Jump to content

Fix For Boating/poptarts


22 replies to this topic

Poll: How this for a fix? (16 member(s) have cast votes)

Is this a viable fix?

  1. Yes (9 votes [56.25%] - View)

    Percentage of vote: 56.25%

  2. No (4 votes [25.00%] - View)

    Percentage of vote: 25.00%

  3. Abstain (3 votes [18.75%] - View)

    Percentage of vote: 18.75%

Vote Guests cannot vote

#1 Brilig

    Member

  • PipPipPipPipPipPipPip
  • Philanthropist
  • 667 posts
  • LocationTexas

Posted 23 April 2013 - 09:19 PM

The idea here is to fix our current issues with boating, and alpha striking without eliminating build diversity, or skill. Again I don't want to nerf aiming. Skill, and by extension pinpoint accuracy for all weapons is something that I think is very important.

So right out the door I want you to know that I don't think There is anything wrong with boating. There is also nothing wrong with alpha striking. Neither boating, nor alpha striking are bad by themselves, they are not bad when they are used together either. They only become an issue when convergence allows all that damage to be put into one specific spot.

Quick break down.
Why is boating effective? because you have only have to worry about 1 weapon range, and 1 type of weapon behavior.
Why do you alpha strike? Because you want to put a ton of damage on a target all at once.

Boating is a valid way to build, and alpha striking is a valid tool. The problem is that convergence makes boating and alpha striking the most effective play style instead of an option. Convergence means that all the weapons on your mech are all aimed at the exact same point when you fire. That is why boating weapons, and alpha striking works so well. Where as with a balanced build different weapons have different ranges, and behaviors so it is much harder even with alpha strikes to get all your damage in 1 spot. So boating + alpha strikes + convergence = the most effective way to put damage on an enemy mech. Take away convergence, and boating + alpha strikes still put out a lot of damage fast, but no more than a mixed build would.

Fixing convergence means coming up with a way to spread damage around. If you nerfed convergence a boat could still alpha strike without having a major advantage over a mixed build. Neither type of build would be able to guarantee all their damage would hit a single section. Boating could still be effective, it just wouldn’t have the major point damage advantage it currently has over mixed builds. Alpha striking could still be useful, just not in every situation like it is now.

On to the fun part! Here is my favorite idea for nerfing convergence , while keeping things skill based.

Posted Image

It’s actually rather simple. In the above picture those little boxes around the crosshairs let you see the status of your weapons groups. Those boxes could keep that function, but I would also make them the crosshairs for the mechs hardpoints. Any torso hardpoints would get their own little box crosshair to the side of the center crosshair. Arm crosshairs could stay circles. For instance, an Atlas with an AC-20 in the right torso would get a box crosshair on the right.

Posted Image

For a Hunchback-4P with 6 hardpoints in the right torso you would get 6 little boxes stacked just like they are in the side torso. So long as you had 6 weapons equipped. If you had fewer you would just have crosshairs for those hardpoints that were occupied.

Posted Image

It would be easy enough to test as well. If a system is doing too much point damage in testing, then spread the boxes out until damage spread is where it needs to be.

Now for mechs with multiple weapon hardpoints in the arms you can do the same thing. Each arm crosshair could stay a circle to differentiate between torso(box) and arm(circle) hardpoints. If the right arm has 2 hardpoints then it gets 2 little circles a small distance apart to spread damage out. So an Atlas-RS with 2 MLasers in each arm could look like this.

Posted Image

If that’s too good then spread them out a little bit. Or better yet make them look like the Hunchback-SP's arms would.

Posted Image

The biggest issue with this idea is that it will make the game a little more difficult to play. I’m going to stress the LITTLE. The crosshairs do not have to be a huge distance apart. With a real training mission, and a trip or two to the testing grounds it shouldn’t be too hard a for a new player to get a feel for it. The above images are just guestimates, the actual size/spread of crosshairs would be something that would need some tweaking.

