Jump to content

- - - - -

Skill Tree Q & A With Paul Inouye And Chris Lowrey

No replies to this topic

#1 Alexander Garden


  • Developer
  • Developer
  • 1,510 posts

Posted 12 May 2017 - 05:38 PM

Greetings MechWarriors,

As we approach the release of the Skill Tree in the May 16th patch, we wanted to take a moment to run through and acknowledge a few of the more common questions and issues regarding the Skill Tree. We’ve made an effort to compile a solid, straight-to-the-point list of questions covering topics such as the initial creation of the Skill Tree, its structure, its impact on balance, our approach to balance as a whole in coming patches, and the economy of the Skill Tree.

For this post we’re going to be using a question and answer format, posing questions to Lead Designer Paul Inouye and Design Consultant Chris Lowrey.

While these questions aren’t pulled from any singular community member or discussion, I’ve taken the liberty of summarizing questions from throughout the community into a core list and in my own words.

So let’s get started.

Q: Can you summarize some the core design goals of the new Skill Tree, and what role the new Tree will play in MechWarrior Online?

Paul: The Skill Tree is being implemented for numerous reasons, all of which bring MWO forward in terms of 'Mech customization, personalization of how your 'Mech performs, and giving us a few more avenues for increasing depth of play. MechWarrior Online and all other MechWarrior titles have always been about customizing a 'Mech and trying the build that you have created.
So the Skill Tree allows us to bring customization into a new level of depth, and really drives home the freedom to make personal decisions as to how your 'Mech will behave when you are at the helm, with the Loadout you have equipped, augmented by Skills you have chosen.

Q: With 91 Active Node slots, and 237 total Nodes across the whole Tree per-Mech in the final release state, the Skill system is considerably more complex than the previous system of Basic, Elite, and Mastery tiers. The Node count and purchase flow also requires a considerable number of mouse clicks, particularly if someone is fully skilling out a number of ‘Mechs in one session. Why is the new Skill Tree structured this way, and was a system with fewer Nodes ever evaluated?

Paul: The number of clicks associated with skilling 'Mechs is something we have and will continue to be investigating. Suggestions have come from the community and we have some internal ideas on how to do this. We want to make sure we reduce the strain that may be caused by skilling large numbers of 'Mechs, so we will be looking into various options for duplicating Skill layouts or quick-pathing Node selections.

As for the size of the Skill Tree and its Node count, those characteristics of the new Tree derive from a problem that we encountered when we were looking at the existing Skill system with the tiered levels, the values associated with them, and the overall grind to acquire those Skills.These shortcomings of the previous system were further amplified by the three variant system, which was a significant obstacle for players when trying to level their ‘Mechs.

We need players to be able to reach the skill and specialization aspect of the game much faster than what we have right now. This is why there are so many nodes in the new system, offering smaller, incremental upgrades. With this system players don't have to play a large number of matches in order to even begin skilling and specializing their 'Mechs; they can start doing this after their first matches in an owned ‘Mech, which would not be the case if nodes had higher upgrade values and associated costs, and especially if they had to simultaneously level two other variants of that chassis.

The 91 node value comes from us wanting players to have the ability to fill out a large portion of the Firepower tree and still have enough nodes remaining to get more nodes in the other branches like Operations, Jump Jets, Mobility, etc. This is where the give and take aspect of the system comes into play, as providing too many points would introduce a level of power creep that we want to avoid, and would lead toward more homogenized builds rather than specialized builds.

Q: In the Public Test builds players saw, there were a few cases where they would be required to purchase Nodes with no relation to their current Hardpoints or Loadout if they wanted to maximize the useful Nodes. In the case of the Firepower tree for example, a player running a 'Mech with a pure Energy build would be required to purchase Ballistic Nodes in order to maximize their access to Heat Gen or Cooldown Skills. Has this been addressed at all?

