Re: Data Mining the Killboard Scenarios | Jan 2023
Posted: Mon Jan 09, 2023 1:41 pm
by Sinisterror
inoeth wrote: Mon Jan 09, 2023 1:31 pm
just remove rampage, problem solved
Just change Rampage to what it was live which is NOT ignoring Guard Mechanic and Morale Avoidances. Slayer is the core dps on order like it was live but here aoe cap is 24 instead of 9 so these 3 things are why it is insanely OP here.
Re: Data Mining the Killboard Scenarios | Jan 2023
Posted: Mon Jan 09, 2023 1:42 pm
by Morradin
rejndjer wrote: Mon Jan 09, 2023 12:25 pm
there's bunch of IC sieges because when destro attack altdorf there's a clear defender advantage during stage 1 for order. because of that funnel at that square, dont remember BO name. funny nobody mentioned this.
P.S. also most of order forts are taken when there's nobody online. maybe make forts not count for city score if less than 500 players on server are online or something like that?
You have the same funnel in IC. All in game doorways are the same size, IC and altdorf.
And I like your idea about using pop to figure out fort loss/win. Instead of players though, it should use percentages, this way as our population gets lower we still have endgame. Say, if a fort has a greater than 15%, 20% or 25% difference in numbers between attacker than defender, then it does not count for realm victory points.
Re: Data Mining the Killboard Scenarios | Jan 2023
rejndjer wrote: Mon Jan 09, 2023 12:25 pm
there's bunch of IC sieges because when destro attack altdorf there's a clear defender advantage during stage 1 for order. because of that funnel at that square, dont remember BO name. funny nobody mentioned this.
P.S. also most of order forts are taken when there's nobody online. maybe make forts not count for city score if less than 500 players on server are online or something like that?
You have the same funnel in IC. All in game doorways are the same size, IC and altdorf.
And I like your idea about using pop to figure out fort loss/win. Instead of players though, it should use percentages, this way as our population gets lower we still have endgame. Say, if a fort has a greater than 15%, 20% or 25% difference in numbers between attacker than defender, then it does not count for realm victory points.
i haven't measured it, but in IC chokepoint BO, there's more space to move around. also there's no height difference like there is on altdorf choke BO.
Re: Data Mining the Killboard Scenarios | Jan 2023
Posted: Mon Jan 09, 2023 2:05 pm
by agemennon675
FFNation wrote: Sun Jan 08, 2023 9:43 pm
This is really cool, thanks for doing this!
My takeaway is. Looks like damage is much more important for winning than healing. Those damage dealers have high 90% win rates (only 1 out of 10 is a healer and 9 out of 10 are order).
While the top healers have much lower win rates.
Or maybe we could conclude Melee is OP, specifically order melee.
I'm new to the game, but seeing 90% win rates seems way out of balance.
This also looks bad too as a new destro player.
Realm Win Rate City Sieges
Since April 2022 Altdorf has been sieged 72 times, IC has been sieged 202 times
I'm sure a lot of this comes down to player skill and looks like all the best are playing for order side.
Data looks like that because those classes are probably played only with a premade farming un-organised opposition in the scenarios / event scenario etc. Try to not take it too seriously, this game is designed that way so premades farm pugs for easy rewards but when it comes to Group ranked those premades are nowhere to be found
Re: Data Mining the Killboard Scenarios | Jan 2023
Posted: Mon Jan 09, 2023 3:51 pm
by Whyumadbro
Detangler wrote: Sun Jan 08, 2023 3:30 pm
Well no wonder I have more fun in pug t4 scenarios, its the only area that Destro actually wins more often.
Devolving into Rampage/GtdC thread by page 3. Calling it.
damn....
Re: Data Mining the Killboard Scenarios | Jan 2023
Detangler wrote: Sun Jan 08, 2023 3:30 pm
Well no wonder I have more fun in pug t4 scenarios, its the only area that Destro actually wins more often.
Devolving into Rampage/GtdC thread by page 3. Calling it.
damn....
I hit it right on the money
Re: Data Mining the Killboard Scenarios | Jan 2023
Posted: Tue Jan 10, 2023 3:21 am
by Phantasm
Spoiler:
MasterYoda wrote: Sun Jan 08, 2023 2:24 pm
Hello everybody,
Shortly after the killboard was released in April 2022 did I write a short essay about how one might be able to use the killboard statistics for insights into class and realm balance.
Yesterday evening I decided to again data mine the killboard over night and write out possible results this morning. The scraped killboard data encompasses scenario and city instance results from the 2nd of April 2022 up to the early morning of the 8th of January 2023. The scraped data is 3.2 Gigabytes in size but zipped up in a convenient 346MB file for anyone interested in checking the data or doing their own analysis.