Multiple crosshairs keeps shooting skill based. It also gives the mech bay a little more utility since you really have to look at where you are mounting weapons, and how you want your groups setup.

Let me address some concerns I have already run into.

This is the big concern I hear a lot. It's the year 3050, mech engineers could build mechs that allow every weapon to converge on one point no matter where they are located. After all Current main battle tanks have the ability to keep their main gun on target, on the move, through difficult terrain. That’s all very true, but what we are forgetting here is that this is a video game. Those same mech engineers haven’t figured out how to make a half ton machine gun with an effective range over 90 meters. In fact the whole universe falls apart when you try to explain it with real life science. Convergence is a mechanic in the game that is leading to broken game play. So convergence needs to be fixed. You can just as easily say that the technology of the time doesn’t support torso mounted weapon convergence because the systems break down too easily from constant stomping around. Maybe targeting computers in this fictional universe are not good enough to handle it.
The other big one is that the suggested fix will not stop people from boating, or alpha striking with boats. Well that is absolutely true. The above fix is not supposed to eliminate boats or the utility of alpha striking. It is just to fix the current issue we have with it being the best way to play the game instead of an optional way to play.

A lot of people have suggested fixes to convergence with RNG, or expanding crosshairs. The problem with those ideas is that they take skill out of the game, or nerf brawlers without doing anything to nerf sniper builds.

Edit:Clarity

Edited by Brilig, 23 April 2013 - 10:13 PM.


#2 Capt Cole 117

    Member

  • PipPipPipPipPipPip
  • 362 posts
  • LocationSeattle Aerospace Defense Command, Terra

Posted 23 April 2013 - 09:33 PM

Makes perfect sense, although mechs with wide torsos might be slightly disadvantaged.

#3 Brilig

    Member

  • PipPipPipPipPipPipPip
  • Philanthropist
  • 667 posts
  • LocationTexas

Posted 23 April 2013 - 09:37 PM

View PostCapt Cole 117, on 23 April 2013 - 09:33 PM, said:

Makes perfect sense, although mechs with wide torsos might be slightly disadvantaged.


Each mech would need some individual crosshair management. There would probably have to be a standard distance between torso crosshairs for all mechs, instead of relying on the mechs appearance to determine the distance.

#4 Szkarlat M

    Member

  • PipPipPip
  • Storm
  • Storm
  • 76 posts
  • LocationKittery, St Ives Compact

Posted 23 April 2013 - 09:38 PM

Hey Brilig I like your idea however I have the following points to bring up: How do you account for people swapping out their stock weapons. Ie. if you replaced a stock large laser with 2 mediums would the HUD then pick up on your build and place 2 laser targets there instead or will they then converge on the one large laser reticule.

#5 Brilig

    Member

  • PipPipPipPipPipPipPip
  • Philanthropist
  • 667 posts
  • LocationTexas

Posted 23 April 2013 - 09:52 PM

View PostMike W, on 23 April 2013 - 09:38 PM, said:

Hey Brilig I like your idea however I have the following points to bring up: How do you account for people swapping out their stock weapons. Ie. if you replaced a stock large laser with 2 mediums would the HUD then pick up on your build and place 2 laser targets there instead or will they then converge on the one large laser reticule.


You know I hadn't thought to add that. The game would have to be able to tell what weapons were loaded where, and put cross hairs on the hud to match. Since there is a hardpoint limit on each mech it should be easily adaptable.

So in your example the LL crosshair would go away and 2 crosshairs for the MLs would replace it.

#6 Cifrer

    Member

  • PipPipPip
  • Bad Company
  • Bad Company
  • 74 posts

Posted 23 April 2013 - 10:00 PM

Instead of crosshairs for individual weapon types, wouldn't crosshairs for the hardpoints work just fine? I imagine that would considerably cut down the workload of the people doing the work to implement something like this.

