43 replies to this topic

[ForceUser]

Quote

Its every player's lifetime stats from Jarl's list.

Those are not lifetime stats, only stats since the introduction of the leaderboard. Basically the last 18 months, not the last 5 years.

The biggest problem with Match Making in MWO is the impact the mechlab has on the match. No other game that I am aware of has a mechanic that changes the power level of your vehicle/avatar so much before you even load into a match, apart from maybe MMOs PvP. But even then there are a lot of mechanics in many of those games to normalise gear stats and many abilities are fixed/inherent to the class.

Basically until you have fixed load outs that are perfectly balanced with zero or little room for modification, no match will ever be particularly balanced. While you can put together equally skilled players, the mechs, and more importantly builds on those mechs, will have as big or perhaps in some cases even bigger impact on the match. Due to the freedom of versatility of the mech lab, no it's isn't realistic to expect a match maker to try and guess the power level of a build. You'd need AI far more advanced than what's currently been developed to differentiate and gauge the strength level of a system where there are quite possibly billions of different combinations.

It'd be different if you could perhaps work under the assumption that people would go into a match the the optimal setup for each mech, but people don't. I'd say well over 90% of the matches I've ever played has been in unoptimal situations. Ranging from fun/unoptimal builds to unskilled mechs or previously mechs without modules/skills or off meta builds.

Still I don't think the PSR is the best solution.

Edited by ForceUser, 14 January 2018 - 08:12 PM.

[Lily from animove]

ForceUser, on 14 January 2018 - 08:11 PM, said:

Those are not lifetime stats, only stats since the introduction of the leaderboard. Basically the last 18 months, not the last 5 years.

The biggest problem with Match Making in MWO is the impact the mechlab has on the match. No other game that I am aware of has a mechanic that changes the power level of your vehicle/avatar so much before you even load into a match, apart from maybe MMOs PvP. But even then there are a lot of mechanics in many of those games to normalise gear stats and many abilities are fixed/inherent to the class.

Basically until you have fixed load outs that are perfectly balanced with zero or little room for modification, no match will ever be particularly balanced. While you can put together equally skilled players, the mechs, and more importantly builds on those mechs, will have as big or perhaps in some cases even bigger impact on the match. Due to the freedom of versatility of the mech lab, no it's isn't realistic to expect a match maker to try and guess the power level of a build. You'd need AI far more advanced than what's currently been developed to differentiate and gauge the strength level of a system where there are quite possibly billions of different combinations.

It'd be different if you could perhaps work under the assumption that people would go into a match the the optimal setup for each mech, but people don't. I'd say well over 90% of the matches I've ever played has been in unoptimal situations. Ranging from fun/unoptimal builds to unskilled mechs or previously mechs without modules/skills or off meta builds.

Still I don't think the PSR is the best solution.

PSR is just a name. What matters are the caclulations behind it. But neither MS (which it is based uppon) nor the rulesets for ups and downs in the PSR are well chose. At this point they started destroying stuff in terms of proper making matches.

Edited by Lily from animove, 15 January 2018 - 07:39 AM.

[sub2000]

Davegt27, on 12 January 2018 - 02:04 AM, said:

---irrelevant and wrong stuff---

When you play FW, first thing i do I press TAB and look at players names. After a while you can often predict well how the game will go, and what will be the result. There is no MM in FW and strongest players tend to group up together. It is possible.

When I play QP lately, I press TAB, look at players names and, "what a surprise!" regularly I see familiar tags and I can predict game outcome before match is even started. More often than not, a couple of strongest players get on one side of the MM pool and fight not enemy really, but compete who collects more kills. Just a week ago I had series of matches where two dudes from the first 100 in the leaderboard (K2D) were in the same team, against me or with me (when I changed mine mech class). From MM point of view such players are indistinguishable from 0.5K2D Lrm "god" who collected enough games to end in the Tier 1.

[UnofficialOperator]

ForceUser, on 14 January 2018 - 08:11 PM, said:

Those are not lifetime stats, only stats since the introduction of the leaderboard. Basically the last 18 months, not the last 5 years.

That's pretty pedantic but ok. Regardless, using just the last 2 seasons stats should br a better gauge of someone's skills and given more weightage.

ForceUser, on 14 January 2018 - 08:11 PM, said:

The biggest problem with Match Making in MWO is the impact the mechlab has on the match. No other game that I am aware of has a mechanic that changes the power level of your vehicle/avatar so much before you even load into a match, apart from maybe MMOs PvP. But even then there are a lot of mechanics in many of those games to normalise gear stats and many abilities are fixed/inherent to the class.

There are many games with their individual mechanics which are a variation of the mechlab. Even COD.

ForceUser, on 14 January 2018 - 08:11 PM, said:

Basically until you have fixed load outs that are perfectly balanced with zero or little room for modification, no match will ever be particularly balanced. While you can put together equally skilled players, the mechs, and more importantly builds on those mechs, will have as big or perhaps in some cases even bigger impact on the match. Due to the freedom of versatility of the mech lab, no it's isn't realistic to expect a match maker to try and guess the power level of a build. You'd need AI far more advanced than what's currently been developed to differentiate and gauge the strength level of a systechm where there are quite possibly billions of different combinations.

It'd be different if you could perhaps work under the assumption that people would go into a match the the optimal setup for each mech, but people don't. I'd say well over 90% of the matches I've ever played has been in unoptimal situations. Ranging from fun/unoptimal builds to unskilled mechs or previously mechs without modules/skills or off meta builds.

Still I don't think the PSR is the best solution.

Yes well I would also like to see every mech have their own pilot rating (you can be a T4 in light but T2 in asssult etc). Then you further combine BV of the mech and use it in MM. Maybe in the next iteration of the franchise eh?