43 replies to this topic

[Asmudius Heng]

VanillaG, on 25 March 2014 - 02:11 PM, said:

What happens if a LRM obtains a lock via LoS , launch your LRMs, and you lose LoS; do the LRMs slow down?
What if you obtain a non-LoS lock, launch the LRMs and then you get LoS before they hit; do the LRMs speed up?

What you are essential proposing is "Ghost LRM speed" since there is no way to know exactly how fast your LRMs are going to be. The easiest solution is tone back the speed a little without introducing yet another overly complicated mechanic to the game.

This is why missile spread is better ... it can change mid flight depending on your LoS

[Rebas Kradd]

Sandpit, on 25 March 2014 - 01:34 PM, said:

If you agree that 6 days of selective match observation, during a tournament, while players are using them in extremely high and unusual (as in not the norm) numbers, is "good" data collection method and how anyone who understands statistics would do it, well then ok.

I'm not going to bother pointing out exactly why each and every single one of those factors skews the data, let alone how they all combine to make it even worse.

I'm a math teacher, so don't talk down to me about statistics. You are making massive assumptions here. You don't actually have even the slightest knowledge of how they gather telemetry. And while I don't need to believe PGI is perfect, whenever you accuse someone of lacking basic high school math from such a position of unawareness as you stand on...well. Such easy and cartoonish assumptions don't stand well with me. You're a person talking about another person, not about the handy idiot that our minds tend to create when we think across the internet gap.

Let's work this out logically.

a) The "patch testing" period is not set at six days. A lot of people noticed the LRM flood tapering off after 2 or 3.

If LRMs were overbuffed, then it would stand to reason that their high usage WOULD REMAIN after the first week of the patch, so Paul would be waiting for a tapering-off that would never come.

c) Tournament time is not necessarily a guarantee that LRMs will rule. I recognize a number of the faction winners as PPC/AC meta-numbskulls.

d) Not every match features high LRM usage. I've played plenty that were pretty quiet on that front, and some that weren't. There are a number of matches that, with a little (gasp) selective observation, could be used for reliable analysis. You're just conveniently assuming that PGI didn't think of that.

e) By the way, matches that are flooded with LRMs still have their own informative value.

f) Part of the announced upcoming changes, screen shake reduction, is hardly a statistical thing and relies mostly on observation.

g) The announced change has been relatively small. It's not like they're undoing the patch. Your reaction is out of proportion to the situation.

Edited by Rebas Kradd, 25 March 2014 - 03:00 PM.

[DeathlyEyes]

Agent 0 Fortune, on 25 March 2014 - 02:34 PM, said:

Paul’s comments lead me to believe that he was seeing hill-humpers and pop-taters (high Elo players) taking the brunt of the missile strikes (players less inclined to change their tactics), while players with less specialized builds avoided the LRMs in cities and tunnels. If this is the case, then of course PGI is going to attempt to build on this rock-paper-scissors dynamic. This also means they will probably not have separate speeds for direct and indirect fire, otherwise the LRM advantage is lost.

Are you suggesting LRMs should work on players behind cover? If that's the case what is the point of cover?

[LauLiao]

Solahma, on 25 March 2014 - 02:41 PM, said:

I've always liked that idea, but not worth the time for them to look into it or implement it. It doesn't change how fast it travels from Point A to Point B if it is still a designed speed, just on a curve.

Yeah but it would at least give short-to-mid range targets a little more warning to hopefully reduce the whining, but still reduce over-all flight time to the most distant targets which is really where LRMs needed the buff anyway.

And I don't know about "not worth the time" but I don't know that I trust them to get it right without breaking something else.

[Solahma]

LauLiao, on 25 March 2014 - 03:17 PM, said:

Yeah but it would at least give short-to-mid range targets a little more warning to hopefully reduce the whining, but still reduce over-all flight time to the most distant targets which is really where LRMs needed the buff anyway.

Oh okay, I see what you mean. Yeah that is definitely a different take on it. The LRMs have to build up velocity due to acceleration. Makes sense that the farther away they are, the faster they will go. I like that idea.

