or 5 mechs and HBK-4SP ML boats ruled the day, I hypothesized that MLs were not, in fact, overpowered because of their damage and heat. I argued that the reason why MLs ruled is because of the the firing mechanics and netcode that made projectile weapons under perform. This was before they increased the speed of all projectiles (they were pretty slow) and before they gave projectiles and "area" (i.e. before the collision box for the projectile was a single point. Now its a box or something).
At the time, I was arguing that high heat was making energy weapons and heavy ACs too weak and that the only reason GRs didn't overpower the MLs is because of the projectile mechanics. Guess what? I was right. Once they implemented those changes (improved projectile speed/hitbox), suddenly the Gauss K2 became dominant.
Moving to the present, my hypothesis is that the only reason people have been bringing lasers at all is because of their ability to hit light mechs and fast mediums. Now, with the improvements in the netcode, I can consistently hit light mechs with groups of GRs/AC5s/UAC5s. I'm sure I'm not the only person who can now, and once people figure this out, they're going to start dropping MLs for more GRs/ACs or GR/AC-supporting equipment.
Currently, the ML has the same DPS/TTon as UAC5 (using average recycle calculation) and less DPS/TTon than the GR and AC5. Its also less than the more comparably ranged AC10 and AC20 (which are inferior to the GR, AC5 and UAC5). Does the hitscan property of MLs make up for the range and DPS/TTon deficiency? With the netcode changes, I'm going to say no.
Quote
*Fancy Charts:
EDIT: how do I post pictures?
Edited by zorak ramone, 11 February 2013 - 11:43 AM.
#33
Posted 11 February 2013 - 11:52 AM
MustrumRidcully, on 08 February 2013 - 12:00 PM, said:
1) There is a maximum amount of weapons and gear you can equip.
2) You get "free" heat sinks. You will have them, whether you want or need them or not. But since you have them, you should utilize them. THis can change a lot. A 35 ton mech with 10 Double Heat Sinks can install 4 Medium Lasers completely consequence-free in terms of heat. That means to get a DPS of 5 with a Medium Laser, you only need to invest 4 tons on this mech. But on the other hand - if you want to increase the DPS of this mech to 10 - it wouldn'T cost you 4 tons. It might cost you 24 (20 DHS and 4 Medium Lasers.) That will probably mean the Light Mech can forget about equpping 8 MLs - but it can also mean that a heavier mech suddenly suffers a sharp efficiency failure.
3) You don't need to negate all heat. A mech has a heat capacity, and you're not gonna fire forever. This makes it again important to consider a weapon in the context of a build. If you had no heat sinks at all, you could fire a Medium Laser for 30 seconds! But with 2, you would only last 15 seconds. So you need to consider how multiple weapons affect the overal heat load of the mech, and figure out how long you realistically need to last.
4) You don't fire your weapons forever. The actual timing of shots can be important. In 4 seconds, a A Gauss Rifle can fire 2 shots for a total of 30 damage. An AC/2 can fire 9 times, for a total of 18 damage. Considering that the theoretical DPS of the AC/2 is 4, and the theoretical DPS of the Gauss is 3.75, this might be unexpected. But it can be important in practice. (And this doesn'T just affect damage - it also affects heat.)
Well, that's ultimately why my mech efficiency charts work very different from the typical "heat neutral" approaches to balancing. I personally think - the more models we have, the more likely we can synthesize one that gets closer to the truth.
What I like for example in zarok's model - an attempt to evalulate things like locks, Scatter Shot and the like, and range. My model just implies - if the range goes up, we want a lower efficiency. But I don't really know yet how much worse it should be. And I still haven't provided something to consider the implications of of scatter effects, hit scan or projectile speeds. I think they should also play a role. I model the advantages of different fire rates in a different manner than zarok (as mentioned above, I use a targeted engagement time and calculate how many shots are really fired. This gives an advantage to weapon with a low rate of fire - and the advantage is exactly the maximum possible level of advancement you could possibly have from it if you're a perfect shot and can really fire uninterruptedly for that time.)
Chain firing heavily effects your energy damage model.
For example my SRM stalker has 6 medium lasers. If it fires all 6 at once i can maybe get 3 volleys off before a 4th would shut me down.
However if I chain fire the mediums quickly enough that 5 are always on cooldown I put out the SAME dps but almost stay heat neutral.
The trade off is its a lot harder to keep 6 separate laser beams on the same location than a single volley.
somethin to think on..
