

#1
Posted 07 February 2025 - 06:12 PM
This is a follow-up to the feature suggestion post that I had made. https://mwomercs.com...5-chat-balance/
P.S. Thanks for all the work Devs! And thanks to all the awesome players out there! GLHF 07
#2
Posted 07 February 2025 - 10:21 PM
Strelok7, on 07 February 2025 - 06:12 PM, said:
What do you mean by "plus/minus 3 players is still acceptable"? I think you are making a statement about your definition of a stomp (steamroll) being a difference in surviving team members of four or more. There has been lots of discussion in this forum, over many years, about what constitutes a stomp. The high level of the subject is that kill count by team, while definitely a signal, does not always tell the whole story.
Can you give us some acceptable outcomes? I ask because I have lots of data (thousands of matches, across multiple tiers) and could provide some numbers.
Out of 3,497 matches of data I have saved:
- About 89.5% (3,133) end with the defeated team having no survivors.
- The most common outcome is the victory team having seven survivors (about 14.6%).
- Given the bullet point above, most (about 93%) of the seven survivors for the victory team outcomes are 12:5 outcomes (victory team destroyed twelve, defeated team destroyed five).
- A 12:0 outcome is rare at about 1.4% (52 matches).
- About 15.6% (548) are +/- 3 survived, interpreted as a three survivor difference between teams (irrespective of victory). Examples: victory seven, defeated four, and victory three, defeated six.
#3
Posted 07 February 2025 - 10:51 PM
#4
Posted 07 February 2025 - 11:01 PM
#5
Posted 08 February 2025 - 01:52 AM
Yeah, I am therefore basically not playing any more.
The matchmaker plain sucks.
#6
Posted 08 February 2025 - 03:36 AM
#7
Posted 08 February 2025 - 06:02 AM
#8
Posted 08 February 2025 - 06:25 AM
Some dead start crying "dont waste time". F**k them, you still alive and try to kill somebody. Bad, they can tell your position to enemies.
Edited by Saved By The Bell, 08 February 2025 - 06:27 AM.
#9
Posted 08 February 2025 - 06:44 AM
#10
Posted 08 February 2025 - 08:47 AM
#11
Posted 08 February 2025 - 09:20 AM
And generally speaking its about 50-50 in games outside that range and within.
Are there days where you can have a variance of that? Absolutely.
But the "GOOD" matches where its close are the rarity.
Its all down to matchmaker honestly, but eh. There is a wall when it comes to that discussion.
Try to enjoy it as best as can. Use comms and do your best and try to change the outcome to the best of your ability.
#12
Posted 08 February 2025 - 11:56 AM
also if you happen to die in a good match that ends up winning, your brain usually counts that as a loss. especially if you quit before the conclusion.
#13
Posted 08 February 2025 - 01:08 PM
- Poor matchmaking can certainly contribute and there are so many variables that this matchmaker, and in some cases any matchmaker anywhere, cannot really account for.
- The nature of this game, the non-respawn and fixed team sizes, combined with loss of damage output as the team loses mechs, does create a snowball effect. For a game to go "neck and neck" means a lot of very even trading and even loss of mechs. These outcomes are rare because these situations are rare.
- Even a perfect matchmaker will not change singling targets out, asymmetry in applied damage output, and so on. Not unless one person from each team go into a corner, pair by pair, and duke it out and the winner sits out the rest of the match or something.
- I generally feel that 12:2 or less is probably a good signal that the teams were effectively unbalanced. Overall, this seems to happen about 15% of the time in my data, starting at about 10% for Tier 5 and increasing to about 18% for Tier 2. I have no Tier 1 games.
Saved By The Bell, on 07 February 2025 - 10:51 PM, said:
In full disclosure, I have no Tier 1 games as the highest Tier of any account I have is 2. That said:
- There is an increasing trend for counts of 12:0 matches with increase in tier. 0.9% in Tier 5, 1.3% in Tier 4, 1.9% in Tier 3, and 2% in Tier 2. I do not have a lot of games in Tier 2, so take that percentage with a grain of salt. Still, 10% seems wildly high, but maybe Tier 1 games see this kind of snowball effect more often.
Meep Meep, on 07 February 2025 - 11:01 PM, said:
12-0 to 12-4 are blowouts in my book
...
Not the majority of games but its not too far off.
Moadebe, on 08 February 2025 - 09:20 AM, said:
And generally speaking its about 50-50 in games outside that range and within.
Globally (Tiers 2 through 5) in my data, using the 12:4 or less definition of a stomp, I see about 38% of matches fall into this category. However:
- I see a similar split for each tier, however there is also an increasing trend similar to the 12:0 match counts. About 30% in Tier 5, about 37% in Tier 4 and Tier 3, and about 41% in Tier 2.
LordNothing, on 08 February 2025 - 11:56 AM, said:
there is a bias, i forget the name, where by people remember the bad games more than the good ones.
Agreed. Bob ate a berry once and then he died, better remember to not eat that berry.
#14
Posted 08 February 2025 - 06:02 PM
#15
Posted 08 February 2025 - 09:58 PM
dubstep albatross, on 07 February 2025 - 10:21 PM, said:
I ask because I have lots of data (thousands of matches, across multiple tiers) and could provide some numbers.
Sure. But don't spend too much time on it... How about One Team 10 survivors vs Another 0 Survivors. Actually... I see, you are saying it is about 15% of the time.
11 to 0, 9 to 0 ?
#16
Posted 09 February 2025 - 01:46 AM
dubstep albatross, on 07 February 2025 - 10:21 PM, said:
Out of 3,497 matches of data I have saved
Are you interested in more data? I could save and send it to you. What do you need?
#17
Posted 09 February 2025 - 02:20 AM
2-12 with me getting an = they even had one disco.
1-12 with me getting an ^
more than half team every time below 200 damage.
I have screenshots if you wanna see.
Same as yesterday or any day really.
Unplayable. And no fun.
What kind of matchmaking is this? Put all the people who can not aim in one team?
Edited by Hagen Zwosta, 09 February 2025 - 02:23 AM.
#18
Posted 13 February 2025 - 09:33 PM
Glymbol, on 09 February 2025 - 01:46 AM, said:
Are you interested in more data? I could save and send it to you. What do you need?
Thank you for the offer. I will keep it in mind. I built a custom OCR pipeline. It requires a specific resolution because of how I built it. I have been considering reaching out to collect data from the community, but it may not necessarily be a simple task in the end. Plus, if I am being honest, I am a bit lazy.

#19
Posted 13 February 2025 - 10:08 PM
Hagen Zwosta, on 09 February 2025 - 02:20 AM, said:
2-12 with me getting an = they even had one disco.
1-12 with me getting an ^
more than half team every time below 200 damage.
I have screenshots if you wanna see.
Same as yesterday or any day really.
Unplayable. And no fun.
What kind of matchmaking is this? Put all the people who can not aim in one team?
If I may ask, what tier are you in?
#20
Posted 14 February 2025 - 09:22 AM
so good luck for today
(most people is off so mm doing job)
Edited by Saved By The Bell, 14 February 2025 - 09:24 AM.
3 user(s) are reading this topic
0 members, 3 guests, 0 anonymous users