After all, whether it's a medium laser or a large laser, it's in the same hardpoint, pointed the same direction. I wouldn't imagine there would be too much difference.

#7 Brilig

    Member

  • PipPipPipPipPipPipPip
  • Philanthropist
  • 667 posts
  • LocationTexas

Posted 23 April 2013 - 10:09 PM

View PostCifrer, on 23 April 2013 - 10:00 PM, said:

Instead of crosshairs for individual weapon types, wouldn't crosshairs for the hardpoints work just fine? I imagine that would considerably cut down the workload of the people doing the work to implement something like this.

After all, whether it's a medium laser or a large laser, it's in the same hardpoint, pointed the same direction. I wouldn't imagine there would be too much difference.


Honestly crosshairs for each hardpoint is what I meant. I need to go up there and clarify. I was just using the Lasers as examples because the mechs I chose have energy hardpoints in the arms.

Edit: Changed the wording in the main post to reflect that each weapon hardpoint that is occupied would get its own crosshair. IE: 2 hardpoints in the right torso with weapons would show up as 2 crosshairs for the right torso on the HUD.

Edited by Brilig, 23 April 2013 - 10:15 PM.


#8 Szkarlat M

    Member

  • PipPipPip
  • Storm
  • Storm
  • 76 posts
  • LocationKittery, St Ives Compact

Posted 23 April 2013 - 10:15 PM

Fair enough, but I guess hard point aiming issues should only apply to torsos as you would imagine that arms would be able to converge due to actuators etc? Therefore it would limit certain mechs. To be honest I am a fan of the expanding crosshair idea which I think will be easy to implement. I agree with you that convergence and expanding crosshairs seem to favor sniper builds which is why I have suggest expanding crosshairs based on speed of movement or as other have suggested throttle acceleration. That way you can still snipe but your only going to be 100% guaranteed to hit if you're sitting still... and while those running at you will have a larger crosshair a non moving target is easier to hold aim so they will still hit the sniping mech (with a reduced load taking in account of expanded crosshair misses). I think this would be a much simpler fix to implement. The conical spread of fire as you increase speed would mean that a shot aimed at CT may hit the arm or leg instead spreading the damage. Just make each individual weapon fire conically then you get a nice "splashed" alpha across a mech rather than concentrated on the CT. Maybe a sound alternative?

#9 Bloody Moon

    Member

  • PipPipPipPipPipPipPip
  • 978 posts

Posted 23 April 2013 - 10:25 PM

Voted no.

For 3 reasons:

1. Requires much work on PGI's part, so it would take a long time to implement.
2. The current 2 recticles we have is already hard enough on new players, no need to make it even more complicated.
3. Considering the nature of boats and jump snipers i don't think this would solve the issue at all, people would simply switch to a variant/loadout with better hardpoint positions.

#10 Brilig

    Member

  • PipPipPipPipPipPipPip
  • Philanthropist
  • 667 posts
  • LocationTexas

Posted 23 April 2013 - 10:26 PM

View PostMike W, on 23 April 2013 - 10:15 PM, said:

Fair enough, but I guess hard point aiming issues should only apply to torsos as you would imagine that arms would be able to converge due to actuators etc? Therefore it would limit certain mechs. To be honest I am a fan of the expanding crosshair idea which I think will be easy to implement. I agree with you that convergence and expanding crosshairs seem to favor sniper builds which is why I have suggest expanding crosshairs based on speed of movement or as other have suggested throttle acceleration. That way you can still snipe but your only going to be 100% guaranteed to hit if you're sitting still... and while those running at you will have a larger crosshair a non moving target is easier to hold aim so they will still hit the sniping mech (with a reduced load taking in account of expanded crosshair misses). I think this would be a much simpler fix to implement. The conical spread of fire as you increase speed would mean that a shot aimed at CT may hit the arm or leg instead spreading the damage. Just make each individual weapon fire conically then you get a nice "splashed" alpha across a mech rather than concentrated on the CT. Maybe a sound alternative?