[LauLiao]

Solahma, on 25 March 2014 - 03:20 PM, said:

Oh okay, I see what you mean. Yeah that is definitely a different take on it. The LRMs have to build up velocity due to acceleration. Makes sense that the farther away they are, the faster they will go. I like that idea.

Thank you. Also, I'm shocked that any discussion of LRMs has gone on as long as this while remaining civil.

[VanillaG]

SLDF DeathlyEyes, on 25 March 2014 - 03:10 PM, said:

Are you suggesting LRMs should work on players behind cover? If that's the case what is the point of cover?

Just because it is cover for direct fire doesn't mean it is cover for indirect fire. The only change is that the missiles get there faster so they have less opportunity to dodge them or find indirect fire cover. If someone popped a UAV over their position or they had a spotter in LoS they would have the exact same issue.

[Sandpit]

Rebas Kradd, on 25 March 2014 - 02:58 PM, said:

I'm a math teacher, so don't talk down to me about statistics. You are making massive assumptions here. You don't actually have even the slightest knowledge of how they gather telemetry. And while I don't need to believe PGI is perfect, whenever you accuse someone of lacking basic high school math from such a position of unawareness as you stand on...well. Such easy and cartoonish assumptions don't stand well with me. You're a person talking about another person, not about the handy idiot that our minds tend to create when we think across the internet gap.

Let's work this out logically.

a) The "patch testing" period is not set at six days. A lot of people noticed the LRM flood tapering off after 2 or 3.

If LRMs were overbuffed, then it would stand to reason that their high usage WOULD REMAIN after the first week of the patch, so Paul would be waiting for a tapering-off that would never come.

c) Tournament time is not necessarily a guarantee that LRMs will rule. I recognize a number of the faction winners as PPC/AC meta-numbskulls.

d) Not every match features high LRM usage. I've played plenty that were pretty quiet on that front, and some that weren't. There are a number of matches that, with a little (gasp) selective observation, could be used for reliable analysis. You're just conveniently assuming that PGI didn't think of that.

e) By the way, matches that are flooded with LRMs still have their own informative value.

f) Part of the announced upcoming changes, screen shake reduction, is hardly a statistical thing and relies mostly on observation.

g) The announced change has been relatively small. It's not like they're undoing the patch. Your reaction is out of proportion to the situation.

let's clarify a couple of things
Math =/= to statistician
Math =/= ESPECIALLY social statistician
balance = math
decisions on how to balance =/= math (in a simplistic form that we're talking about)


I DO have knowledge of how they collected telemetry because they TOLD US how they did it. Paul posted it.
6 days
a very small sample size of matches reviewed (these are NOT assumptions, these are directly stated by the person collecting the data)
Data reviewed while in an upward trend right after the release of a new buff (in other words they should have waited and reviewed data in a much more "stable" example)

that's NOT a good, solid, and reliable way to collect data,period.
if you're trying to support that junk as quality statistical data collection then you're quite simply wrong. period

basing any kind of decision on junk science like that is going to cause the decision is going to be skewed. period

There are no assumptions there. period

I NEVER accused them of lacking basic high school math. I accused them of NOT using good science in collecting their data. I accused them of using piss poor data collection, sampling, and making in turn, a piss poor decision based on piss poor data. I accused them of basing decisions on junk. I accused them of making a decision based on statistical data that would be laughed out of ANY statistical forum. period

A) PAUL STATED he watched for 6 days. Take it up with him

Whish is why you DO wait to see if they DO taper off. That's called good data collection as opposed to bias

C) Tournament means there were a LOT more and different playstyles outside of the norm. The norm is what you want to collect data on. Not outliers, special instances, etc. Again, this kind of stuff is stats 101, not math 101

D) Again, I'm not assuming anything, I'm basing this DIRECTLY from Paul's statement. Although you're pointing out yet again WHY his statements on how he collected the data is bad.

