88 replies to this topic

[Mcgral18]

I've tried doing a little bit of testing about the new Match Score, and have had some success...and some seemingly random numbers.



*KMDD seems to remove a Hit and Run (one H&R per KMDD), otherwise the stats stack as you would expect, completely additive. That one is probably a bug, or my numbers are off.

Wins stayed at 20 points.

From my limited testing, we found most bonuses (Flanking, Hit&Run, Scouting, etc...) to be worth 2 points.
Not entirely true, Hit and Run provides a 1 bonus, as does Flanking, UAV Detection and it seems most others as well.

Suicides counted as -5 points. More specifically, Death counts as -2, and Suicide itself only 3...but you can't have one without the other. CW might be different.

Spoiler

One abnormality was after a Disco on my part, I lost 2 points. It wasn't Suicide (5 points, and I didn't kill myself), yet a win should yield 20 points.

A weird value...damage seemed to be 48% towards Match Score...I don't feel too confident in this one because it's just so random, but it kinda adds up (before throwing any other bonuses in it). Previously, it was 10%.
Tested by shooting 100 damage, then being killed. 0 points added from the win, and substract any Scouting, Flanking and Hit and Run (which can be avoided by being shot and being targeted).

50% of damage (not 48%) is counted towards Match Score. A death takes 2 points away, which caused this confusion at 3 in the morning.

Team Damage also seems off...as in, it didn't seem to affect the score at all, and especially not to the tune of 48% 50%. This hasn't changed, team damage doesn't seem to do a thing.
Team Killing however...that's a big hit. It's around a negative 50 points...which is pretty big.

Issues with this testing? Does dying (yes), or losing (no), count against your score? I haven't tested that yet (but there is a CW bug in which apparently if you don't die, you get exceedingly high scores, over 1000 in extreme cases).

In CW, it uses nearly the same score system as normal matches, with the exception of the Win+20...I only have 1 entry so, potentially rife with flaws, but my math had a 1 point difference compared the actual match score, which could be explained by damage rounding.

Damage test:

Spoiler

Spoiler

Both seem pretty conclusive (-2 for scouting on the second).


Now, there are some things I'm not too clear on, which include Component Destruction (assuming 2), Killing Blow (assuming 2), Solo Kill(presumed 4), Kill Most Damage Dealt (assuming 4), and Saviour Kill (utterly unknown).
Killing Blow seems to be worth 4, Component Destruction 2, Solo Kill seems to be around 8 (High, but makes up for KMMD being 0). Kill Most Damage Dealt seems to do nothing, and actually (sometimes) removes a bonus (Hit and Run) from your match score.

When I try to break down a live match and taking all the bonuses, it never adds up. Saviour Kill might be the thing that's needed, or the killing stats are completely wrong. Didn't spend much time on those.

Other oddities include Stealth Tag not counting towards Match Score (or only 1 point?1 point is par for course in regards to bonuses), so those may very well not be a blanket 2 score each.


Has anyone done testing of their own? I've activated a month of Premium Time if someone wanted to help with some more in depth testing.

Something I'd like to test is damage on a win, which theoretically should just add 20 points, but it helps either confirm or debunk the above values.


Anyone have Data? Any ideas on how to test (and if possible isolate) certain values? We had some ideas...but then we discovered formation bonuses, Protected Light/Med, and other things that kept adding (seemingly) +2 Match Score per reward.


Anyone want to help FOR SCIENCE?!

Edited by Mcgral18, 26 August 2015 - 12:05 AM.

[Kristian Radoulov]

You have done God's work, but I find myself wondering if you could make a cliffnotes version for the 4AM crowd?

Also I'll help you tomorrow night if you still need people.

Edited by Kristian Radoulov, 24 August 2015 - 12:06 AM.

[ApolloKaras]

Kristian Radoulov, on 24 August 2015 - 12:05 AM, said:

You have done God's work, but I find myself wondering if you could make a cliffnotes version for the 4AM crowd?

Also I'll help you tomorrow night if you still need people.

Cliff notes do not exist in the realm of, SCIENCE!

[Kjudoon]

For science? Quick! Someone call Cave Johnson!

Throw science at the wall and see what sticks!

[Mcgral18]

Kristian Radoulov, on 24 August 2015 - 12:05 AM, said:

You have done God's work, but I find myself wondering if you could make a cliffnotes version for the 4AM crowd?

Also I'll help you tomorrow night if you still need people.

I'm thinking some of these numbers are wrong...but at the moment, any bonus is at +2 points, nearly half of your damage is Match Score, and killing values are unknown to me.



More SCIENCE is needed. And scientific methods.

