78 replies to this topic

[Ruar]

Rampage, on 27 April 2017 - 08:40 AM, said:

PGI is trying to control min/maxing where every single Mech takes the same "best" skills and every meta Mech is a carbon copy of each other.

There are two ways to do that. The first is to drastically limit the amount of Skill Points that you can use so that there is no way that you can take all the "best" skills and must make choices of which one you will take at the expense of something else that you really want.

The second way, and the way PGI has chosen, is to give you lots of skill points but raise the cost of using those points by spreading the "best" skills out and separating them with skills of lesser value that you must work through to get the maximum benefit of wanted "best" skill. There is actual benefit to the system that PGI chose because the majority of those "useless" skills do provide performance benefits when they are activated in the process of getting to the "best" skills. The only time this is not the case is when you really, really want to max out something like weapon cooldown and have to choose a node for a weapon that you do not have equipped to get to additional cooldown nodes. You are not "forced" to do that. You choose to do that to pursue the min/maxing and therefore pay additional SPs to get that extra benefit.

Min/maxing has led to power creep over the years. I understand and support PGI's efforts to control it. I do wish there were less SP awarded and less nodes to choose from. However, I also enjoy the added performance that those "useless" nodes give to my Mechs. I may not have specifically wanted increased sensor range and target info gathering when I was trying to get my Seismic Sensor or Radar Deprivation but it is nice in game to be able to see the enemy and his load-out from further away. It is also nice not to come to a almost complete stop when climbing a grade on Canyon when I have a couple of "useless" Hill Climb nodes that got in the way of me getting Cool Run.

I have currently leveled up and used 7 Mechs in the PTS. Each and every one of them is slightly different. Even in the case of using two Mechs that are the same chassis and variant, the skill tree nodes that I selected to optimize for the performance that I wanted are different in more than just the Weapon Tree. In the old system, they would be carbon copies of each other with the exception of of the Weapon Modules that were equipped. I find that this new system adds some diversity to the builds.

Is it perfect? No, but I can see what they are doing here and I find this to be a workable if not refined way of achieving their goals of limiting min/maxing and providing an improved level of optimization for the Mechs.

In effect PGI has limited the number of skills, it just doesn't seem like it because they went with 91. Then they took those limited skills and plugged them into a web system. They used both methods and this is why people are upset because it throttles customization twice instead of just once.

Given we have limited skills then we don't really need a web system to ensure we have to make choices on how to spend those skills. They just need a linear system with the costs balanced so that we have to make choices. See my post a few above yours about radar dep and how it's the perfect example of this cost balance.

With a web system it will take a lot of effort to keep balanced. A linear system is easy to balance as we go along.

Honestly, there is no logical reason to use the web system because the linear system can accomplish all of the same goals, be organized in a much easier to understand method, is less daunting to new players, is easier to click across, easier to balance, easier to add/subtract new nodes in the future, and is more flexible overall to changes in meta.

The web system offers the players nothing other than change for change sake, the linear tree design offers everything the devs are trying to accomplish in a system the players like.

I just can't figure out why they insist on using the web other than someone high up at PGI thought of it first and now refuse to acknowledge it's inferior.

[Breadmachine]

Kiran Yagami, on 26 April 2017 - 06:32 PM, said:

What do you mean they are "too valuable"? We already have those skills on our mechs, and they're the cheapest and fastest ones to get. Too valuable my ***. They work just fine in their current iteration. There's no need for the ridiculously complicated and stupid implementation they're trying to sell us on.

The intent is to fundamentally change the way the game is played such that there aren't these maxed out "meta builds" to the degree where you could run a specifc chassis in many different skill configurations that each reflect a different purpose in a play style.

This affords a light the ability to be a brawler, or a quick recon, or a brawling recon, or a quick brawler, and affects the basic gear it can carry, for example.

It looks like the lesson from Test #1 was learned, and this Test #2 puts the skills we will want behind skills that we will use, rather than the previous tree putting literally worthless nodes in the way of valuable assets.

The real trick, now, is to get out there and test combat.

[Baba_Yaga]

One simple solution, dump this pile of garbage!