Yep the type of actuators an arm has would determine how the arm crosshairs behaved just like they do now.

As well as expanding crosshairs work for infantry simulators I just don't like the idea of having it in MWO. My main issue is that it tends to give snipers an advantage over brawlers. Mechs are already really easy to hit with the current system. Anything heavier than a light is not very difficult to hit even when its on the move. The other problem I have with expanding crosshairs is that it does too much to mitigate a pilots skill. If I can place my crosshair at speed over the section I want to hit, then I should hit it.

As for it being easier to implement/balance I really couldn't say. Either idea would probably take a lot of time and effort.

#11 Brilig

    Member

  • PipPipPipPipPipPipPip
  • Philanthropist
  • 667 posts
  • LocationTexas

Posted 23 April 2013 - 10:37 PM

View PostBloody Moon, on 23 April 2013 - 10:25 PM, said:


1. Requires much work on PGI's part, so it would take a long time to implement.


As opposed to what?

View PostBloody Moon, on 23 April 2013 - 10:25 PM, said:

2. The current 2 recticles we have is already hard enough on new players, no need to make it even more complicated.


Yeah it really would add another layer of difficulty. There are a couple only things that could mitigate some of that. A good training mission could go a long way. Then not putting the crosshairs so far apart that it made hitting the mech difficult, but still having them far enough apart that hitting the same section with everything wasn't possible except for point blank.

View PostBloody Moon, on 23 April 2013 - 10:25 PM, said:

3. Considering the nature of boats and jump snipers i don't think this would solve the issue at all, people would simply switch to a variant/loadout with better hardpoint positions.


We already have mechs/varients that are better snipers/boaters than others. There isn't a way to change that, some mechs/varients lend themselves to a certain play style.

The idea is that with the hardpoints having separate crosshairs no mech could do the 40-60 pinpoint damage we are seeing now.

#12 Caustic Canid

    Member

  • PipPipPipPipPipPip
  • Ace Of Spades
  • 256 posts

Posted 23 April 2013 - 11:34 PM

I like this idea, but there are a few thing I would like to address.

1. The crosshairs would have to be dynamic, in that they would have to shrink/grow/move depending on the range of what they are looking at, to give the pilot an idea of where they will actually hit.

In the example you give, the weapons in the torso would be incredibly hard to hit with since, despite firing directly forward, they would still give the impression of convergence due to the event horizon. Since firing all weapons directly forwards will create a "wall of fire" about the size of that mech, the crosshairs will have to reflect this.

2. Built in/fixed convergence. During ww2, fighter planes machine guns were mounted to converge on a single point, creating a "kill zone" where pilots would try to place enemy planes before firing. Too close or too far and the rounds would spread. It is natural to assume that the mech designers of this era would take into account this advantage and build some level of fixed convergence into torso mounted weapons as well.

People could either manually adjust the convergence point of torso mounted weapons (in the mech lab, not in game) or it could simply be set to the weapons effective range. (More likely the former as people will want to customize it)

The nice thing about this is that 6xppc stalkers and the like could still do massive damage to one location, but would have to be at a specific range to do so. (Giving them less to complain about since this wouldn't make their build completely unviable)

Edited by Caustic Canid, 23 April 2013 - 11:44 PM.


#13 Bloody Moon

    Member

  • PipPipPipPipPipPipPip
  • 978 posts

Posted 23 April 2013 - 11:39 PM

View PostBrilig, on 23 April 2013 - 10:37 PM, said:

As opposed to what?


Opposed to simply slowing the convergence speed for example? Similar gameplay effect, but much easier on the devs.

View PostBrilig, on 23 April 2013 - 10:37 PM, said:

Yeah it really would add another layer of difficulty. There are a couple only things that could mitigate some of that. A good training mission could go a long way. Then not putting the crosshairs so far apart that it made hitting the mech difficult, but still having them far enough apart that hitting the same section with everything wasn't possible except for point blank.


