EvenMatchOmni

Discuss and provide feedback on Maps.
User avatar
pooty
Posts: 4358
Joined: Sat Apr 03, 2021 10:22 am
Location: Michigan
Server Sponsor: Yes
Server Admin: Yes

Re: EvenMatchOmni

Post by pooty »

I think its still a bit odd when bots are involved. Not sure why that matters, but I've noticed it doesn't work quite right with bots.
User avatar
McLovin
Posts: 1165
Joined: Sat Apr 03, 2021 12:54 pm
Location: Salt Lake City, Utah
Server Sponsor: Yes
Server Admin: Yes

Re: EvenMatchOmni

Post by McLovin »

captainsnarf wrote: Fri Oct 07, 2022 10:45 am if teams are even numbered, new joins go to the weaker team
What does "weaker team" mean?

1. The team with the least amount of points in the moment?
I think Jubei took the first round, so his team was ahead in round score when Dognuts joined. But in the moment Dognuts joined I had all the nodes.

2. The team with the lower overall player history scores?
Not sure what Jubei's overall history score is but it can't be higher than mine on this map.
User avatar
captainsnarf
Posts: 2632
Joined: Tue Jul 06, 2021 1:51 pm
Location: Washington
Server Sponsor: Yes
Server Admin: Yes
Contact:

Re: EvenMatchOmni

Post by captainsnarf »

Weaker team = team with lower points per hour rating. Game score is not used when determining balance.
User avatar
captainsnarf
Posts: 2632
Joined: Tue Jul 06, 2021 1:51 pm
Location: Washington
Server Sponsor: Yes
Server Admin: Yes
Contact:

Re: EvenMatchOmni

Post by captainsnarf »

Also, PPH fluctuates as you play. It might have been that at that moment, Jubei's PPH was higher than yours. Maybe he just finished building a node while you were driving around when dognuts joined.
User avatar
pooty
Posts: 4358
Joined: Sat Apr 03, 2021 10:22 am
Location: Michigan
Server Sponsor: Yes
Server Admin: Yes

Re: EvenMatchOmni

Post by pooty »

So Enyo was asking about the Minus-KS balancer:

Code: Select all

EvenMatchDebug: Balancing team sizes and PPH...
EvenMatchDebug: Assigning leon (691.23 PPH) and Halloweeny)o( (518.83 PPH)
EvenMatchDebug: leon will be on red (now 691.23 PPH), Halloweeny)o( will be on blue (now 518.83 PPH)
EvenMatchDebug: Assigning nautikal (410.31 PPH) and Super_Sanka (408.22 PPH)
EvenMatchDebug: nautikal will be on blue (now 929.14 PPH), Super_Sanka will be on red (now 1099.46 PPH)
EvenMatchDebug: Assigning AnKeeDo (390.85 PPH) and EbNo (363.54 PPH)
EvenMatchDebug: AnKeeDo will be on blue (now 1319.99 PPH), EbNo will be on red (now 1463.00 PPH)
EvenMatchDebug: Assigning Staff_Sergeant (300.51 PPH) and p)o(Oty (275.30 PPH)
EvenMatchDebug: Staff_Sergeant will be on blue (now 1620.50 PPH), p)o(Oty will be on red (now 1738.29 PPH)
EvenMatchDebug: Assigning Thomas_Jefferson (255.02 PPH) and DW>RESURRECTION (199.36 PPH)
EvenMatchDebug: Thomas_Jefferson will be on blue (now 1875.52 PPH), DW>RESURRECTION will be on red (now 1937.65 PPH)
EvenMatchDebug: Assigning PhRoZeN (192.24 PPH) and ZDragonMan (173.96 PPH)
EvenMatchDebug: PhRoZeN will be on blue (now 2067.76 PPH), ZDragonMan will be on red (now 2111.61 PPH)
EvenMatchDebug: Assigning plecostomus (154.89 PPH) and Jubei (152.87 PPH)
EvenMatchDebug: plecostomus will be on blue (now 2222.65 PPH), Jubei will be on red (now 2264.48 PPH)
EvenMatchDebug: Assigning Coldcut (144.19 PPH) and 7 (135.75 PPH)
EvenMatchDebug: Coldcut will be on blue (now 2366.84 PPH), 7 will be on red (now 2400.23 PPH)
EvenMatchDebug: Odd player count, Red has higher PPH, assigning I'm_Sorry to blue (77.43 PPH)
Looks pretty decent.

I think we should go back to map based PPH. Some players like Levis/Tanks and get good scores and aren't much for DM, others like DM and don't do much with minos/levis.... and remember its not purely Map based or overall its a blend of it. Plus the knownplayer multipliers are on good and stains.
Plus like AnKeedo, he likes megas and racks up a huge PPH playing with 3 people early in the day, but I don't think his 390+ PPH holds up later...on non Mega maps.

