Tuesday, February 25, 2014

A measure of effectiveness.

       Several of my recent posts have generated rather heated discussions.  Both because of my defense of unconventional builds, and my habit of seeing things differently.  But there is one common thread.  When you design a build to do damage, you are wrong because high damage means sloppy mech destruction.  A lower damage count, many times, results in more kills.  This cannot be denied.

       The other side of the coin is that when you look at kills, that number may only be high because you are stealing kills from teammates, or you are running a glass cannon, and despite your impressive count, your team loses.  This also is true.  Your win/loss ratio could be used as a measurement, but to us PUGgers, due to the variability of teammates, and random disconnects, it really doesn't mean much.  So where to begin?

       While the tendency is to make a build, play test it, and bounce it off other players for opinions, that method is far from scientific.  I want something based on gathered data, that can conclusively show a build, mech, or pilot, is good, or bad.  There are more variables to cope with than can possibly be gathered together, so which ones should be the center of focus.

       In my line of work I deal with several regulatory agencies.  My department must keep track of several metrics, each of which tell only a part of the story, but together, tell everything.  Preventive maintenance goes up, correctives go down.  Things like that.  It would seem, at first, that the same dynamic could be captured here.

       I have previously drawn up spreadsheets with odd combinations of metrics expressed at ratios, because such things can be quite useful.  Damage per second / heat per second, is a good example of this.  Damage per ton is another.  I felt I could come up with some magic ratio that would indicate your effectiveness as a pilot, your skill as a mech designer, or at least the effectiveness of a build.

       After all, the kills/deaths ratio doesn't tell the story completely.  And, as it has been painfully pointed out to me, the job of some mechs is to wreak havoc and die noisily in the center of the field.  Drawing attention to yourself, while your team either picks apart the enemy, or captures objectives.

       At first glance it would seem like kills/damage might work.  Using a great game and a terrible one, to maximize the spread, I came up with (7 kills / 1208 damage {I know right, great game.  Not typical though unfortunately} x 100 for usability) yields 0.579 for a match in my "terrible" atlas build.  The same math for a bad match in my shiny new CN9A(C) yields (1 kill /200 damage, x 100) 0.500

      Clearly this ratio does not tell much of anything, since one of my best matches, and one of my worst, are separated by such a small amount.  Maybe that few hundredths (Ten thousandths, really) means a lot.  My experience likely is not indicative of everyone else's.  Also, matches can be won or lost on the capture of objectives.  Wins pay better, all else being equal.  So that guy in the Jenner didn't kill anyone, he did win you the match.  His ratio couldn't even be calculated, but without him, you all lose.

       The more I thought about it, the more it seems it can't be done.  When you think about it, if there was one ratio by which an effective assault pilot could be rated, would that ratio be valid for a light pilot?  If the effectiveness of a medium is judged by the kills/damage ratio, does that hold up for a heavy, or a light?

       In the absence of any reasonable, or even mildly irrational, metric to measure effectiveness as a pilot, or the effectiveness of a build, there is only one, long-term, way to measure your efficacy.  The win/loss ratio.  If you have been doing your part, regardless of poor-performing teammates, you will win matches.  This is because bad players will, over the long-run, appear on both teams, more or less, equally.

       But then ELO is always seeking to keep that win/loss ratio close to 1.  Whether it works or not; it does invalidate the win/loss ratio as a usable metric.  So here we are, back at the beginning, bouncing ideas off of each other, and having no real numbers.  Not for the PUG world anyway.

Friday, February 21, 2014

Submarines and you.

       I am setting out to change your mind.  I say this up front because, hopefully, you will be less resistant to honesty than coercion.  Most of the debating I do with people about this game is because I think of things in a completely different way than most.

     Most seem to regard mechs as either giant soldiers, or only in the terms described in canon.  The roles of the mechs are as defined by a small game company who thought giant stompy robots with axes and guns were kewl.  I don't disagree with that sentiment.  And who doesn't dig giant robots? But these were not military geniuses.  I am not either.  But I know enough to look for analogs of mechs in the modern world to see what military geniuses think they should be doing.