A training mission is won't make this even remotely easier, this change would ensure that FPS players have to completely relearn aiming and everyone else will have an even harder time understanding each variant.

View PostBrilig, on 23 April 2013 - 10:37 PM, said:

We already have mechs/varients that are better snipers/boaters than others. There isn't a way to change that, some mechs/varients lend themselves to a certain play style.

The idea is that with the hardpoints having separate crosshairs no mech could do the 40-60 pinpoint damage we are seeing now.


I know what your goal is, yet if a variant has favorable hardpoint placements then it could do pinpoint damage even if your "fix" gets implemented.

For example this build would have 2 convergence points:

http://mwo.smurfy-ne...f95489e24e8f29e

But even if the weapons converge by your suggestion those points will be so close they might as well be the same point from gameplay perspective.

Edited by Bloody Moon, 24 April 2013 - 12:08 AM.


#14 Brilig

    Member

  • PipPipPipPipPipPipPip
  • Philanthropist
  • 667 posts
  • LocationTexas

Posted 23 April 2013 - 11:57 PM

View PostCaustic Canid, on 23 April 2013 - 11:34 PM, said:


1. The crosshairs would have to be dynamic

In the example you give, the weapons in the torso would be incredibly hard to hit with

2. Built in/fixed convergence.


Chopped it up just to make the post smaller.

In response to 1. Having dynamic expanding crosshairs is what this idea is supposed to avoid. Aside from having to manage multiple crosshairs, I am not sure where the difficulty in hitting an enemy mech would come from. If its under the crosshair the weapon will hit it.

In response to 2. I like the idea, but at the same time I think it would be hard to juggle if you used more than 1 type of weapon. Which would lead us back to everyone boating 1 type of weapon. There is also the issue with knowing where your weapons will hit when they are not in the sweet spot of convergence.

#15 Brilig

    Member

  • PipPipPipPipPipPipPip
  • Philanthropist
  • 667 posts
  • LocationTexas

Posted 24 April 2013 - 01:04 AM

View PostBloody Moon, on 23 April 2013 - 11:39 PM, said:

Opposed to simply slowing the convergence speed for example? Similar gameplay effect, but much easier on the devs.


That's is actually what this idea evolved out of. There are a couple of issues I have with just slowing down convergence. 1 is knowing where the weapons are going to hit between crosshair hitting the target, and the weapons are fully converged. It also doesn't do much to bring mixed builds back on par with boating. Then you still have the issue of alpha striking being the best thing to do. All it really does is add some time between shots.


View PostBloody Moon, on 23 April 2013 - 11:39 PM, said:

A training mission is won't make this even remotely easier, this change would ensure that FPS players have to completely relearn aiming and everyone else will have an even harder time understanding each variant.


There is nothing to relearn about aiming. I think your underestimating the average player by a lot.
Having weapons on the right side of the mech correspond with crosshairs on the right side of the center reticle doesn't seem like a difficult concept to grasp.

View PostBloody Moon, on 23 April 2013 - 11:39 PM, said:

I know what your goal is, yet if a variant has favorable hardpoint placements then it could do pinpoint damage even if your "fix" gets implemented.
For example this build has 2 convergence points:
http://mwo.smurfy-ne...f95489e24e8f29e
But even if the weapons converge by your suggestion those points will be so close they might as well be the same point from gameplay perspective.


Sorry took a minute to find a picture with a big enough mech under the crosshairs.

Posted Image

So with the above build we could end up looking at something like that. Not a bad catch. It's possible to swing the arm crosshairs over to the right torso crosshairs, and for a moment they would be lined up. At the right distance you might be able to get all 40 points of damage on 1 section. Past a few hundred meters you would end up hitting multiple sections or missing, but up closer its possible. Actually I think there might be an Awesome variant that could do that too.

Defiantly a potential problem, but I'm not convinced its a big enough snag to make multiple crosshairs a bad idea.