[Jack Shayu Walker]

Ruar, on 27 April 2017 - 08:42 AM, said:

I tried hill climb on my ENF because I'm flanking with it all the time. In the end I dropped it to try out seismic which has some value as well. I'm not really sure which is better for me overall because they both seemed to be equal for my style. I put radar dep on every mech though because it has value to me all the time.

How do they make hill climb or seismic more valuable to me? They could increase the return, but at some point that has to stop. If seismic showed me the mech movement at 800m I would take it but that's a bit out of control. If hill climb let me go up an angle I couldn't previously go up then I would take it, but just keeping up my speed isn't enough really.

However, if those were both one point skills then I would consider taking them. The problem though is if they are one point skills they can't really be used as gates because everyone would have them. Which leads back to the problems with the web system instead of a linear route. In linear we can have a bunch of one point skills sitting on the side to round out a build and still have a high cost on thing like radar dep. In a web there is a limit to how many gates you can create before everyone ends up looking exactly the same.

I think its best to consider that hill climb is primarily useful for assaults and maybe slow heavies. I wouldn't be opposed to causing the climb to affect the actual angle, its just not what it does currently. The point is though, it's a legitimate helpful investment for mechs that have a hard time climbing certain slopes. You just need to charge the pilot the maximum amount of skill points they'll be willing to pay to equip it.

What is that maximum? I don't really know, but a few tests once things go live wouldn't be hard.

[Ruar]

Jack Shayu Walker, on 27 April 2017 - 08:52 AM, said:

I think its best to consider that hill climb is primarily useful for assaults and maybe slow heavies. I wouldn't be opposed to causing the climb to affect the actual angle, its just not what it does currently. The point is though, it's a legitimate helpful investment for mechs that have a hard time climbing certain slopes. You just need to charge the pilot the maximum amount of skill points they'll be willing to pay to equip it.

What is that maximum? I don't really know, but a few tests once things go live wouldn't be hard.

Imagine how hard it's going to be to adjust the cost of hill climb in the web though. Move it up higher so it's cheaper and now you have to move something else down making it cost more. There is no simple way to make adjustments on these tertiary skills that add flair but aren't core the majority of builds.

Linear tree solves that problem though.

[Jack Shayu Walker]

Ruar, on 27 April 2017 - 09:00 AM, said:

Imagine how hard it's going to be to adjust the cost of hill climb in the web though. Move it up higher so it's cheaper and now you have to move something else down making it cost more. There is no simple way to make adjustments on these tertiary skills that add flair but aren't core the majority of builds.

Linear tree solves that problem though.

Not advocating for web. Want tree. Re-read postz plz.

[Ruar]

Jack Shayu Walker, on 27 April 2017 - 09:03 AM, said:

Not advocating for web. Want tree. Re-read postz plz.

Understood, never said you did. Just continuing to stump for my favorite candidate despite the entrenched party doing their best to ignore his value.

[Mike Barnes]

So basically what I'm getting from the half that are mad is that you just want the exact same thing we have now, with little nodes. Instead of making choices about your 'mech and its abilities in 'mech operations overall, you just want to play the same build you have now, make sure that no change gets let in, and the game can just be stifled, stuck in the same state is has been in for five freaking years. Quit your crying put on your big boy/big girl pants and take your hill climb.

[Ruar]

Mike Barnes, on 27 April 2017 - 09:08 AM, said:

So basically what I'm getting from the half that are mad is that you just want the exact same thing we have now, with little nodes. Instead of making choices about your 'mech and its abilities in 'mech operations overall, you just want to play the same build you have now, make sure that no change gets let in, and the game can just be stifled, stuck in the same state is has been in for five freaking years. Quit your crying put on your big boy/big girl pants and take your hill climb.

You obviously have not read the replies. Nice try trolling though. 4/10

[Jack Shayu Walker]

Mike Barnes, on 27 April 2017 - 09:08 AM, said:

So basically what I'm getting from the half that are mad is that you just want the exact same thing we have now, with little nodes. Instead of making choices about your 'mech and its abilities in 'mech operations overall, you just want to play the same build you have now, make sure that no change gets let in, and the game can just be stifled, stuck in the same state is has been in for five freaking years. Quit your crying put on your big boy/big girl pants and take your hill climb.

Okay first, you obviously haven't been reading a damn thing. Nobody wants the old system. We want a linear skill tree that will promote actual choice and build diversity. The web design severely -and quite tragically- bottlenecks customization. If you don't understand why, read around, there are plenty of posts about why, including in this thread.