       Some have suggested that mechs are like bombers and fighters in WWII.  The assaults being the bombers, and that they needed protection from enemy fighters.  The lights and mediums being the fighters, and the heavies... well it goes off track here because this comparison is flawed.

       Mechs are analogous to ships.  Think of the role a submarine plays in naval warfare.  Scouting, hit and run, harassing (and hopefully sinking) larger ships, disappearing after a single shot only to reappear and shoot again.  Sounds like a light mech, doesn't it.

       Frigates are smaller than destroyers. They are designed to protect larger ships, and perform anti-submarine duties.  naturally they also get used for basic attack and defend scenarios.  Sounds a lot like mediums.

       Destroyers are larger and have a primary role of defending larger ships against attacks by smaller, faster ones, like frigates and submarines.  Originally called "torpedo-boat destroyers" it was shortened to just, destroyers.  Many have become the primary surface combatants of the navies of the world.  This fits nicely with Heavies.

       Cruisers are the largest of ships.  They have the most displacement and exist solely to lay down some hurt, in the most efficient manner possible.  (Battleships and Battlecruisers are no longer in main use by the world's navies, and in fact only three countries use cruisers at all)  This is obviously the Assaults.

       So if we play the mechs to the roles of ships, we should, in theory, be an efficient, powerful, battlegroup.  You will note that the role of a cruiser is not to draw fire away from the rest of the group.  Nor is it to lead the charge.  The role of the cruiser is to fire on targets, while smaller ships see to its protection.

       So when I am confronted by angry voices saying that the most angering sight in the game is a barely damaged Atlas, that the job of the assault is to be a laser-sponge, that the entire point of an assault is to die in a "loud, grotesque, military manner", (extra internets if you get that reference), I reflect on two things.  First, that the person saying this, while passionate and well-versed in the game's lore, probably hasn't looked at ships.

       Second, that they don't realize that a mech is not a soldier, it is an expensive piece of hardware.  A captain's responsibility is their very expensive ship.  If this game were as real as some of you think, having your mech destroyed would be worse than hiding or running, despite what the lore says.

Reference:
http://en.wikipedia.org/wiki/United_States_Navy_ships


Wednesday, February 19, 2014

The Atlas Experiment part duh

       So.  After much criticism, many helpful suggestions, and some soul searching I have completed another week in the refitted Atlas.  I now sport an AC20, (having given up on the Gauss) and a machine gun in the ballistic points.  One ER PPC on the side opposite the AC20, so I always have something pointed at someone, and a flamer.  I also placed three LRM 5s.  I will explain these choices.

       The AC20 is because it's a beast.  The ERPPC, because the heat is worth the extra range, and damned-sure worth not having a minimum.  The 3 LRM 5s are because three 5s weigh less than a 15.  Every little bit helps.  Also it kind of matches the tube layout.  Originally I did not include the MG or flamer.

       But then I was forcibly reminded about the nature of PUGs.  My team mates rushed headlong into the mists ahead and left the slow Atlas and his massive artillery behind.  And then I get light swarmed.  I didn't exactly have the time to type a request for help while backpedaling, so it didn't end well.

       Adding the Flamer and MG helps keep lights at bay long enough to dispatch them, or for help to arrive. They aren't exactly a waste of space in close quarters fighting either.  This brings me though to why I built the, admittedly gimmicky, 2xAC2, 2xLL, damage machine I was using.  In a PUG your teammates are rarely to be counted on.  No matter how well they play as an individual, they lack coordination.  This is where the "role specific" mech methodology fails.    