E) yes they, as does every other match played, as opposed to a small sample size watched by one person over a few days...
again, not assumption, EXACTLY what Paul stated and EXACTLY why you dont' collect data like that

F) Observation was done to collect statistical data. That's EXACTLY what it is. They were observing to collect data to make a decision. You're absolutely right though, it wasn't statistical data, it was a casual observation which is exactly why it's wrong

G) That's a matter of opinion and I don't think it is. If it were an isolated incident I might be more inclined to agree.

Edited by Sandpit, 25 March 2014 - 05:30 PM.

[Marj]

LRM's are really difficult weapons to balance. If you have 3 assault missile boats + a spotter or UAV all 3 assaults can focus their fire on one target with little downside. Add in the fact that some maps have no effective LRM cover and they can be devastating. AMS won't save you from 180 LRM's. If you only have an LRM20 trebuchet they may do very little though since AMS shoots down a fixed number of missiles per volley.

I think the speed buff is good, it was needed. Any slower and LRM's would go back to being niche weapons. They need something to offset the buff so they don't become absurdly OP under certain circumstances though. I don't think a speed decrease is the way to do this. A large increase to spread on targets out of LOS would help, with the spread increasing the longer the target can't be seen by the firing mech. Tag + narc should offset this, but not completely. Lowering the arc a bit so some but not all missiles get stopped by cover would help too, but it would be difficult to get the angle just right.

[VanillaG]

Sandpit, on 25 March 2014 - 05:18 PM, said:

I DO have knowledge of how they collected telemetry because they TOLD US how they did it. Paul posted it.
6 days
a very small sample size of matches reviewed (these are NOT assumptions, these are directly stated by the person collecting the data)
Data reviewed while in an upward trend right after the release of a new buff (in other words they should have waited and reviewed data in a much more "stable" example)

I think the main assumption on your argument is that that he relied solely on his observations of matches to come this conclusion. We all know that they have per match data on weapons used because of the stats page that is part of our profile. If anything the stats that he could look at are probably the most pristine that we will ever see because all of the previous stats were archived as this change was put it in. If you want to believe that Paul based solely on his observations I am pretty sure that there is nothing that anyone can say to change your mind.

Personally I think the process went down like this:

• Change goes in and forums explode with "OMG nerf lrms/LRM are OP" threads
  
• Looks at the back end stats to see if something is wildly out of whack.
  
• Drops into random games at different ELO ranges to observe the actual gameplay experience.

I think based on those three things occurring the decision to lower the speed comes about. Balancing gameplay is still more art than science because somethings are hard to capture with just numbers.

[Alex Warden]

Charons Little Helper, on 25 March 2014 - 01:21 PM, said:

1. The main complaint of the LRM buff is indirect fire.

2. Most people who seriously used LRMs don't use indirect fire consistently. (otherwise they wouldn't bother with Artemis)

3. So - the simple solution is to make LRMs when fired while the target is in LOS fire at 175m/s, while those fired indirectly move at 120m/s(perhaps a smidge more)

4. Done

<<< likes it

[Charons Little Helper]

VanillaG, on 25 March 2014 - 02:11 PM, said:

What happens if a LRM obtains a lock via LoS , launch your LRMs, and you lose LoS; do the LRMs slow down?
What if you obtain a non-LoS lock, launch the LRMs and then you get LoS before they hit; do the LRMs speed up?

My intention was that the LRM speed be determined entirely by LOS at the time that the LRMs were fired. Hence my wording "...when fired while the target is in LOS...".

I'm sorry that I wasn't more clear.

Of note - this change would also make AMS considerably more effective against indirectly fired LRMs than directly fired LRMs.

[mike29tw]

Lyoto Machida, on 25 March 2014 - 01:45 PM, said:

We could also try getting rid of indirect unless the target is TAG'd, NARC'd or UAV'd.

I like the idea. Light mechs would have a role if TAG, NARC, or UAV is mandatory for indirect LRMs.

[Alex Warden]

yea the "no indirect fire unless a spotter tags or narcs the target" idea was something that i broought up several times in "feature suggestions" for a while... it would be way more BT/MW style than what we have now and bring more role warfare elements to scouts... just raise the tag/narc boni (even give the narcing scout the bonus for the damage the LRMS deal on that target while narced) by a good ammount and LRM players will have enough targets most of the time... then apply the OP´s idea to that and you still have more incentive to go for direct fire (which is the only way that Artemis works btw, thus more damage per salvo) while keeping indirect fire viable for surpression

it has been widely ignored...