#16 Bloody Moon

    Member

  • PipPipPipPipPipPipPip
  • 978 posts

Posted 24 April 2013 - 01:23 AM

View PostBrilig, on 24 April 2013 - 01:04 AM, said:

There is nothing to relearn about aiming. I think your underestimating the average player by a lot.
Having weapons on the right side of the mech correspond with crosshairs on the right side of the center reticle doesn't seem like a difficult concept to grasp.


One can never underestimate enough in this case.

Also the PPCs are much closer to eachother vertically than they are on your pic.
Anyway i'll stop arguing, you have your point of view and i have mine.

#17 Kahoumono

    Member

  • PipPipPipPipPipPip
  • 306 posts

Posted 24 April 2013 - 01:43 AM

Thanks Brilig, you have done what I have been wanting to do for a long time. To simplify things lets just make weapons on the same body part converge regardless of type, it would really make things easier that way. Also arms reticles should converge as well, although not instantaniously. I.e. when to move your reticle for the arm weapons they start out apart as you [stop or slowed the cursor sufficiently] the arm reticles will converge after x seconds. x could be affected by your pilot tree or a targeting computer. I have no problems being sniped by 2PPCs on one body part or even 4ppcs from both arms if they have to take an extra second to line it up. With the changes as you have described you'll add another layer of strategy to the game where once you have figured out where the enemy's main armaments are you can use it to your advantage.

#18 Brilig

    Member

  • PipPipPipPipPipPipPip
  • Philanthropist
  • 667 posts
  • LocationTexas

Posted 24 April 2013 - 01:46 AM

View PostBloody Moon, on 24 April 2013 - 01:23 AM, said:

Also the PPCs are much closer to eachother vertically than they are on your pic.
Anyway i'll stop arguing, you have your point of view and i have mine.


I mentioned it up in the original post. I don't think you could have the crosshairs on the hud match up exactly with the weapon positions on the mech. They would have to be adjusted to keep some mechs viable, and keep others from being over powered. Maybe even have a standard pattern that all mechs would adhere to as closely as possible.

And honestly I welcome the arguments. This idea has been bugging me for a while and its good to get it out where it can be knocked around. I really think convergence needs to be fixed and the more people thinking about it the better.

#19 MustrumRidcully

    Member

  • PipPipPipPipPipPipPipPipPipPipPip
  • Legendary Founder
  • Legendary Founder
  • 10,644 posts

Posted 24 April 2013 - 01:50 AM

I think the hardest time would be figuring out which crosshair refers to which hardpoint/weapon.

Maybe it would be enough to just have crosshairs for each mech hit location? So yeah, that Hunchback Weapon Torso rocks, because all the guns in there still converge on one spot. But that's neatly compensated by the fact that it's a single hit location that everyone knows to focus on. (Just like it's now, except you have no advantage from it compared to someone that has his weapon hard points split across different sections).

Edited by MustrumRidcully, 24 April 2013 - 01:51 AM.


#20 Brilig

    Member

  • PipPipPipPipPipPipPip
  • Philanthropist
  • 667 posts
  • LocationTexas

Posted 24 April 2013 - 02:01 AM

View PostMustrumRidcully, on 24 April 2013 - 01:50 AM, said:

I think the hardest time would be figuring out which crosshair refers to which hardpoint/weapon.


I've been thinking on that for a bit, and I think it would take a combination of things to make it easier to distinguish which weapon goes to which crosshair. A combination of color, and crosshair shape could do the trick. red + for ballistic, blue square for energy, green triangle for missiles.

Then whichever weapon is on top in the mech bay takes the highest crosshair for that section, and on down the line. If they are stacked next to each other then maybe they could go from right to left.

There is supposed to be a weapon grouping option coming for mech bay 2.0, and if you can assign the weapons to their crosshairs there, then that should help out.





11 user(s) are reading this topic

0 members, 11 guests, 0 anonymous users