       But I digress.  After another week I am more comfortable with the "more traditional" Atlas I now run.  The other night I had a game that came down to 2 on 1, with me being the 1, and an Atlas being one of the other 2.  Looked mostly intact too.  Great.  My AC20 was nearly out, missiles long used up.  I won't bore you with the details, but I pulled a win right out of my ass.  (OK, by request details below)

     HPG map.  I just had my last teammate and the mech we were engaging go down nearly simultaneously.  I had already used my LRMs and had gotten notice that my AC20 ammo was 25%.  I peeked over a ramp that I was on to see an Atlas and another mech that, sadly I didn't recognize.  Both because I don't have them all memorized, and because I was focused on the bigger threat.  (other mech was tore up).

     I fired the AC20 into the Atlas' back, and backed up so as to not be seen.  4 count and peek.  Back was to me again.  Fire again and back down.  6 count this time (Figured they'd take a longer look around.)  Peek, back to me again, fired... Just as the Atlas turned.  OK time to back into a corner to waited for them to come to me.

       Smaller guy was first.  I fired the flamer, MG and AC, which then was out of ammo, in a vain attempt to at least make the battle even.  While I would like to dazzle you all with the story of how I didn't take a hit, that would be a lie.  I was pounded.  Torso twisting like an epileptic doing the Humpty-hump, I managed to not lose anything more than armor.  At this point I was mostly keeping my right side, with the now inert AC20, and fairly useless flamer toward them.  

       Honestly if they had headsets I would not have pulled that out.  I managed to keep both of them where I could see them most of the time.  Firing the flamer/MG combo at the smaller opponent seemed to be keeping him from hurting me too much.  I lined up on the Atlas, was thankful I opted for the ERPPC for the 100th time that match, and fired.  His left arm came off.

       Now I felt like I could win this, for the first time since seeing that dreaded skull.  Circling the Atlas, (and absentmindedly shutting down once myself), it took three more shots with that ERPPC to down him.  The smaller mech, (that I really do feel like an idiot for not knowing) was backed against the wall, I don't think he was even firing at that point.  Maybe out of ammo, or guns blown off, since he was short a limb or two.  

       While I would like to say it was a clean death, that would be a lie.  Apparently they added shake to the aiming reticle for close matches.  It took me three shamefaced shots to take that thing down.  But I did take it down.  I don't exactly think it was a statue worthy performance, but it was a win.   

       The end result was 4 kills, 5 assists, just over 700 damage.  So almost as well as I was doing after months with my other build.  I suspect with enough time I would hit the same levels as my "terrible" build.

       I am curious why there was such a visceral reaction against this Terrible Atlas build Does it look wrong? Yes.  Is it counter-intuitive? Yes.  Is the chassis, in fact, a terrible fit for the build? Yup.  Did it work? Absofreakinlutely.

       See the build was born out of frustration with my Jagers.  They kept overheating when I needed them up and running, and I was tired of compromising on that.  More heat sinks meant less ammo.  I tried the infamous AC40 build, but I felt dirty running it.  Also, it lacked the range, and honestly the fun of the AC2 builds I ran.  I wanted a mech with enough weight allowance to not ever worry about heat.  So I made that Atlas that you all hate.

       So I had made the longest range machine I could.  LRMs, as I know from using my own ECM, aren't much good against some mechs.  One person asked; with all of the ammo I carried, if I ever ran out, because if not, that meant it was too much.  In reply:  Yes.  Yes I run out of AC2 ammo frequently.  Even with that much.  That happens when you are shooting people from further away than they can return fire.

       So my point, is that if the performance is the same, but more c-bills are awarded when you do more damage, why not use it?

Sunday, February 16, 2014

The Atlas Experiment

       Based on conversations I've had on various online forums, I have been playing my Atlas all wrong.  See, given it's consummate lack of mobility, and relative height advantage, I had been using it as a sniper.  A sniper with low power guns.

       It is no secret that I favor simple reliable builds.  On paper that UAC 5 sounds great.  In the real world (Well, the game world) it fails spectacularly every time I need it.  This propensity to fail 1/5th of the time makes it undesirable to me.  My methodology is always the same.  Mathematically speaking, any weapon system you use will be taking space, and weight allowance from another.  Since mass drivers (ballistics and missiles) are the more powerful, and less heat generating, I like those.