Edited by Alex Warden, 26 March 2014 - 07:18 AM.

[Bagheera]

Rebas Kradd, on 25 March 2014 - 02:58 PM, said:

I'm a math teacher, so don't talk down to me about statistics. You are making massive assumptions here. You don't actually have even the slightest knowledge of how they gather telemetry.

Two things here.

Your profession is irrelevant because PGI' failure to adequately gather and evaluate data isn't a mathematical failure. It's a failure in the application of scientific method.

They chose to use a tiny sample set taken during a short period of skewed behaviour and declared it representative of the norm. Apply all the math you want, but if you are starting from a false premise the math doesn't matter.

Second: We actually have plenty of observational evidence of PGI's methods to generate a quite safe hypothesis that they are not up to par in gathering and examining their metrics. Why they do this is a mystery, but they continually fail to consider the data they collect outside of a very small vacuum of a single game element.

The "team" demographics are the best example. No in game LFG, no chat, a crap friends list and teaming interface, several of the largest "groups" are awol (Kaos, et al), all sanctioned tourneys are solo events, and then they say things like "84%" are playing solo.

Yeah, no shit PGI.

I could go on. How long did it take for them to "figure out" a solution to the dominance of PPCs - oh wait, they didn't. They just made it so you can't alpha large groups of them anymore.

The math isn't the problem, it's their continued inability to examine the numbers within a proper context that is the problem.

Edited by Bagheera, 26 March 2014 - 08:20 AM.

[General Taskeen]

As Alex W, Fup, and others mentioned, those are good ideas. I have posted the same myself as well.

A lot of people worry that lrm's would be useless if only TAG/NARC are the tools that allow people to fire missiles over an obstacle, but they would only be useless in direct fire if they are not balanced properly to be useful for direct fire which is exactly a current problem. If they are fast missiles in LOS and if they simply gave the ECM a "increase lock on time buff" in LOS (rather than complete lock out), then it would be fine. A reasonable concern is the no-damage min-range as well. I am in favor of completely removing any last notion of no-damage min-range in a MW game for any weapon. In MW4, for instance you can still do damage with LRMs at close range. The Devs also think "clans with no min range lurms" is stupid, well then make IS and Clans LRMS do full damage in min-range, jeez problem solved (again MW4 did this). We already have tons of weapons that do full damage at min-range doing pin-point and are way more deadly.

First, and most obvious, if these changes were take to place, you almost instantly have roles. Light Mechs suddenly have a more major reason to use NARC/TAG, for more than missile converge/accuracy bonuses, but rather simply to provide fire support and they are the quickest mechs to get into position. Where as ECM has its role of shutting down NARC... like its supposed to in that case. BUT the caveat is, they need to give NARC its alternate use with SRM's as well.

They could also even make indirect fire even more bonus with TAG/NARC with these types of changes, by making it so the other player with the LRM's does not have to target the TAG/NARC target, but rather point their LRM's in the air in the general direction of the target and the LRM's then auto-seek the target giving the target no advance warning. The only warning would be situational awareness or your AMS suddenly going off.

I also see a lot of posts saying LRM's are fine and to stay in cover... well when you look at maps like Alpine and those other maps with no cover, you are screwed by all those missiles coming from a Mech you can not even see. On the city maps or an occasional map with cave, its not as apparent, or on River Forest map which occasionally has a high mountain spire.



They take a week to "look at data" but don't understand how the game-play mechanics they made ARE the problem. Raising or lowering a few stats is fickle. And obviously many things would need to be reversed or overhauled in the event a change like this occurs at all (being the pessimistic type with this game, I don't see that happening). The reason I think they won't change it is, because they don't want to add "thinking" stuff, since the current design favors more Arcadish Mechanics, being that this is "f2p."

Edited by General Taskeen, 26 March 2014 - 08:43 AM.

