5 replies to this topic

[ExAstris]

The Law of Large Numbers.



I'm going to write a lot, use math and logic, and be critical of design decisions MWO has made. So please be patient or move on. It is my hope as always that PGI (especially Paul) reads this, as it highlights critical factors in my enjoyment of this game.



Let me explain my answer.

12v12 has done one major thing to every factor in gameplay balance, it has put it against a slightly larger denominator, minimizing its impact to any individual match.

This is to say that things that previously caused individual matches to be completely wrecked, now are far less capable of completely wrecking a match. Before, if a premade 4man was 50% of your opposition, that meant half their team had coordinated their builds and tactics. If you dropped solo, that meant there were only two other slots on each side where 4 mans could fit, meaning if there were 2 4mans in the drop, 50% of the time you'd have one on each side, but 50% of the time both would be on the other team.

In short, we all know that premades have the largest statistical impact on who wins a match. And this situation was just quite unfair to solo players. So how did 12v12 improve this? Well, it put more people on both sides, so now instead of a 4man meaning half of an entire team is already coordinated (and in all likelihood, on a voice com), it now just means that 1/3 of their entire team is coordinated.

Further, by simple probability mechanics, we can show that far fewer matches will become the horrendous all-premade vs all-solo drops because every time a 4man drops it has a chance of dropping on either side. So in the previous 8v8 scenario, the odds were 50-50 that a solo player would vs an entire team of premades instead of a balanced premade + 4pugs vs premade + 4pugs. Now with 12vs12, there is the same chance each 4man will drop on either side, but now we have to have 3 4mans drop on the same side, with each beyond the first having a 50% chance to land on either side (because the first will always be counter to our solo player). That gives us a 1/4 chance that 3 premades will get stacked together on one team.

Further, adding or subtracting one from the total number of premades guarantees no team will be fully premade vs a team of full pugs.



Thus, the case of the matchmaker delivering a full premade team vs a full pug team is vastly, vastly less likely than before. And thus one major factor in the more closely matched 12mans stands revealed.



Any factor that could be previously exploited to gain a huge advantage has been minimized by simply adding more players, but not changing any of the other parameters those players drop by.






Two more factors whose influence has been reduced.
Disconnects.
ECM.

These are the next two contenders on the list for statistically viable ways to predict winners. Disconnects are obvious. One team with 8 players is more likely to win than their opposition with only 7 players. The likelihood of the disconnect being the deciding factor when its 12vs11 is quite a bit lower than when its 8v7. Again, simply making the numbers larger has ironed out a match balance issue (which is fortunately far less prevalent thanks to netcode improvements).

ECM. I don't know if PGI still doesn't realize how incredibly powerful ECM is, or if they just don't care anymore. Every one of my ECM mechs has a hilariously inflated K/D and W/L ratio compared to its non-ECM counterparts, even when those are clearly superior in every other way (my 3M vs my 2A). I've noted multiple times elsewhere why this is the case, and what needs to be done to correct it. (Hint: the slower pace your game runs at, the more powerful information control is)

But anyways, ECM is also a diminished factor in single-handedly winning games. Just like premades and disconnects, it still does from time to time. But with the larger player numbers, the likelihood of one team dropping with 4 ECM and the other none is much smaller than it was before, meaning a much higher chance of fair matches. This doesn't change the fact that ECM still needs to be changed, and has needed it since its introduction, however, its horrendous effects being minimized has made a noteworthy contribution to my enjoyment of the game.






The Lessons we should learn:
There are clearly things about MWO's match setups that make one team more likely to win than others simply in virtue of how the MM dropped them. While disconnects can't be entirely addressed (local connectivity issues aren't MWO's problem), the other two issues can.

Premades: How we could make it worse.
We could screw this all up by letting premades of 6 join regular matches. We'd be right back where we started, letting half of a team be prepared and coordinated in a way that regularly dominates a match. The reward for playing with friends shouldn't be pugstomping, it should be getting to play with your friends.

Premades: How we could make it better.
Add premades to the matchmaker algorithm. It doesn't have to be terribly strict, but just prevent silly mismatches like two 4mans on one side vs none on the other. Make sure each 4man that enters has a 3 or 4man on the other team. Make sure each 3 man has a 2, 3, or 4 man on the other team (in roughly the same elo bracket), and that each 2man has a 2 or 3man opposing it.

Another option to improve match balance would be to simply make the maximum premade size smaller. But this isn't a good option for the game in my opinion. People are use to dropping with 4, and cutting that back would only be whine inducing when there is a much better solution in the previous paragraph. Similarly, making a solo only queue would improve the match balance for solo players, but would segregate the community and drop queues making drop times and variety far worse for everyone. The solution above trumps both of these.