Second I want my hill climb! I just want it to be an actual specialization i can focus in on! Not a 15% 3 node filler on the way to cool run!

Third I like my little boy pants thank you very much

Edited by Jack Shayu Walker, 27 April 2017 - 09:15 AM.

[Chound]

Kiran Yagami, on 26 April 2017 - 06:32 PM, said:

What do you mean they are "too valuable"? We already have those skills on our mechs, and they're the cheapest and fastest ones to get. Too valuable my ***. They work just fine in their current iteration. There's no need for the ridiculously complicated and stupid implementation they're trying to sell us on.

In the current live system you pay gxp to unlock a skill then pay more for a module so you can use the skill. advanced zoop and Radar derp are both 15Kgxp and 3M C bills

[Tesunie]

Jack Shayu Walker, on 27 April 2017 - 08:14 AM, said:

I think ultimately if you clearly label the varying node costs with numbers in big font, a linear tree with varied costs would work just fine. I think it's actually more rewarding to not get all of the free extra stuff on the side, and just to get what you were after, Yeonne.

I have found, that doesn't always seem to work.

Where I work (generalize merchandise store) I can write a sign in big letters, and no one will see it. I could make it bright pink, and no one will see it. I could tattoo it to my forehead and no one would read it.

Every day, I get asked when the store's hours are. The hours are posted right at eye level on the front doors. In order to walk into the store, you have to look at it. And yet, every day I get asked the same question.

We have a note by our credit machine that says "please wait to swipe". Everyone still swipes their card, which messes up the machine. We have colored the note pink, and still no one reads it.

We have our return policy posted on the counter in front of the register. I still end up with people who ask what it is, and/or try to return something that is against the store policy (and normally, we even give verbal warnings if we can't take something back, such as crossbows which tend to be mishandled and break, because people don't read the instructions).

Etc.

Not saying I'd be opposed to your concept, but there are just so many people who just wouldn't seem to know... Because they don't read...

However, I don't overly understand the issue here. You want instead of having to go through 3 skills to get the skill you want (which still will provide you some bonuses), you want your skill to cost you 3 skill points. I personally would rather just take the other skills if it's going to cost me the same anyway, personal opinion here. Though I go grasp what you are saying, make low skills cost less and higher skills cost more, instead of spread out and locked behind other skills.

Scynonymz, on 27 April 2017 - 08:26 AM, said:

We need to be able to save multiple loadouts for a single variation of our mechs and switch seemlessly between them. To include mech builds and skill point perks associated with the builds.

Thing is now, that is where you are suppose to own different mechs, so you can place different loadouts with different skill sets on them. As each mech will now use it's own skill sets, instead of sharing with all other mechs of the same variant.

I personally (which I may be unique with this) have always bought a new chassis if I had more than one build for a mech. (Typically after testing on a chassis, where I would buy another copy and place the old build on that one. But this is in the current system, with shared skills.)

I don't even like sharing engines or modules... I like my mechs to be "grab and go" ready.

Ruar, on 27 April 2017 - 08:42 AM, said:

In linear we can have a bunch of one point skills sitting on the side to round out a build and still have a high cost on thing like radar dep. In a web there is a limit to how many gates you can create before everyone ends up looking exactly the same.

Here is the issue, each system has it's pros and cons. It's a matter of which system works best. Of that, no idea. Each has the potential to work as well as the other as far as I can see it, but I do think for game balance and to try and fight the Min/Max that we know will be attempted at every corner, I think the web has the best chance at this time. Do note, this is my opinion, not a bashing of your idea.

[Y E O N N E]

Ruar, on 26 April 2017 - 09:59 PM, said:

I'd rather spend one point on cool down, one on laser duration, and then spend the third on whatever I choose instead of being forced to take a skill the devs are forcing me, and everyone else, to take. They've simply built a new cookie cutter instead of actually allowing us to customize the way we want.

I have no problems with higher cost skills, but there must be no filler nodes. Let us pick only what we want. Its easy enough to do if the just admit they've messed up on the tree.

What it sounds like you are really lamenting is that the payout isn't worth the skill point investment.