Chris: We have directly addressed this with the release version of the Skill tree. The Firepower tree has been completely reworked, removing the majority of hard gating situations where access to universal weapon nodes were gated by hardpoint or weapon specific nodes. The Mobility tree has also been re-tuned, with the full removal of Arm-based skills.

Posted Image

In the final Firepower tree seen above, players will be able to unlock the majority of universal Firepower Nodes without needing to commit to any weapon-specific Nodes. A sole point of decision remains in the lower-middle portion of the Tree, as seen above.

With that said, the Skill Tree is still balanced around the concept of progression toward acquiring high value nodes, and the amount of nodes needed to max out high value Skills. So while hard gating of Hardpoint-specific nodes has been addressed, this does not mean that all optimal paths to a more desired node will be free of certain specialized nodes, nor does this mean we have removed various support skills on the path to high value skills. In other words, Hill Climb or Improved Gyros are still a required investment if you want to reach further into the Operations tree and acquire more desirable Skills such as Heat Containment and Cool Run.

Q: So with the node pathing in mind, can you expand on the decision to use a nonlinear Skill Tree? Was a linear system ever evaluated?

Chris: Linear systems were evaluated, but didn’t accomplish the intended design goals that we were building towards. Ultimately, we are not aiming to allow unrestricted a-la-carte access to the skills as we know that this will only result in min-maxing of the best skills from every individual tree, rather than promote specialization into general roles.

The nonlinear system allows us to set up nodes in a way in which we can design around high initial up-front costs for maxing out high value skills, but subsequent investments in the same specialization become much more cost efficient as you invest more into the tree.

Just as a hypothetical example, If a linear system was set up and gave a cost of 25 points to max out Cooldown or Heat Gen individually, then the cost of taking both nodes would be doubled if you wished to spec into both of them, which often leads to people only taking one or the other. With a non-linear setup, we can set it up so acquiring one of those skills carries a high 30 point cost, but you also pick up secondary nodes along the path, which means that if you wished to target a second skill within the same tree, maximizing that second skill sees a heavily reduced cost.

Q: In its final release state, players will need Skill Points to unlock Skills. If the player doesn’t have any legacy Skill Points to use from their Refund Ledger, each Skill Point will cost 45,000 C-Bills and 800 XP. If a player wants to later re-equip a Skill they already own for that ‘Mech, doing so will require 400 XP and no C-Bills.
How did you arrive at these values?

Paul: After the first round of PTS feedback we wanted to make sure that we're not hitting players too hard in terms of cost. Maxing out a 'Mech is not meant to be a cheap process and we want players playing the game with the 'Mechs they're building.

We took a data pull from all Tier 4/Tier 5 players without Premium Time, but who are active players, and looked at their average C-Bill and XP earnings per-match. The player data showed us that the average C-Bill earning per-match for a player in that bracket was 119,315 C-Bills. The XP average was 818. With the feedback from the initial costs of the Skill Tree and an overall re-evaluation of total costs, we felt that around 30-35% of the average C-Bill earning per-match was an appropriate value for contributing to a single Skill Node. When it came to XP similar principles applied, and we felt the single average match earning in XP was an appropriate value for contributing to a single Skill Node.

When it comes to respecing nodes, we wanted to work with the community feedback that we got but at the same time keep a maintenance cost that kept players playing and using their builds on the battlefield. This is where we came to 400 XP for re-applying an owned Node for your ‘Mech.

Q: What spurred the decision to forego the original plan for a full C-Bill refund of Owned Modules? Why the deadline of December 3rd for owned Modules being refunded as C-Bills?