       Unfortunately this isn't a movie and you will run them out of ammo.  So a backup energy weapon is advisable.  I like to set the maximum number of weapon types to 2 if at all possible.  This is because it allows the most damage to be dealt, while still affording flexibility.  My build was an Atlas DDC, with ECM, (the whole reason you get the DDC) AMS, standard 300 engine, 2 AC2s, 2 large lasers, (best heat management for the damage they do) over 20 Double Heat sinks, and a metric butt-ton of AC2 ammo.  I also added advanced zoom.
     
       You will notice there are no missiles.  Any other weapons added would mean carrying less ammo, and running out; or fewer heat sinks, and constantly overheating.  Overheating is not a problem I have in this build.  The dual AC2s, which fire together do huge amounts of damage/second.  My range, given the advanced zoom, is far enough that I can target the missile boats and not be fired back on.  When they do the AMS takes care of most of that problem.

       The Lasers are great for upping the damage count, or when my side is selectively targeted, which is always.  My post game stats averaged 400-600 damage and 3-5 kills with a truly random amount of assists.  Not as good as many of you, but good for a casual PUG player like myself.  I have had games with 5 kills and 1200+ damage.  But I'm not going to lie, they were rare.

       So I switched my build to a guass cannon, (for the range) and PPCs.  I did badly the first few games because I kept getting light swarmed.  This was not a problem for me when I ran AC2s, since they fire about 2x/second.  So I dropped 2 tons of ammo, (I had too much anyway) and added a machine gun and ammo.  (it was either that or a flamer)  I also swapped out the PPCs for ER PPCs.  More heat, but no minimum distance.  Now the lights didn't swarm around me and I could begin my experiment.

       I gave myself a week to get used to the new build.  After a while I really felt like I was pounding the other team.  Right until the end of game report.  I was now averaging only 200-300 damage and 1-3 kills with fewer assists.  So the punchline is: While I see where the impression would arise, that bigger guns equals more damage; that is simply not the case.

       Also, with my first build, I was able to keep entire lines of enemy mechs behind a ridge, or other cover, while my team closed the distance.  They usually had to send at least two of their own to take me out, but by that time my team was on the rest of them.  This worked much better than you'd expect.  Again, math.  Double coverage on me, means an imbalance in our favor in the main fray.  The other advantage is that I could keep pounding away while any mech I was engaged with at close range was overheating.  More than once this has made all the difference.

       So the short story is: smaller, rapid firing long range weapons do more damage consistently, than their larger bore counterparts.  Add to this the inherent advantage of not needing an additional "Light-broom"  and the case becomes clear.  The AC2 is the current king of the battlefield.

Sunday, February 2, 2014

The People of PUGs

       There are some truly excellent players out there Pugging away.  Guys (and presumably gals) who's names you remember because you hope to have them on your team.  There are some amazing shots, crazy yet-effective builds, and positively genteel pilots.

And then there are these guys...

Captain chat spam:  Thankfully this baboon seems to have a 9:00 bed time, but you'll know him when you see him.  Whether its random Nazi nonsense, or insulting teammates mothers, he makes his presence known.  Some do this well enough that it does irritate enemies into making mistakes; but usually they just concentrate fire on the mouthy guy, he becomes a bad memory, and moves on to the next match.

Mechcommandon't:  The guy who passes lobby time by repeatedly assuming, then relinquishing command.

Blaze of Glory:  Wants to be first to the front.  Puzzled at how often he dies 1 minute in.

Streak of brown:  Not playing a sniper or LRM boat: hides in back anyway.  Often seen blocking the retreat of a teammate instead of switching places or getting out of the way.

Cheese with that:  Never has made a mistake, but boy are his teammates a disappointment, and he will tell you all about it, if you bother to look at the chat.  Like how they are never in the right place to finish that enemy he almost got.  Or how they can't manage to capture five resources with the two assault mechs remaining.  If he were still in the game he totally could.

