Please upgrade your browser for the best possible experience.

Chrome Firefox Internet Explorer
×

***Official PvE Progression V2***

STAR WARS: The Old Republic > English > Server Forums > The Ebon Hawk
***Official PvE Progression V2***

Bombbuster's Avatar


Bombbuster
06.22.2013 , 08:27 AM | #461
Quote: Originally Posted by KeyboardNinja View Post
Coming up with a fair numerical ranking for the bosses in the instances is a project that I will be tackling next week.
I'd welcome input on that. My only question would be, how are you going to assign a mathematical value to boss's difficulty if you haven't even had pulls on that boss?

Difficulty is a very subjective opinion, and that opinion changes with time. If you had asked last week how difficult OP-IX NiM was I would have said very difficult because it was our first week on him and it took some 20 pulls to kill it. But if you had asked this week I would have said just something like, 'tough but doable' because we one-shot it. How do you assign a value to that?
Quote: Originally Posted by KeyboardNinja View Post
With all that said, we will be skipping ahead each week to farm gear from the later bosses after we have deemed that we're done pulling DG for a particular lockout period. None of us like this idea, but since every guild on the server is doing it, we're in a bit of an involuntary arms race as a result.
"every guild" is a bit of an exaggeration, only 4 guilds have even killed the first boss to be able to 'skip' anything, and only one guild has killed OP-IX that I'm aware of. I think guilds assumed that just because DnT steamrolled the last three bosses that we would all be able to do the same, but are now finding out all of the TFB NiM content's difficulty has been significantly increased over the PTS versions.
Mal: Appears we got here just in the nick of time. What does that make us?
ZoŽ: Big damn heroes, sir.
Mal: Ain't we just.

KeyboardNinja's Avatar


KeyboardNinja
06.22.2013 , 08:47 PM | #462
Proposed Point System

Quote:
Rank bosses from 1 - 10 in difficulty (10 being NiM DG; 1 being HM Writhing Horror)

Bonus for clearing all bosses == Sum_bosses difficulty * 3
Bonus for clear cardinality == 10 - card

Instance bonus compensates for 3 ranks of deficit on all bosses. If three other guilds clear 6/7 on the instance but can't clear the final boss, your guild could still come out on top if you're first to 7/7. If four clear 6/7 and you're fifth on the first few bosses, it will depend on the difficulty of the final boss. If the final boss is hard enough, you could still come out on top, but it's not assured.

Boss ranks cannot be changed after they have been cleared by *any* guild. They can only be changed pre-clear with 3 votes from guilds which have pulled the boss.
So what does this mean. Well, let's imagine that we give the following boss ranks to the (unbugged) bosses in NiM TfB:
  • Writhing Horror: 3
  • Dread Guard: 10
  • Operator IX: 5
  • Kephess: 2
  • Terror From Beyond (unbugged): 6

Now let's imagine that UWA clears the entire instance *except* for Dread Guard, and they do it server first. Their score would be (3 + 5 + 2 + 6) * 10 = 160. Let's further imagine that Anger Management comes in later and clears *everything*, including Dread Guard. Their score would be (3 + 5 + 2 + 6) * 9 + 10 * 10 + (3 + 5 + 2 + 6 + 10) * 3 = 322. Thus, the points from clearing the other bosses second, then the points from clearing Dread Guard first, and finally the instance bonus for clearing the entire instance first. The scale is tuned to reward the situation where a really good guild came late to an instance where everyone else is 4/5 and then clears 5/5.

We can pick scores for bosses based on convention (Bomb, feel free to use your judgement here). Guilds that pull these bosses and see how hard they really are should have the ability to vote on a new score. Once a boss has been cleared, the score is FINAL and can no longer be changed.

Initial boss scores can be assigned based on guild knowledge. We all have a pretty good idea of what bosses are going to be harder than other bosses in (e.g.) NiM S&V. This can be a starting point. Once guilds have pulled the bosses a few times, we can vote on new scores to refine the ranking until the boss is cleared. If a boss is cleared on the first night that it's pulled, but its score is universally deemed to be insufficient, multiple guilds corroborating the difficulty could push for a change in the score. At that point, it's Bombbuster's choice. The main thing is to ensure that we don't change a score after it has been tabulated into a guild ranking. Once the boss score has been added to any guild's score, it cannot be changed. This prevents gaming the system.

Note that this scale is designed to allow clearing of bosses in a non-linear fashion. There is no particularly strong reward for skipping ahead and clearing a later boss, since they have no special bonuses. We would use a slightly revised scale for an instance which enforces linear progression (e.g. the next hard mode). Also note that this scale only rewards cardinality for at most 10 guilds.

There are a couple disadvantages here. First, and most notably, it would be possible to rack up points by clearing multiple easy bosses, exceeding the value of a single hard boss. This is annoying, to be sure, but also a problem that corrects itself since the cardinality scalar is multiplicative (meaning that you get proportionally more points for clearing a hard boss first and the easy bosses second, rather than the easy bosses first and the hard boss second). It just means that intermediate rankings may be a bit odd. The other notable disadvantage is that there is no penalty built into the system for skipping forward and farming later bosses. In fact, the system encourages guilds to do this, since you can get points. We can impose an artificial restriction that time-of-kill information on later bosses will not be considered until an earlier boss is cleared, or we can simply allow skipping. Dealer's choice.