Alex: Our original approach for the refund and transition of player accounts and inventories into the new Skill Tree failed to properly account for the the full range of player accounts, inventories, and skill statuses. You can see a deeper summary of those missteps in the original approach in our post here.
With a subsequent re-evaluation of the refund and transfer system as a whole after the initial announcements and PTS we introduced the full 1:1 transfer of 'Mech Skill statuses. Those changes were outlined in the Refund Ledger section of the most recent Skill Tree PTS post.
With the new mandate of ensuring the retention of player progress, the Module system was additionally re-evaluated and improved to positively account for a greater overall percentage of the playerbase than what informed our original evaluations. The Module > C-Bill refund system as originally planned failed to account both for the role of Modules as Skill system items and for the essential role of XP within the new Skill System, leaving a greater portion of players with enough C-Bills to put toward Skills, but not enough XP to actually acquire those Skills. This represented a loss of player skill progress we needed to address. With the use of General Skill Points (global points which can be consumed to acquire Skills for any 'Mech Skill Tree, without requiring the use of any C-Bills or XP) we are ensuring a majority portion of the playerbase will be fully capable of skilling their 'Mechs with no loss of previous skill status or associated progress.

However, Owned Modules purchased as of December 3rd 2016 are still being fully refunded purely as C-Bills. December 3rd marks the date of the announcement of the new Skill Tree at Mech_Con and some of our preliminary messaging regarding owned Modules. We know players may have based their Module purchasing decisions thereafter on those preliminary announcements, and it is important that we address that.

In all other areas, the refund and transfer system outlined in the most recent Skill Tree PTS post remains the same. We strongly feel that the system we have arrived at is the fairest, most comprehensive solution for the greatest portion of the active MechWarrior Online playerbase.

Q: Taken as a whole, acknowledging known outliers and some of the legitimate Clan versus IS imbalances, the current state of balance is arguably in one of its better states, relative to what we’ve seen in the past. How does the Skill Tree play into the balance question? Was it designed to promote better game balance?

Chris: Yes and no. Some core system level imbalances are directly being targeted for improvements with the release of the Skill Tree. The Engine desync brings greater mobility balance between ‘Mechs with smaller to larger engines. The Skill Tree itself offers more opportunities for people to customize ‘Mechs in different ways, which could see some ‘Mechs more apt to specing into certain trees compared to others.

To us this is an improvement over the relatively rigid Pilot Tree and Module system which tended to leave less mobile ‘Mechs trapped due to the focus on Mobility Skills in the prior system. Multi-weapon ‘Mechs definitely benefit more through the new Skill Tree system then they ever did in the Module system with the introduction of the universal weapon nodes.

That said, it is critically important to stress that we are not looking at this Skill Tree as a way to directly address ‘Mech to ‘Mech imbalances or Inner Sphere and Clan tech imbalances. While we do have ways in which the IS benefit more from the Skill Tree by virtue of either different values, or with the way native Quirks interact with the Skill Tree in cases like Armor/Structure Quirks and the Survival Tree, we cannot rely on a progression system to be the solution to something that must be fixed at the base level.

To address these points we are targeting other changes that will see improvements to balance along these lines in future patches.

Q: So how do you account for the fact that a number of already high-performing ‘Mechs, particularly on the Clan side, are perceived to disproportionately benefit from the Skill Tree? These are already strong ‘Mechs with no offensive Quirks, but under the new Skill Tree players will be able to apply a list of optimized offensive Skills. Alternatively, many IS ‘Mechs rely on their current Quirks to be viable; why the blanket reduction of existing Quirks?

Chris: The current Quirk system is not producing the intended results for the experience we are attempting to provide with MWO. While Quirks have served us well up until this point as a system which assisted toward balance - fairly well in some cases - the results ultimately ran contrary to ‘Mech customization; one of the key features of MWO.

Overall, current Quirks have introduced a level of power creep that we feel is the wrong direction for MWO, with heavily Quirked ‘Mechs acting more as puzzles then avenues of personal customization. With often only loadouts that fully optimize around a ‘Mechs given quirks able to be considered in a “balanced” state, while loadouts which ignore, or do not care to customize into the very ridged quirks often massively underperforming loadouts that play exclusively to pre-canned quirks.
Loadouts often needed to be fully optimized around the ‘Mechs inherent Quirks in order for the ‘Mech to be considered in a “balanced”, optimal state. Loadouts which didn’t lean into the inherent Quirks would be taken with the knowledge that they’re sub-optimal. This was a restrictive system that needlessly caused entire Loadouts to underperform.