What is going to happen if PGI changes to a linear tree is they will count the number of points you need to spend to get to the node you want, then either add that to the price of that one node or subtract the difference from your total pool. In the end, you will be spending the same amount you are now to get exactly the same benefit and you will have nothing extra to spend. To boot, you will only spend down each linear progression until the point where you feel the return is no longer worth the expense, meaning you still end up with a cookie-cutter set of nodes that everybody buys because that's the most potent arrangement because there is no real trade-off here.

None of the problems you have with the current tree are ultimately caused by its arrangement. They are caused by the value on the nodes themselves not being worth the number of skill points you have to spend to get there.

Honestly, linear is not the way to go. It doesn't even mean anything. Grouping skills by intended role is the most desirable way to go if we aren't going to implement a Perk-Drawback type of system instead.

[Ruar]

Yeonne Greene, on 27 April 2017 - 07:31 PM, said:

What it sounds like you are really lamenting is that the payout isn't worth the skill point investment.

What is going to happen if PGI changes to a linear tree is they will count the number of points you need to spend to get to the node you want, then either add that to the price of that one node or subtract the difference from your total pool. In the end, you will be spending the same amount you are now to get exactly the same benefit and you will have nothing extra to spend. To boot, you will only spend down each linear progression until the point where you feel the return is no longer worth the expense, meaning you still end up with a cookie-cutter set of nodes that everybody buys because that's the most potent arrangement because there is no real trade-off here.

None of the problems you have with the current tree are ultimately caused by its arrangement. They are caused by the value on the nodes themselves not being worth the number of skill points you have to spend to get there.

Honestly, linear is not the way to go. It doesn't even mean anything. Grouping skills by intended role is the most desirable way to go if we aren't going to implement a Perk-Drawback type of system instead.

I completely disagree.

I know they can't simply make it so the most used skills equal 91 nodes exactly. The reason I know that is because it would be very boring and players would have no interest. They have to leave enough room for people to pick and choose. There will always be something extra to spend after selecting the highest value items.

The key is making sure there are enough high value items that you have to pick and choose from among them. That's something that is very difficult to determine right now because the tangle mess doesn't show which items are the most important and which are just being taken because they are close by.

As for the part about only spending down to the point I feel is worth the return... absolutely. Which isn't really any different than the tangled mess except the linear path is clear on what is being chosen and why instead of being forced to take unwanted/unneeded tax items.

The tangled mess will lead more to cookie cutter builds than a linear system. It's already happening if you read through the threads on this forum. It's destined to happen because the tangled web makes people feel like they are incredibly limited in their choice and it's better to take an optimum route than to actually take the path they want to take. Because all of the extra nodes that would be available in a linear tree are being sucked up with gates on the web.

I can agree with you that grouping skills by role is more effective than grouping them by category. Imagine how horrible a tangle role web is going to look compared to a series of linear branches that can be filled in based on role.


There is literally no argument you can make that the linear design can't do better than the web design. None. You can choose to deny the superiority of the linear design in a broad scale argument, but if you compare point by point the linear tree shows it's worth. Feel free to point to any of the sections, roles, or branches and I can show you how the linear concept is better than the web for the intended goals. All I ask for is specific examples instead of broad concepts because broad concepts are just theorycraft but specifics are easy to compare.

[Tesunie]

I wish to say first that each side are making excellent points. But I do wish to just remind people to keep it civil. I understand passion for a subject (I too have my passions), but please don't let it lead to insulting and name calling. (And no, the person I have quoted under this is not being picked out, I just had something specific to say to their post.)

Ruar, on 27 April 2017 - 09:54 PM, said:

There is literally no argument you can make that the linear design can't do better than the web design.

I actually would like to disagree with you here. I'm not going to call one system out as better over the other though (surprise!), as actually each system has it's merits, as well as their downfalls.

For a web, better skills can be locked behind less valuable, but still useful, skills. If carefully laid out and plotted, a web design can very easily prevent Min/Maxing of skills and improve diversity. However, if the web is arranged incorrectly, than certain skills will be easy to grab and will always be taken, making diversity less likely. (For example, if all the laser quirks was easy and cheap to unlock, now laser weapons all have an unwanted boost, making them and their skills the new min/maxed weapon for the game. Otherwise known as "meta" or "flavor of the month".)