The main mathematical disadvantage of this framework is the linear ranking scalar causes a "compression" of the top-end of the scale. What do I mean by this? The scalar for clearing a boss first is 10, and second is 9. Thus, a guild clearing a boss first will receive 11% more points than a guild clearing a boss second. The scalar for clearing a boss third is 8, which means a guild clearing a boss second will receive 12.5% more points than a guild clearing a boss third. See where I'm going with this?

The scale is inversely quadratic as it approaches the maximum. In practice, this means that the top-tier guilds will have comparatively close points, while guilds which cleared bosses a bit later will have dramatically lower scores and be almost mathematically incapable of "catching up". Now, this may actually line up with how progression is informally tallied. I'm not sure. This is where the community needs to weigh in. It is possible to fix this mathematical flaw and make all progression ranks linearly distributed from the others, but it makes the expressions more complicated and can lead to non-integral scores (which makes Bomb's life a bit harder).

Disclaimer: I haven't done a game-theoretic analysis of this system, so it may encourage some undesirable Nash Equilibria. We have a small enough raiding community that this is unlikely to be the case, but I thought I would bring it up. If someone else wants to do this analysis, that would be great. Otherwise, if the community likes, I can ensure that the system has been designed to encourage sane behavior among selfish players, but it's a tedious analysis and I didn't feel like doing it today. :-)

Thoughts?

(side note: 16 mans and 8 mans will be given the same difficulty ranking by default, unless there's a boss which has a radically different difficulty on one mode vs another)
Computer Programmer. Theory Crafter. Streaming Dilettante on The Ebon Hawk.
Tam (shadow tank) Tov-ren (commando healer) Aveo (retired sentinel) Nimri (ruffian scoundrel)
Averith (marksman sniper) Alish (lightning sorcerer) Aresham (vengeance jugg) Effek (pyro pt)

Bombbuster's Avatar


Bombbuster
06.23.2013 , 08:22 AM | #463
Quote: Originally Posted by KeyboardNinja View Post
  • Writhing Horror: 3
  • Dread Guard: 10
  • Operator IX: 5
  • Kephess: 2
  • Terror From Beyond (unbugged): 6
This illustrates why assigning numerical values to bosses you haven't personally experienced is so problematic, and why I am reluctant to adopt such a system.

You have OP-IX as less than twice as difficult as TWH, but it is easily many times more difficult. Likewise you have Keph as less difficult as TWH! We one-shot TWH the very first time we pulled it, with no deaths, but we've had some 20+ pulls on OP-IX with 3% being our best attempt. Only someone who has never had multiple pulls on OP-IX would rank the very easy TWH above OP-IX.

I guess what I'm saying is a mathematical ranking system is certainly feasible, but the values would have to be assigned by, say, the first 5 world guilds to clear the instance in order to be relatively accurate. I like the idea and the formula, but it only works with correct values assigned to bosses, and those of us who haven't experienced the content are unqualified to do that.
Mal: Appears we got here just in the nick of time. What does that make us?
ZoŽ: Big damn heroes, sir.
Mal: Ain't we just.

KeyboardNinja's Avatar


KeyboardNinja
06.24.2013 , 10:32 AM | #464
Quote: Originally Posted by Bombbuster View Post
This illustrates why assigning numerical values to bosses you haven't personally experienced is so problematic, and why I am reluctant to adopt such a system.

You have OP-IX as less than twice as difficult as TWH, but it is easily many times more difficult. Likewise you have Keph as less difficult as TWH! We one-shot TWH the very first time we pulled it, with no deaths, but we've had some 20+ pulls on OP-IX with 3% being our best attempt. Only someone who has never had multiple pulls on OP-IX would rank the very easy TWH above OP-IX.

I guess what I'm saying is a mathematical ranking system is certainly feasible, but the values would have to be assigned by, say, the first 5 world guilds to clear the instance in order to be relatively accurate. I like the idea and the formula, but it only works with correct values assigned to bosses, and those of us who haven't experienced the content are unqualified to do that.
This is why boss weights can be changed by guilds who have pulled the boss. Do you really want a scale which provides the same reward for clearing WH as for clearing the DG? I certainly don't, and your post sounds like you don't either. Having some sort of difficulty weighting is the only way to represent that problem, though the exact way in which that rating is assigned can be tuned.