#34
Posted 11 February 2013 - 11:24 PM
zorak ramone, on 11 February 2013 - 11:42 AM, said:
You must upload them somewhere, for example something like photobucket or imageshack or imgur, and then you can use the [ IMG ] [ /IMG ] tags and enter the link between. (it's also a button in the text editor. Just read the tooltips.)
LordBraxton, on 11 February 2013 - 11:52 AM, said:
Chain firing heavily effects your energy damage model.
For example my SRM stalker has 6 medium lasers. If it fires all 6 at once i can maybe get 3 volleys off before a 4th would shut me down.
However if I chain fire the mediums quickly enough that 5 are always on cooldown I put out the SAME dps but almost stay heat neutral.
The trade off is its a lot harder to keep 6 separate laser beams on the same location than a single volley.
somethin to think on..
No, if you're really "almost heat neutral", you are firing your weapons slower than they can. You can't escape addition.
I suspect that the chain fire delay is simply long enough that you are not firing each ML every 4 seconds. And thus, you are also not achieving their full DPS.
zarok ramone said:
I think there is a good reason why people still use medium lasers.
For example, if you use 5 Medium Lasers and have Double Heat Sinks, just with the engine sinks you have a neat heat production of 3 Heat Per Second.
You also have a heat capacity of 50. That means you can fire for 16 seconds before you overheat, dealing a total of 80 damage.
Sure, now you're mech is really hot, but you already dealt 80 damage to your enemy. 2 Gauss Rifle in the same time would have dealt 120 damage, but it would also have costed over 30 tons to achieve that. You just invested 5 tons! No wonder that Garth and Paul love their Cicadas and Jenners.
Most combat engagements are not over theoretical infinite time spans. So it really matters to not overheat too quickly.
This would radically change if their were heat penalties. Then suddenly those 16 seconds might come with such harsh penalties towards the end that you cannot hope to achieve this damge potential, and your 5 tons of weight investment are wasted.
The problem is just - if you, say, go up to 10 tons of medium lasers, this no longer works either. You may have a theoretical DPS of 12.5, but you also gain 8 heat per second, soyou'd last only 6 seconds. Which is only 75 damage - less than you could deal with 5! It'sin a shorter time frame, so it might still be okay if you're a fast striker, but it'll take you 25 seconds before you can repeat it.
THis leads to the "lower heat" energy weapons be in a sweet spot. And it also leads to light and medum mechs in a sweet spot.. Heavier mechs need to invest a lot more weight to raise their damage output (usually in the form of ballistics or missiles). A mech that can't do this, like the Awesome 8Q, is basically screwed.
Edited by MustrumRidcully, 11 February 2013 - 11:37 PM.
#35
Posted 11 February 2013 - 11:36 PM
you need weight, heat, fire rate, range, and size. On top of that you have to understand how much it can be boated, what chassis can carry it, and what weapons can be mixed with it on each chassis. so just comparing one small part of the weapon and saying it is not balanced is somewhat narrowly viewed.
#36
Posted 12 February 2013 - 05:06 AM
#37
Posted 12 February 2013 - 05:28 AM
#39
Posted 12 February 2013 - 05:40 AM
Unfortunately it is difficult to tell which model the devs are actually using. If you knew then I'm sure you could more accurately convey balance on their own terms.
#40
Posted 12 February 2013 - 09:21 AM
zorak ramone, on 08 February 2013 - 09:17 AM, said:
One of the problems with weapons balance discussions is the lack of a consistent method for comparing weapons with different firing mechanics. For example, we know that lasers should overall be less damage / tonnage efficient than a similarly ranged projectile weapon (e.g. LL vs AC10) because of the laser’s lack of ammo and the increased accuracy. The question is … by how much?