However, a linear tree also has the same kinds of strengths and weaknesses. It's strengths are a clear path to follow. You want this skill, than by all means unlock it. However, it's downfall is that it can make it real easy to min/max, making some highly valued skills very easy to possibly get and/or max out. Sure, it's easier to find what you want and skill it out, but it's also easier to end up with a "super strong meta", and even more "cookie cutter" builds.


Each has their flaws. Neither one is better over the other, and neither is perfect and infallible.


Sorry if this looks "passive" and wishy washy of a post. I understand each side of the debate. Each side has very excellent points being made. Overall, I may support the "web" style, but a linear style also has merits as well, so I don't want to just dismiss those either.


Edit: The best system in my opinion would be a web style, with similar role skills grouped together. This would make those "lock" skills still very useful for the intended role, and would help to enforce roles over "web crawling".

Edited by Tesunie, 27 April 2017 - 10:17 PM.

[Ruar]

Tesunie, on 27 April 2017 - 10:14 PM, said:

snip

You are wrong though. You keep talk in general and avoiding the specifics I so clearly said I can prove.

Goal- let players customize their mechs

Web failure- forces players to accept dev predetermined skills instead of actually customizing their mech.

Proof- You've seen the linear sensor tree and saw how it was superior for customization over the web. Forcing every player who wants to get radar dep to also take target retention and sensor range is not customization and is the epitome of cookie cutter.

Goal- Prevent min/maxing

Web failure- Locking skills behind gates effectively reduces the number of skills players feel they have to spend thus making them even more likely to select only the paths that provide maximum return even if that return is not what the player truly wants

Linear success- Creating sufficient number of nodes (since we can't simply increase the cost of a single node) so that the higher value skills cost more to complete. Yes there will be min/max builds, that's gamer nature, but the key difference is a linear path will also allow more customization options that the web simply can't allow.

Goal- avoid super strong meta

Web failure- The web design is so complex and complicated that making small changes to node layout or value will have repercussions throughout the entire web requiring significant testing every time a change is made to account for shifting meta.

Linear success- As the meta changes so too can the number of nodes in a skill and the value of those nodes separate and apart from any other node in the tree. Additionally adding new node series is very easy compared to adding anything to a web design.


Like I said. No argument where the linear tree isn't better.

[Tesunie]

Ruar, on 27 April 2017 - 10:24 PM, said:

Like I said. No argument where the linear tree isn't better.

No argument in your opinion.

Your method makes it easier to select only the skills you want, true. However, letting players "choose only the skills they want" I find normally isn't always a good thing. Sometimes, you (as a designer) need to block specific things off for a while so people have to "earn" it.

In this case, if you wanted "maximum" Radar Deprivation, you could go ahead and unlock all the skills you just mentioned. But, what if I was content with only taking half the Radar Deprivation skills, so I could place more skills in something else? With your proposal, depending upon how it's done of course, it may be too easy to max out Radar Deprivation (for example). Then again, this isn't saying that the web is perfect either, depending upon the skills being used along the path.

As I said before, either can be possible and could work very well. The web may be better if the skills are spread out by role, and those skills that fit in more than one role could be split between the two. This would force people to "dual role" if they want the max skills, or be content with half of the maxed skill.


Right now, I can't give you anything specific for a build. I need to get on the test server first and then provide a build as well as a skill set I'd be interested in for said build. I'll probably choose my Huntsmen Prime, with a preference for Adv. Target Decay and Cool Run styled skills, as well as a few LRM aid skills. But I need to see the skills before I can say what ones I'm going to be looking into specifically. I'm kinda hoping to get on the test server tomorrow...

[Ruar]

Tesunie, on 27 April 2017 - 10:35 PM, said:

No argument in your opinion.

Your method makes it easier to select only the skills you want, true. However, letting players "choose only the skills they want" I find normally isn't always a good thing. Sometimes, you (as a designer) need to block specific things off for a while so people have to "earn" it.

In this case, if you wanted "maximum" Radar Deprivation, you could go ahead and unlock all the skills you just mentioned. But, what if I was content with only taking half the Radar Deprivation skills, so I could place more skills in something else? With your proposal, depending upon how it's done of course, it may be too easy to max out Radar Deprivation (for example). Then again, this isn't saying that the web is perfect either, depending upon the skills being used along the path.