When one of the core pillars of the MWO experience is the ability to customize and personalize your ‘Mech through the MechLab, this is something we cannot ignore any longer. We intend to break this blind adherence to narrow offensive quirks as being what dictates an entire Chassis’ performance and balance.

This is not a decision taken lightly, and we are fully aware of the concerns raised about these changes. With the Skill Tree being released, we felt strongly that now was the time to address this long standing concern for the following reasons.

First and foremost, we do not want the quirks to be one of the primary factors for considering ‘Mech viability on the IS side. We further do not want said Quirks to add to the massive performance gulf between ‘Mech Loadouts optimized entirely around the Quirks, and those that are not.
With new tech coming, and a heavy focus on balancing the tech levels through the baseline equipment, we knew that we would have had to reduce the Quirks in this way regardless of the Skill Tree. We decided to do it now to ensure that players go into their Skill Tree in the context of this Quirk reduction now, rather than later. While many of our top veteran players would have been able to adapt from a change such as that if they had a large amount of HSP should we have kept the Quirks in, this is not something that most of our player base would have been able to do. And we did not wish to give them months of investing Skill Points into a Loadout optimized around the inherent quirks only to have that Mech’s Quirks severely reduced shortly after.

Additionally, we feel players may be underestimating some of the boons that the Inner Sphere will receive through the Skill Tree system, the engine desync, and the way these interact with the current Quirks. Engine de-sync in particular is going to have a much bigger impact than many players may believe at a glance. With Inner Sphere ‘Mechs being the primary recipient of many baseline Mobility increases, and Clans being the recipients of many of the decreases, this will have an impact. While the Mobility chart offers more percentiles than the previous pilot lab system, in cases of Clan ‘Mechs such as the KDK-3 which saw a 25-30% reduction to its base values, even with the added Mobility from the tree you will not see higher performance than you have under the previous system. Alternatively, many IS mechs will see greater boons then they have had previously.

This does not mean that we feel that the net result of all this will solve all of the Clan and IS balance concerns that are present. For this we are still laser focused on further baseline changes to address ongoing concerns of balance between the IS and Clan factions.

Q: If the skill tree system is not meant to address Clan and IS balance, what is your plan for balance going forward?

Chris: Base level balance is what we are working towards. And to this end, we are continuing to work on improving this. To give you guys a sneak peek as to what is coming, these are some of the things currently being worked on to better balance the game:
  • June will see an extensive balance pass performed on Energy weapons, which will touch nearly every weapon currently in the energy lineup. This balance pass will focus on two major points:
    • Better weapon role definition between different weapon types.
    • A heavy focus on better baseline balance between IS and Clan Energy weapons.
  • Beyond this, Quirks are being extensively re-evaluated as we get data from Skill Tree release. We will be immediately focused toward those IS chassis’ receiving new Hero variants in June. We’ll possibly also see new Quirk passes on existing standard variants along with those Hero releases. In addition to this, we are looking at introducing new non-weapon Quirk types to add more variety to the way we currently Quirk mechs, though at this time, there is no ETA on the new Quirk types.
  • The Skill tree will be monitored heavily after release. We will have new nodes introduced with the new tech in July, and we will continue to refine the overall Skill Tree balance as we move towards introducing the nodes for that new tech.
  • The final balance point we can briefly highlight now involves Engine types. Better balance across the Standard, XL, and Clan XL Engine types is a major change that I am heavily pushing for. It's currently in the very early stages of evaluation, with gameplay and technical considerations to be made. At this time, there is no solid ETA or info I can provide, but it will definitely be something we look at addressing after the introduction of new tech.
Although these are the only things that we are able to address now, many more things are being looked into behind the scenes.

We want to assure everyone that balance remains a core focus of the design team, and we will be working heavily to monitor and improve it after the release of the Skill Tree in May, the release of new tech in July, and on into the future.

1 user(s) are reading this topic

0 members, 1 guests, 0 anonymous users