I have developed a model for weapons balance based off of the weapons that we know are top tier (GRs, UAC5s, SRM6s, SSRM2s, LRMs) or near top tier (LLs, MLs, SLs). By basing the model on existing weapons that we know are good, it provides us with preliminary experimental validation of the model. First a TLDR on the model:
- Damage efficiency decreases linearly for all classes of weapons by range
- Damage efficiency is boosted or decreased based on negative or positive (respectively) characteristics of the weapon
- The linear scale, and efficiency boosts are based on fitting the model to the top and near-top teir weapons
- Standard ACs: Damage boost for AC10 and 20, reduction in heat for all ACs but the AC5, reduction in recycle for the AC5
- LBX10: Damage boost to 14
- UAC5: Slight recycle increase to bring in line with the AC5
- GR: No change
- Lasers: Reduction in heat to all standard lasers
- Pulse Lasers: Reduction in heat and increase in range to match standard lasers
- PPCs: Reduction in heat, minimum range on PPC removed
- LRMs: Standardization of recycle time and reduction in heat for LRM10
- SRMs: Reduction in heat for SRM2 and SRM4 and standardization of recycle times
- SSRM2: No change
The Model
Overall, damage efficiency should decrease with increasing range. Currently, the decay of damage efficiency by range in MWO best fits a linear model. So, in the model, damage efficiency decreases linearly with range, at a rate consistent with current MWO. Details are below in spoilers:
Altering damage efficiency by weapon firing mechanics
Not all weapons are directly comparable in MWO due to different firing mechanics. Some mechanics, like missile spread on SRMs, are penalties (i.e. a weapon with identical stats, but no missile spread is clearly superior) and others, like the hitscan mechanics of lasers, are bonuses (i.e. a weapon with identical stats, but without laser-hitscan is clearly inferior). Based on weapon mechanic bonuses and penalties, damage efficiency is boosted or decreased. Details are shown below in spoilers:
Rebalancing weapons based on the model
Combining the linear model with adjustments based on weapon mechanics allows us to calculate a predicted damage efficiency for each weapon. I did this using the following guidelines:
- Top tier weapons (GR, UAC5, SRM6, LRMs, SSRM2) remain unchanged or receive slight nerfs
- Near top tier weapons (LL, ML, SL) receive minor buffs
- Base tonnage is never changed
- Final damage efficiency should be within 0.0175 of expected (see model details)
Ballistics
Energy Weapons
Missiles
Adding new weapons
One of the advantages of having a model is that it makes adding new weapons very easy. You already have a guideline of where that weapon should fall. Below are the same weapons charts, but with some CBT weapons that haven’t been added yet, but probably will be later. To do this, I followed the pattern of the other weapons (i.e. how MWO translated them from CBT to MWO) as a starting point, and then adjusted to the model. So, for the new weapons the “old” stats are the direct translation from CBT to MWO, following the pattern of other weapons, and the “new” stats are adjustments to fit the balancing model.
Ballistics
Energy Weapons
Missiles
Advantages of the model
The advantage of the model is that it essentially gives you a testable hypothesis for weapons balancing. The numerical modifiers also make it easy to tweak expectations. If after balancing weapons according to the model they still feel weak or overpowered, then you go back to the weapons mechanics modifiers and tweak them up or down respectively and try again. I’ve already mentioned that it makes adding new weapons easy, but it also makes adjusting current weapons to fit new roles easy. A few examples follow:
Case study: what if PPCs are two powerful according to the model?
Case study: Pulse lasers and infighting weapons instead of accuracy tradeoff weapons
Conclusion
I have presented here a mathematical model for balancing the weapons in MWO. It allows for modifiers based on different weapons mechanics and is easily adaptable in case the predictions of the model don’t end up matching up with reality. This may not be the perfect model, but I think that having a model is a better way to approach weapons balance than by dropping or adding points of heat and damage here and there and hoping you make things balanced. The model as it stands predicts bold changes, but these are changes that are at least internally consistent, and can be changed in an internally consistent manner.
There are some problems with this post. The main one is that it seems to pretend that all of the post is based on math, which it isn't. The formula and the modifers are all based on gut feeling numbers and can very well be wrong. This is fairly obviously why lasers for instance ended up with different numbers despite most people agreeing that lasers in general are fine in MWO.
Another issue is that the model doesn't consider critical slots at all, despite several weapons having that as main balancing factor, like the AC/20. That is somewhat countered with that Ammo is given a rather low modifier and it is mostly ammo based weapons that also takes up many critical slots.
In the end so do I consider it more accurate to model things around only weight and crit spaces by considering heat as simply needing heat sinks to counter and thus ending up also having a weight and crit worth. Thus a PPC+HS should be compared with an AC/10+Ammo+HS. But energy weapons and ammo weapons will end up with different advantages in the end and wont be totally equal. It will simply be innefficient to try and have enough HS to continiously fire energy weapons, but on the other hand so will it be equally innefficient to have ammo for unlimited shooting of an ammo weapon. This is a difference from the TT as there so could you manage both of those, but the double armour and increased rate of fire shaked up that balance.
12 user(s) are reading this topic
0 members, 12 guests, 0 anonymous users