[Kiiyor]

I'M IN.




I'll rip some screens soon, i've been keeping track of my own games. I don't normally snap the detailed stuff though, so i'll have to start doing that.

Kjudoon, on 24 August 2015 - 12:13 AM, said:

For science? Quick! Someone call Cave Johnson!

Throw science at the wall and see what sticks!

[Mcgral18]

Now...obviously this doesn't work for CW, else anytime you breach 2000 damage you'd have over 1000 Match Score.

However...perhaps that ties into dying.

[Alistair Winter]

Looking at average damage, KDR and WLR in this sample, I have determined that you're all Tier 5 pilots.

Also, kind of curious how PGI will handle this, since they didn't want to release the information.

[Mcgral18]

Alistair Winter, on 24 August 2015 - 07:49 AM, said:

Looking at average damage, KDR and WLR in this sample, I have determined that you're all Tier 5 pilots.

Also, kind of curious how PGI will handle this, since they didn't want to release the information.

They'll probably laugh, because all of it is likely wrong.

Probably some hidden values, or even different rewards having different values (which will take a long time to test).



Then there's the CW aspects, where not dying seems to follow the right formula, but then if you die even one it cuts it down significantly.



Weird stuff.

[Lily from animove]

hmm, later when i am at home, we can test some stuff in privates,

would take soem time, but bets thing is to make a list of what actions a match can have:
spotting, flanking, assits etc.

make only One of each per match and see how scores are affected. but this would take long because you have to wait for the 15 mins tumer to run out

also build testmechs, one with 10 alpha's so you amke sure the damage applies correct in the correct numbers.

I remember in the odl formula, we had for 15 damage = 1 score. so your 48% can come from some roundings.

@Alistar, PGi can't do anything, aside banning it from the forum, yet it won't help, because then people just test it internal and give it on 3rd party pages.

would be more clever giving the numbers at all, we could then at least better test if the system works correctly.

Edited by Lily from animove, 24 August 2015 - 08:19 AM.

[Mcgral18]

Lily from animove, on 24 August 2015 - 07:55 AM, said:

hmm, later when i am at home, we can test some stuff in privates,

would take soem time, but bets thing is to make a list of what actions a match can have:
spotting, flanking, assits etc.

make only One of each per match and see how scores are affected. but this would take long because you have to wait for the 15 mins tumer to run out

also build testmechs, one with 10 alpha's so you amke sure the damage applies correct in the correct numbers.

I remember in the odl formula, we had for 15 damage = 1 score. so your 48% can come from some roundings.

@Alistar, PGi can't do anythign, aside banning it from the forum, yet it won't help, because then people just test it internal and give it on 3rd party pages.

would be more clever giving the numbers at all, we could then at least better test if the system works correctly.

Rewards (and how to get them) were listed in the Patch Notes for Rewards 2.0. Should be more or less accurate, but some might have changed since then.

https://mwomercs.com...tch-notes-13345


Damage does seem to be worth a large chunk.

[Lily from animove]

Mcgral18, on 24 August 2015 - 08:00 AM, said:

Rewards (and how to get them) were listed in the Patch Notes for Rewards 2.0. Should be more or less accurate, but some might have changed since then.

https://mwomercs.com...tch-notes-13345


Damage does seem to be worth a large chunk.

l
its not hard to figure out, yet time consuming:

make the follow: make private match with 1vs1 and meet up.

two mechs, best is DWF's with some energywepaons eahc with a alpha of 10 damage. (to avoid any ammo explosions) bets is havign guns able to create specific damages

(no one tageting the other !!!)

Round one:

DWF1 meets DWF 2 and makes clear 100 damage.

DWF 2 meets DWF 1 does 50 damage.

then you have
Score 1 = 100damage*Y+C (C= tie scores)
Score 2 = 50 damage*Y + C

let it end in a tie (timeout).

you then clearly can figure out the delta score is the damage delta which defines the Score gotten from damage. and you also know the tie score value. It just will take 15 minutes to wait, quite long.

Round 2

Meet with an IS mechs1, high ST Internal HP. Put in each sidtorso a 2 weapons (probably alsers) . and hav an AC5 as "gunweapon".

try to destroy a single component on the other mech with the AC 5 (should work best) Do it on both torsi, in case the one torsi blows off not geetting critted.

The other mech can do a single targetting action.

let time run into a tie. Susbtract the tie scores and the damage score from both results and you hav the scores for a spotting and the scores for a component destructed.

This science is safe, yet a bit time consuming. But should be worth the time.

could be done for other boni as well, would require some more discipline keeping pilots to test further bonuses like tag damage and such.