[Haji1096]

Solahma, on 25 March 2014 - 02:25 PM, said:

Easiest isn't always the best though. There is already a LoS detection mechanic in the game, it would simply call on that mechanic to determine what the travel speed is. The two scenarios you mention could be handled either way, but I think it makes more sense that it would be a parametric speed that changed in real-time. So you would have to maintain LoS for it to continue traveling at that speed or else it would slow down. On the other hand, it would make Jump Jet mechs like the Catapult and Highlander different and more effective if the speed held regardless of maintaining LoS. They could jump, lock/fire, drop down and maintain the faster speed if it was set-up like that.

The only issue I have is justifying it from a fluff view... The missiles travel at different speeds... why? to better acquire a target and track better? Otherwise it doesn't make much sense to have a missile, with a pre-set engine that provides a set amount of thrust to change depending on the condition like this. I'm fine with writing that off for gameplay, but still makes little sense for reality.

I think this would be great. Jump, fire LRMS...maintain LOS for higher velocity. Loose LOS for lower velocity, but tighter turn radius...which means tighter tracking. Seems like it would be a fun mechanic.

[Sandpit]

VanillaG, on 25 March 2014 - 08:12 PM, said:

I think the main assumption on your argument is that that he relied solely on his observations of matches to come this conclusion. We all know that they have per match data on weapons used because of the stats page that is part of our profile. If anything the stats that he could look at are probably the most pristine that we will ever see because all of the previous stats were archived as this change was put it in. If you want to believe that Paul based solely on his observations I am pretty sure that there is nothing that anyone can say to change your mind.

Personally I think the process went down like this:

• Change goes in and forums explode with "OMG nerf lrms/LRM are OP" threads
  
• Looks at the back end stats to see if something is wildly out of whack.
  
• Drops into random games at different ELO ranges to observe the actual gameplay experience.

I think based on those three things occurring the decision to lower the speed comes about. Balancing gameplay is still more art than science because somethings are hard to capture with just numbers.

Even if i went with that assumption...

6 days is not good sampling
Basing balancing on the "hot" period where their use is abnormally high (that's not an assumption that IS directly what Paul stated) instead of waiting until they do taper a bit
During a tournament where ALL data is skewed a bit because of the tourny

This is NOT good data collection. Period.
This isn't the first time we've seen things like this either. This is a trend

You do not balance games like that (unless you WANT data to reflect your biased vision which then you take data in manners that will support that) because it's junk science at best and not how ANY statistician would collect that data.

That data collection would get you a C-F grade in any stats 101 class depending on how generous the instructor was feeling. So even in a best case scenario you're balancing a multi-million dollar project on "C" level work in the absolute best case scenario but more likely a "D". I'm not talking about the buff, nerf, or balancing decisions themselves.

I'm talking about how the data is collected. It's horrible. It's bad. It's unreliable.
In short, it needs to be done properly

[Bobdolemite]

Mining reliable data about a newly introduced system, within the first 6 days of it being released. And then using that information to make changes is a little short sighted.

the above example would be like someone doing statistics about the daily running habits of a specific city, on a day when the city was doing a marathon. There would be more runners than normal ( a lot more) and the data would be skewed.

In this case it would have taken more time (and in the way things move in this game it could have been literally weeks before meaningful data would be achieved)

Though something has to be done or the baseless whiners will have a field day... amirite?

I will say though that I applaud PGI in their restraint, this nerf is a minor one (IMO) and I am happy to say that although they did overreact by moving so quickly, they did not overdo it this time (which is surprising)

[Sandpit]

Bobdolemite, on 26 March 2014 - 10:14 AM, said:

Mining reliable data about a newly introduced system, within the first 6 days of it being released. And then using that information to make changes is a little short sighted.

the above example would be like someone doing statistics about the daily running habits of a specific city, on a day when the city was doing a marathon. There would be more runners than normal ( a lot more) and the data would be skewed.

In this case it would have taken more time (and in the way things move in this game it could have been literally weeks before meaningful data would be achieved)

Though something has to be done or the baseless whiners will have a field day... amirite?

^ this guy gets it