ECM: How to make it worse.
Let ECM fully cancel missile locks again. Or introduce another piece of equipment with pitifully small mounting costs that completely dominates information warfare and horrendously messes with the balance of an entire category of weapons. (seriously, BAP is a tax for all streak users now. Imagine if every PPC/Gauss boat had to mount an extra piece of equipment to be able to fire its weapons just because an enemy unit got within range and had a Magnetic Field Disruptor on board. And that's not even ECM's biggest problem!)

ECM: How to make it better.
Remove the stealth field. Information warfare is 95% ECM right now. Every other piece of information warfare is either a half-measure to pull back the overly effective ECM or a superfluous nicety that gets overrun by it.

Secondary alternative: Let BAP see through the stealth field to obtain locks at all ranges, but still only full counter at 150m. The stealth field is really that powerful.

Baseline requirement: Items that are way out of line in terms of the benefits they bring to the field vs the cost it requires of your mech's loadout cause imbalance in the matches. Normalizing that ratio for all equipment removes the need for large matches to obscure their nasty effects.






TL;DR. Matches in MWO tend to be more fair with 12v12. But its only because putting more people on both sides has minimized the impact that the existing match-destabilizing features have. We could further improve match quality by normalizing unbalancing mechanics (ECM) and/or putting the most statistically significant match determiners into the matchmaker itself (premades).

[AlmightyAeng]

There's a simpler explanation, too. There's more fodder between the good players who take advantage and shoot...you have longer to make a mess of the enemy, so the numbers go up. Just sayin'.

[Joseph Mallan]

Your explanation of Groups in 12 man may be off a bit. a Team could be comprised f any combination of smaller teams 1-4 man strong or 2-3 4 man Sync drops... Which to my observation hasn't happened yet.

Other than that I think you covered it pretty good.

Edited by Joseph Mallan, 14 August 2013 - 11:48 AM.

[Frisk]

Good post, complete thought sentences, paragraphs... really good stuff.

I think ECM isn't nearly the OP win button it was when it was first released. PPC counter for ECM (imo) really leveled things out.

There have been suggestions in separating the stealth and jamming capabilities, I think this is a good solution.

[Levi Porphyrogenitus]

12v12 does do wonders for mitigating problems just by the magic of dilution. I've definitely found it to be an improvement over 8v8 in most respects.

My opinion about premades and such rests almost entirely on a combination of pre-game lobby (after the MM has found a team for you) and in-game voice comms. Once those go live a lot of what elevates a premade will go away, as PUGs will be able to coordinate team comps, call targets, etc. Even the simple expedient of adding a "focus target" hot key to lance and company commanders would go a long way.

ECM is mostly a problem for the simple reason that it is a hard counter to anything (actually, a hard counter to several things), and is only countered in turn by other hard counters.

The fix for it is to change the whole system into one based on soft counters.

1 - Have ECM impose penalties on enemy missiles and electronics.

1a - It should educe missile guidance capability (easier evasion since its targeting updates from the launching mech are suffering interference).

1b - It should increase lock-on times (signal interference causes the computer to take longer to get a firm firing solution).

1c - It should increased missile spread (individual missiles deviate more from their proper flight path as they get bad data due to ECM interference).

1d - It should reduce the broadcast range of NARCs that are inside its envelope.

2 - Have BAP provide bonuses that mitigate the penalties of enemy ECM (or provide outright boosts to uninhibited electronics).

2a - It should extend base sensor range for the carrying mech.

2b - It should allow for detection of any mechs regardless of LOS when inside 90m and shut down mechs in LOS inside 180m.

There is no need for BAP to counter ECM as ECM does not prevent anything. BAP's primary strength should be its LOS-free detection and sensor contacts inside a very short range.

3 - Give Artemis IV, TAG, and NARC specific roles by giving each one a unique boost to missile performance that interacts with ECM in a different way.

3a - Artemis IV should require LOS to work, and should improve tracking capability (more maneuverability) and tighten missile groupings. It's a very heavy investment for missile carriers and its benefits cannot be shared, so it deserves to be potent.

3b - TAG should decrease lock-on times. Since it and Artemis IV give different benefits it would be fine for them to stack, but not necessary.

3c - NARC should broadcast a lockable contact to all friendlies inside 750m (375m if the stuck target is under ECM). It should last for one full minute or until the stuck location is destroyed.

This gives each piece of gear its own role, and allows them to interact with ECM in their own way. Nothing entirely counters anything else, as it becomes entirely a matter of degrees of performance.

[keith]

to make it not a wall of text. now a good player will take 1 random alpha from a bad player because they can't see them. this in turn makes them red. when they come across a decent player they are already half hurt in their CT. while the dice roll game of the other good player being hit was not. the one player who was hurt dies. this in turns makes the game less skill based. so far everything PGI has done makes the game less based and more pug friendly.