As I said before, either can be possible and could work very well. The web may be better if the skills are spread out by role, and those skills that fit in more than one role could be split between the two. This would force people to "dual role" if they want the max skills, or be content with half of the maxed skill.


Right now, I can't give you anything specific for a build. I need to get on the test server first and then provide a build as well as a skill set I'd be interested in for said build. I'll probably choose my Huntsmen Prime, with a preference for Adv. Target Decay and Cool Run styled skills, as well as a few LRM aid skills. But I need to see the skills before I can say what ones I'm going to be looking into specifically. I'm kinda hoping to get on the test server tomorrow...

That's fair. And to the point about radar dep, I think I've answered it previously in it's own thread and in this one. I completely agree a linear tree could have the cost for something like radar dep too low. The advantage of the linear tree though is that such a mistake is very easy to correct by adding more nodes or reducing some nodes depending on what needs to be done. The same can not be said of the web design.

I'll stop here though because I feel I've made my point. When you do have an example or two of where you think the web is better please let me know and I'll see if the linear tree can hold up like I think it will.

When you do get on test make sure you try some hot builds to see how slow heat dissipation is and then feel the frustration of trying to work the firepower tree only to realize the only thing it's really useful for is the 15% laser duration. I spent a good 30 minutes today building and rebuilding a Roughneck that ended up costing me about 125 total GSP. It's going to seriously suck to be on live trying to tweak a mech and constantly having to buy new SP to figure out which weapon combinations are going to work and which don't.

[Tesunie]

Ruar, on 27 April 2017 - 10:40 PM, said:

When you do get on test make sure you try some hot builds to see how slow heat dissipation is and then feel the frustration of trying to work the firepower tree only to realize the only thing it's really useful for is the 15% laser duration.

First, I have to test the build without skills a few times. This is to establish a base line, as well as to test how an unskilled mech operates next to a fully skilled mech (and lets face it, a lot of people are just max skilling a mech, and going right on it).

Once I have a feel for that, then I'll purchase mastery level of skills and see what I come up with. Last time, I selected my skills and ended up leaving them the first selection...

[Y E O N N E]

Ruar, on 27 April 2017 - 10:24 PM, said:

You are wrong though. You keep talk in general and avoiding the specifics I so clearly said I can prove.

I am not wrong, I'm just three steps ahead of you. Your concept is based entirely around being able to alter the value-per-node of each ability such that you can get the performance you want for what you think is a fair price in nodes. First, PGI has already told you what they want the value to be (i.e. 100% Radar Deprivation being worth 17 out of 91 nodes). Second, being able to alter that value does not inherently require changing the architecture, you can do that simply by increasing the values on the nodes as they stand.

As for increased customization, no. You've fallen for the illusion of choice. There are certain capabilities that are essential to 'Mech performance (i.e. Cool Run, Heat Containment, Anchor Turn) and they will always get taken first and always to the the maximum value possible before continued investment begins harming performance. You'll continue grabbing nodes down the priority list and the only time those mediocre capabilities will be chosen is when there's nothing more useful to spec, which depends entirely on how generous you are with your value-per-node. Ultimately, most competent 'Mechs will end up looking very similar except for the last 10-20 nodes which will be dependent upon the weapons payload (I do think firepower was better off split into separate weapon families). And do note, there is an upper limit to how powerful a capability can be because the tree, being universal, has to make sure it doesn't become abusive when applied to certain 'Mechs.

Now, whether or not we pick up some extras we would not otherwise go for has zero bearing on whether or not you get cookie-cutter selection. You will get cookie-cutter selection regardless because the nodes in the web and in the linear version just creep power upwards for each one you acquire rather than deviate power toward a particular niche, because killing 'Mechs is the single overriding objective of the game, and because some 'Mechs are inherently flat-out better than others but still have the same tree.

Quote

Like I said. No argument where the linear tree isn't better.

There are plenty of arguments when your "proof" has all of the same high-level deficiencies as the problem to be solved and when your defined successes are such only because you've said so and not because they actually achieve it. I've seen many skill trees in many games and the only one that is remotely appropriate for this game is the one from Heavy Gear 2, not what PGI has constructed and not the piddly reshuffle you are pitching. Those both belong in PvE games, not PvP.