Snarf, I'll turn it back on and look to make it an .ini file option.
User avatar
captainsnarf
Posts: 2632
Joined: Tue Jul 06, 2021 1:51 pm
Location: Washington
Server Sponsor: Yes
Server Admin: Yes
Contact:

Re: EvenMatchOmni

Post by captainsnarf »

I was wrong when I said I commented out code. It's a config value.

Code: Select all

[EvenMatchOmni.MutTeamBalance]
bIgnoreMapSpecificPPH=True
change to false to include map PPH
User avatar
Enyo
Posts: 1626
Joined: Mon Apr 05, 2021 11:27 pm
Server Sponsor: Yes
Server Admin: Yes

Re: EvenMatchOmni

Post by Enyo »

pooty wrote: Mon Oct 24, 2022 10:47 pm Looks pretty decent.

I think we should go back to map based.
:lol: :lol: :lol:
Looks pretty decent???

There's a whole lot wrong with those numbers. First of all, where are they coming from? Why is leon at 691 and me at 518? Let's just start with where the Hell does the balancer get these goofy numbers?

And Staff Seargent has a higher PPH than you? In what universe? You have a multiplier and it's still lower??? Something is amiss with the numbers it's storing and using. These aren't remotely close to real PPHs. Seems like it's using some short term calculated PPH rather than overall player PPH.

Before knee jerk turning map specific back on, maybe figure out exactly how these PPH scores are being calculated by the balancer first?
“Never argue with stupid people, they will drag you down to their level and then beat you with experience.”
― Mark Twain
User avatar
pooty
Posts: 4358
Joined: Sat Apr 03, 2021 10:22 am
Location: Michigan
Server Sponsor: Yes
Server Admin: Yes

Re: EvenMatchOmni

Post by pooty »

Code: Select all

EvenMatchDebug: Assigning AnKeeDo (390.85 PPH) and EbNo (363.54 PPH)
EvenMatchDebug: AnKeeDo will be on blue (now 1319.99 PPH), EbNo will be on red (now 1463.00 PPH)
EvenMatchDebug: Assigning Staff_Sergeant (300.51 PPH) and p)o(Oty (275.30 PPH)
Other than perhaps swapping Ankeedo and Me, the rest is pretty decent.
Before knee jerk turning map specific back on, maybe figure out exactly how these PPH scores are being calculated by the balancer first?
I'll post it up later, its past PPH plus recent PPH, subject to multipliers. I think the recent might be throwing it off (and specing skews it as it counts the time if you join, so if you spec for 5 minutes, then play a minute it divides your points by 6 not 1). Recent can have much fluctuation and perhaps that should go and just use historical values-- meaning sure some players have an occasional great game, but we want to balance by average not by outliers.

Maps based is not knee jerk. Why is Ankeedo's numbers so high? Because he racks up huge PPH playing Megas/Badgers all day -- which doesn't translate to other maps, and some players do better on certain maps so I think that's valid.
User avatar
pooty
Posts: 4358
Joined: Sat Apr 03, 2021 10:22 am
Location: Michigan
Server Sponsor: Yes
Server Admin: Yes

Re: EvenMatchOmni

Post by pooty »

Snarf, is the latest in Git? I can't find bIgnoreMapSpecificPPH
User avatar
Enyo
Posts: 1626
Joined: Mon Apr 05, 2021 11:27 pm
Server Sponsor: Yes
Server Admin: Yes

Re: EvenMatchOmni

Post by Enyo »

pooty wrote: Tue Oct 25, 2022 11:25 am I'll post it up later, its past PPH plus recent PPH, subject to multipliers. I think the recent might be throwing it off (and specing skews it as it counts the time if you join, so if you spec for 5 minutes, then play a minute it divides your points by 6 not 1). Recent can have much fluctuation and perhaps that should go and just use historical values-- meaning sure some players have an occasional great game, but we want to balance by average not by outliers.

Maps based is not knee jerk. Why is Ankeedo's numbers so high? Because he racks up huge PPH playing Megas/Badgers all day -- which doesn't translate to other maps, and some players do better on certain maps so I think that's valid.
There's a recent PPH being added with past... huh? Why would it use anything other than just average PPH? If it's something that can be easily changed, I'd say do that before turning back on map specific.And how is it calculating past PPH? Is it calculated over a long-term, the past month, past week?

I said it was knee jerk to turn map specific back on because there's something wonky going on with the numbers it's using for PPH even without map specific. And changing more than one variable at a time means you don't know which one has a positive or negative effect.

And for guys like Ankeedo, wouldn't a "divider" on players like him help with that? Is that in effect? Him being scored above you for balance is a great point of how flawed the balancer is because it can't account for someone like him. The two bad matchups of where it put Ankeedo and Staff surely throws off the math for the rest of the balance, completely changing the outcome. Maybe give Ankeedo a 0.5 multiplier... but why in the Hell is Staff scored higher than you??? That should never happen.

Bad players are generally bad on all maps, and good players are generally good on all maps, therefore map specific is unneeded. The only time good players struggle is when they're on a team full of stains.
“Never argue with stupid people, they will drag you down to their level and then beat you with experience.”
― Mark Twain
Post Reply