I did not have the time or the interest in repeating a wide-ranging data analysis of all classes as I did the last time - even though class balance has not changed significantly since i last performed it - there is merely more data available now. I therefore largely reused the original code and will keep myself brief. I expect people will certainly use this data to speculate about game balance and I do think that to a degree it contains indications about such, however please feel free to check out the information that killboard data mining provides from the github repository. You'll see that this information is limited and obviously by nature does not include rvr performance data. Though I do think it is useful and interesting if taken with a grain of salt.
Realm Win Rate T4 Standard Scenarios
A 'stomp' by either realm was defined by if the opposing realm had less than 100 points and therefore received the minimal amount of crests from the scenario. The percentage behind the 'stomps' specifies the percentage of how many wins are actually stomps, e.g. 56% of all T4 standard scenarios that order wins are done so with destro players receiving the minimal amount of crests.
Furthermore, since I oftentime see people asking if scenarios are popping, here an overview of the average waiting times accumulated over the weeks, days or hours of the day, allowing among others to see if scenarios pop faster/slower than a couple of months ago or at what time they pop fastest:
Scenario Queue wait time T4
Spoiler:
Scenario Queue wait time Mid-Tier
Spoiler:
Personal Feedback to RoR
Lately RoR has become incredibly toxic and bitter - particularly on destro side where I experienced it the most since I lately play mostly destro on my YodaXYZ twinks, though I certainly saw it on order too. I don't want to propose solutions or anything the like since I am sure the core of the problem is the availability of developer hours and community manager hours - since this is a free and volunteer-run game. And I am very thankful and grateful to those volunteers providing such a fun experience and that I understand what a monumental undertaking it is to run an MMO as volunteer devs. As usual would I like to point out that I'd be happy to help if I can.
However I would also like to point out that destro global chat - and I suppose order global chat as well - is incredibly hostile and bitter and oftentimes insulting to the devs themselves or anyone trying to defend the game. Oftentimes from the same players and usually spurned on when the realm imbalance is too great - rightfully blaming the ease of xrealming, class imbalances or the possibility of completely unfair playstyles like stealth-gank parties. It degrades into enormous bitternes that I am sure is not only driving new players reading it in global chat away in droves but is just not fun to be exposed to in a game that is intended to be fun in ones own free time. Trying to provide counter points and defending the game so that new or other players don't get exposed to it and leave is only met with more 'horde mentality' - putting those people on ignore results in the same negative effect for new players or others who just want to enjoy the game. I genuinely consider this lately a huge issue driving new players away.
You may be aware of this, you may not be and I'm sure a lot of players and devs probably disagree with me. I honestly do not know. But since I rarely use the forums since I don't like the time-based response ordering instead of an upvote-based response ordering (e.g. reddit) did I just think that I'd leave my personal feedback here as well.
Again, thanks for the hard work!
Cheers!
agemennon675 wrote: Mon Jan 09, 2023 2:05 pm
Data looks like that because those classes are probably played only with a premade farming un-organised opposition in the scenarios / event scenario etc. Try to not take it too seriously, this game is designed that way so premades farm pugs for easy rewards but when it comes to Group ranked those premades are nowhere to be found
Is it possible Yoda to conclude that from gathered data? Whats the ratio of premade vs premade standard scenario and premade vs pug? That could confirm or debunk a myth that playing premade scenario queing system pair you against pugs often then vs other premade.
Re: Data Mining the Killboard Scenarios | Jan 2023
Posted: Tue Jan 10, 2023 8:54 pm
by Emissary
Detangler wrote: Sun Jan 08, 2023 3:30 pm
Well no wonder I have more fun in pug t4 scenarios, its the only area that Destro actually wins more often.
Devolving into Rampage/GtdC thread by page 3. Calling it.
Mine showed at the end of page #2 ... so close
edited to say you did say "by" page 3 so I stand corrected sir you are a God among men
Re: Data Mining the Killboard Scenarios | Jan 2023
Posted: Tue Jan 10, 2023 10:29 pm
by Paxsanarion
GrishnakBlagtoof wrote: Mon Jan 09, 2023 8:27 am
Order is just flat dominating. And years ago when it was Destro "dominating" people went to Order....and none have come back...why?
The numbers don't lie....
Typically I agree with you....however I can assure you Order is not dominating during EU prime time......it really depends on when you play more than anything else.......and who logs Order and who logs Destro at a given moment For example right now it is 530 pm NA east coast time.......and T4 is 58% Destro
Re: Data Mining the Killboard Scenarios | Jan 2023
Posted: Wed Jan 11, 2023 4:52 am
by Sofong
SLAYER SLAYER SLAYER WHITE_LION WHITE_LION WHITE_LION
great.... thats what i expected.