Edited by Lily from animove, 24 August 2015 - 08:21 AM.

[Revis Volek]

Kristian Radoulov, on 24 August 2015 - 12:05 AM, said:

You have done God's work, but I find myself wondering if you could make a cliffnotes version for the 4AM crowd?

Also I'll help you tomorrow night if you still need people.

These test were done between 1 and 3 am...so this IS the shortened version. We were all sleepy, so pardon any thing we missed.

Another note, didn't see it in McGrals post but Flanking could only be achieved by a light mech in our testing while two mechs were targeting and shooting each other the light had to shoot the back of an untargeted mech to get the FLANKING trigger. No idea what it cost more then likely 2 pts.

More science would be needed to confirm or debunk most of the stuff we tested anyway....we need more stompy robot scientist! Please report to McGrals testing facility....

and like McGral stated Tag Stealth was only applied to the final score on a win (loss had no pts added for tag stealth) and it only gave one (1) point as far as i could figure as well.

Edited by DarthRevis, 24 August 2015 - 08:41 AM.

[Bilbo]

DarthRevis, on 24 August 2015 - 08:37 AM, said:

These test were done between 1 and 3 am...so this IS the shortened version. We were all sleepy, so pardon any thing we missed.

Another note, didn't see it in McGrals post but Flanking could only be achieved by a light mech in our testing while two mechs were targeting and shooting each other the light had to shoot the back of an untargeted mech to get the FLANKING trigger. No idea what it cost more then likely 2 pts.

More science would be needed to confirm or debunk most of the stuff we tested anyway....we need more stompy robot scientist! Please report to McGrals testing facility....

I get the flanking bonus regularly with an assault. I've never paid attention to whether or not the mech was targeted, but I always target mechs when possible. Could be that the flanked mech doesn't have you targeted.

[Jman5]

A few things you may want to try.

1. UAV detection and UAV kill
2. Assault mode base capture
3. Counter ECM

[stjobe]

Mcgral18, on 24 August 2015 - 08:00 AM, said:

Rewards (and how to get them) were listed in the Patch Notes for Rewards 2.0. Should be more or less accurate, but some might have changed since then.

https://mwomercs.com...tch-notes-13345

I made an easier-to-read version here: Rewards 2.0: What you need to do to get XP and CB.

If you get any use out of that one it might not have been all wasted effort for me...

[Mister Blastman]

Deaths must be a fractional divisor of some sorts whereas other values are weighted with a multiplier or to a power (or in a weighted spread).

I imagine they're using some sort of formula similar to what they use in tournaments with some tweaks to it.

[Revis Volek]

Mister Blastman, on 24 August 2015 - 08:52 AM, said:

Deaths must be a fractional divisor of some sorts whereas other values are weighted with a multiplier or to a power (or in a weighted spread).

I imagine they're using some sort of formula similar to what they use in tournaments with some tweaks to it.

Yea deaths or mechs left is some kind of multiplier....but 3 mechs left and 2000 dmg seems to very often break down to a score of over 1000....i just wish i could put my finger on it.

Bilbo, on 24 August 2015 - 08:44 AM, said:

I get the flanking bonus regularly with an assault. I've never paid attention to whether or not the mech was targeted, but I always target mechs when possible. Could be that the flanked mech doesn't have you targeted.

I know i have gotten it in game before too, but we did not have time to do any further testing...we were some sleepy giants out there.

But more testing will be done by myself to confirm this....Not sure if targeting is important or if its just shooting an already engaged target from the back.

[Bilbo]

DarthRevis, on 24 August 2015 - 08:59 AM, said:

Yea deaths or mechs left is some kind of multiplier....but 3 mechs left and 2000 dmg seems to very often break down to a score of over 1000....i just wish i could put my finger on it.




I know i have gotten it in game before too, but we did not have time to do any further testing...we were some sleepy giants out there.

But more testing will be done by myself to confirm this....Not sure if targeting is important or if its just shooting an already engaged target from the back.

According to stjobe's marvelous list, all you have to do is shoot them from behind without having been targeted by them or be engaged:

Flanking:

• Move behind an enemy Mech without being targeted by it and without being engaged.
  
• Deal damage to the enemy Mech from behind (out of sight).
  
• Reward has a cool-down of 20 seconds.

[ApolloKaras]

Kiiyor, on 24 August 2015 - 02:45 AM, said:

I'M IN.

I'll rip some screens soon, i've been keeping track of my own games. I don't normally snap the detailed stuff though, so i'll have to start doing that.

lol Its like shining the bat signal and Kiiyor arrives