All alone again:  Somewhere between Blaze and Cheese, this troglodyte consistently rushes into traps, knots, or deathballs of enemy mechs, far ahead of his team, then whines about a lack of help or backup on the chat.

That isn't "R":  aaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaxddddddddddddddddddddddddddddddddddddssssssssssssssssssssssssssssssssssssssss

oops sorry.  

Hit T by mistake

Four figure ping:  You see him in the lobby.  You hope that it's just a momentary misread.  It will go down.  Nope.  DC.  Better luck next time.  Not like we wanted an even number.

Participle Projecting Can:  Knows more euphemisms for homosexual than he does other words put together.  Uses them all repeatedly.  Methinks the laddie doth protest too much.

Thankfully the following fools are not much seen anymore:

Team Killington:  Apparently believes himself to be a double agent.  

Notabot:  Moves and shoots randomly.  Either a keyboard spambot, or someones toddler sibling.  

Captain Mal Content:  Complains in chat about how bad this game is.  Yet still playing, oddly...

I'm sure I've missed a few of these endangered feces.  I'm sure there are new types (d)evolving as I type this.  Happy Hunting.  




Saturday, February 1, 2014

Weight Limits and You

       I am taking a break from my ill advised series on the roles of each weight class in PUG matches to present some thoughts on how weight limits would change the way we play the game.  Much has been thrown around on this already, so I will try not to focus on that.  Also I will be focusing on only the PUG matches.

       This is because I play exclusively PUGs, and pre-made teams will have an even larger advantage once weight limits go into effect.  This is because they can plan, in advance who is in what mech.  For arguments' sake I am going to assume 720 tons as a base.  Mostly this is what the player base seems to have settled on as right, (more or less) and because it is as good a number as any.

       The inclusion of a weight limit means several things.  Firstly, there will likely not be any teams with 4 Atlas on them anymore.  Second, at an average of 60 tons per mech, everyone needs to pull their weight.  There will be more lights, and more mediums.  This is because a single Atlas will make the rest of the team average 56 tons.  There needs to be that offset.  2 Atlas makes the rest of the team 52 tons.  It gets bad quick.

       So to all the assault pilots: "you need to bring the noise big time."  To the lights: "one of you needs to babysit each assault on your team."  Sorry guys, but when there are fewer big guns on the team, the ones you have need to be preserved.  The only way around this is our newest MVP; the mediums.  Between 40 and 55 tons they represent a group that is below the average tonnage.  Each medium frees up tonnage for a heavy or an assault.

       Heavies:  Your role is basically unchanged.  But remember that each ton over 60 will cost teammates tonnage.  Given the Min/Max nature of most heavy builds, and the cost/ton of the damage they bring, I predict that this class will lose popularity faster than assaults.  My reason is simple.  With a proliferation of lights, and medium mechs, slow-moving glass-cannons will not be able to pull their weight.  They will be spotted, and disassembled too quickly to do much good.

       To the mediums: "you are now the bread and butter of a team."  At your tonnage and speed, plus the ability to pack some firepower, you will finally get the respect you deserve.  Still help out that assault pilot, but remember too that you have the ability to swarm and swamp enemy mechs.

       Think about this: a team made of all 6 assaults and 6 lights will do well, but it will be very hard to coordinate.  A team of 6 mediums, and 6 heavies will be easier to coordinate, and bring the same firepower, but not much will be different.  But a team of 3 Atlas, 6 Blackjacks, and 3 Hunchies scares the hell out of me.

       Mediums are great platforms for attrition fighters.  Guns, speed, and not too much armor make them hard to hit, and bad to be hit by.  I am predicting that you will not see too many assaults without ECM.  A map full of fast, mean attrition fighters means that being a big, slow target makes you die fast.  Unless you remain hidden.

       While it will mean some adjustment, and many min/max builds will be forgotten, I'm with the lore guys in saying this will be a welcome change, even if it mixes skill levels more.  Provided there are an equal number of green recruits on each team.

       Like we all are then, I remain cautiously optimistic.  Good hunting.