No replies to this topic

[Victor Morson]

I've seen a lot of talk/arguments about the way matchmaker is working right now, and thought I would throw out detailed alternative again, based in a BV-style system. I'm going to break it down into a lot of details here, but please try to remember these numbers are placeholder, and obviously would be subject to adjustment! The core concept, however, I think would be an easy solution to a lot of problems.

The point of this system is to balance matches based on both equipment and player skill, while also making it easier for the matchmaker to put together teams.

---

Determining Player Value

Each player's Value can be altered by a few things:

Skill Value - Using ELO or another method to track pilot performance, this could account for a small portion of the total value. The more skilled the pilot, the more points. Say for the purposes of this example we will say it will offer a 0.1 multiplier for each "rank"/bracket the player goes up.

Mech Value - While there are a milion ways to calculate this, I will use some simple examples for a starting value. If done properly, this can literally be used to balance matches around powerful and weak 'mechs even if they aren't adjusted in-game.

• Base Value: A base value based on tonnage. We will say 100 points per ton.
  
• Tech Value: Units with Ferro may get a 1.1 multiplier while units with Endo a 1.3 multiplier. This means an Awesome with Endo would be 104 points so far. This could be applied past this, to things like Double Heatsinks, which means Single Heatsinks greatly help in not raising the Mech Value.
  
• Weapon Value: This gets trickier to provide examples without spamming a list, so I'll pick one: Say the PPC. We'll assign the PPC a value of 20. So an Awesome with 4 PPCs would gain 80 Value base.
  
• Weapon Multiplier: To reflect the power of boats, this can be applied to the Weapon Value. For example, let's say each PPC costs 20% more than the last over the first one. So instead of the above 80, the Awesome would pay 80 for the first, 96 for the second, 115.2 for the third, 138.24 for the fourth, meaning 4 PPCs now cost the Awesome 429.44 Weapon Value instead of 80. That is enough for an entire base Cicada, in terms of points.
  
• Tech Base Multiplier: Clan tech can take a direct 150% cost on the final BV, meaning that Clan Tech that's optimized and large will take a dramatic amount out of the team compared to lighter models.
  
• Other Multipliers: There's room for infinite tweaking here. Values for modules, certain equipment (Artemis for example could give an increase in Weapon Value to missiles that use it), etc. Any aspect of the 'mech could adjust it's Value. They could even be applied to boost the value of organized groups based on their size, increasing their value slightly to reflect their organization. Exceptionally good or bad chassis would also be given modifiers- i.e. the Cataphract could cost 10% more than a normal 70 tonner, but the Dragon may cost 20% less.

The best part is all of this could be handled completely transparently in the UI. In fact, I suspect competitive leagues would swap to this system over tonnage in very short order.


The Matchmaking

You're probably wondering how this will help Matchmaker. Simply put, this is all the actual Matchmaker now needs to do:

• Look at the available players & groups, but purely as their Player Value. After the number of players / average value is checked, Matchmaker can begin filling a lobby with a target BV set specifically based on what's available.
  
• At this point Matchmaker need only sort out the various Player Values (including Team Player Values) onto the teams from what's available, within a certain margin of error: One that could change depending on the number of active players at that time. All skill, weight, and tech advantages were calculated before Matchmaker even got involved. Thing of this as the computer playing Tetris with numbers - something it's very, very good at.
  
• Begin the match!

Advantages


The first advantage is matchmaking time. No longer would it need to worry about perfect ELOs or weight classes. It would all, in the end, be about everyone having a set value number, and stuffing them into whatever team they'll fit on.

It also offers a way to nerf powerful setups (without impacting gameplay), reward lower-tier mechs, etc: If a top tier Firestarter offers a similar Player Value to a bottom tier Cataphract, it makes for a hard decision. Keep in mind PGI could lower/raise the Weapon/Chassis value costs based on metadata pretty easily once this is put in.

This also, most importantly, offers a way to balance things like Clantech or even large groups of players in the public que, as they could be given inflated numbers to reflect these things.

In Practice

I'm convinced if this were to be the way of the future matchmaker, you'd effectively have very asyemtrical matches (which are very interesting - far more than 3/3/3/3) while still offering a very balanced experience - probably far more than what we have with the current system, given a Dragon and Mad Cat are utterly interchangable for the Heavy Slot as just one example.

Powerful boats could be driven, but they would either take much out of the team's overall resource pool, or be pitted against a team of a similar composition. In fact, it'd be easy to set a maximum Team Value for team launches, specifically to enforce this.

In Conclusion

I've long held that I've been less in favor of this system because of it's roots in BattleTech (BV 2.0 is far superior to BV 1.0 though!) and more because it can be modified into working VERY well for matchmaking a game like this. Add in to the fact that it using a far simplier method for the actual matchmaking, and there are very, very few downsides. I think casual players would appreciate this greatly, and competitive players would easily shift towards Team Values from Tonnage.

All in all I think this would speed Matchmaker dramatically, offer far more varied and interesting matches, and also help balance out problems in the game without even touching gameplay directly. I'd love to hear what folks think, and please again remember all numbers contained within are examples for the purpose of demonstration and not suggested final values.

Edited by Victor Morson, 22 September 2014 - 02:19 PM.