For example, we can rank bosses from 1 to 10 such that each rank is twice as difficult as the previous one (exponential rather than linear). I thought this was a somewhat unnecessary complexity, but it would be an option.
Computer Programmer. Theory Crafter. Streaming Dilettante on The Ebon Hawk.
Tam (shadow tank) Tov-ren (commando healer) Aveo (retired sentinel) Nimri (ruffian scoundrel)
Averith (marksman sniper) Alish (lightning sorcerer) Aresham (vengeance jugg) Effek (pyro pt)

Ordo's Avatar


Ordo
06.24.2013 , 10:54 AM | #465
I am wondering though Tam. . . . you have UWA and AM mentioned there. . . . where's Aisthesis in that hypothetical gedanken?
It ain't about how hard ya hit. It's about how hard you can get hit and keep moving forward. How much you can take and keep moving forward. That's how winning is done!

MrOscarMonster's Avatar


MrOscarMonster
06.24.2013 , 07:32 PM | #466
In my humble opinion, perhaps it would be best to avoid a numerical value assignment to each boss and simply leave ranking as they are right now, and only allow ranking changes once a guild clears the entire raid. This circumvents disagreement over boss values, skipping DG, etc. For instance, suppose we currently (Pre 2.0) have:

1. <Yoda Soda Club>
2. <Wookie Cookie Bakers>
3. <Han Solo Burger Truck>
4. <Ewok BBQ Pit>
5. <Deep Fried Mon Calamari>

Suppose <Ewok BBQ Pit> gets server first clear of ALL of NiM TfB and <Wookie Cookie Bakers> gets the 2nd on server. While <Han Solo Burger Truck> might be further in the NiM TfB raid then <Yoda Soda Club>, that ranking won't actually shift until they clear all of NiM TfB. In this example, rankings would adjust (movement highlighted in green) to:

1. <Ewok BBQ Pit> (Server First!!!)
2. <Wookie Cookie Bakers>

3. <Yoda Soda Club>
4. <Han Solo Burger Truck>
5. <Deep Fried Mon Calamari>

Essentially, ranking is based on time of full TfB clear. No ranking advantages from skipping bosses or cherry picking which boss to focus on to garner a higher point score (If one were completely ranking obsessed, under you proposed number system, the quick and dirty way to jump rank would be to skip both DG and Op9 for an easier 11 points while other guilds struggle against Op9). In addition, this preserves some style of ranking system for guilds that aren't doing NiM TfB at all, yet still actively raiding and progressing at their own rate.

Just my thoughts on how we might approach ranking. Just out of curiosity, what are y'all's thoughts on nerf to NiM TfB?

And obligatory Rakata of Niceness:
Kleric

Apfelschorle - Envii - Amandeur - Artifice
<Sanctuary>

Ordo's Avatar


Ordo
06.24.2013 , 09:33 PM | #467
Quote: Originally Posted by MrOscarMonster View Post
In my humble opinion, perhaps it would be best to avoid a numerical value assignment to each boss and simply leave ranking as they are right now, and only allow ranking changes once a guild clears the entire raid. This circumvents disagreement over boss values, skipping DG, etc. For instance, suppose we currently (Pre 2.0) have:

1. <Yoda Soda Club>
2. <Wookie Cookie Bakers>
3. <Han Solo Burger Truck>
4. <Ewok BBQ Pit>
5. <Deep Fried Mon Calamari>

Suppose <Ewok BBQ Pit> gets server first clear of ALL of NiM TfB and <Wookie Cookie Bakers> gets the 2nd on server. While <Han Solo Burger Truck> might be further in the NiM TfB raid then <Yoda Soda Club>, that ranking won't actually shift until they clear all of NiM TfB. In this example, rankings would adjust (movement highlighted in green) to:

1. <Ewok BBQ Pit> (Server First!!!)
2. <Wookie Cookie Bakers>

3. <Yoda Soda Club>
4. <Han Solo Burger Truck>
5. <Deep Fried Mon Calamari>

Essentially, ranking is based on time of full TfB clear. No ranking advantages from skipping bosses or cherry picking which boss to focus on to garner a higher point score (If one were completely ranking obsessed, under you proposed number system, the quick and dirty way to jump rank would be to skip both DG and Op9 for an easier 11 points while other guilds struggle against Op9). In addition, this preserves some style of ranking system for guilds that aren't doing NiM TfB at all, yet still actively raiding and progressing at their own rate.

Just my thoughts on how we might approach ranking. Just out of curiosity, what are y'all's thoughts on nerf to NiM TfB?

And obligatory Rakata of Niceness:
Interesting thought. As an extension, what if we didn't worry about rankings for this boss or that boss and instead had some sort of marker per-boss?

So, lets say that Yoda Soda Club has cleared all the bosses, wookie cookie bakers has only cleared the first boss and the last three, Han Solo Burger Truck has only beaten the first boss, but had the server first to do so. Ewok BBQ Pit has cleared the first boss, but is on a different faction from the others (Faction first) And Deep Fried Mon Calamari has only cleared the first boss. It'd look something like this:
*#* = server first
~#~ = faction first but not server
Green = go
Red = stop

1. <Yoda Soda Club> 1 - *2* - *3* - *4* - *5*
2. <Han Solo Burger Truck> *1* - 2 - 3 - 4 - 5
3. <Ewok BBQ Pit> ~1~ - 2 - 3 - 4 - 5
4. <Wookie Cookie Bakers> 1 - 2 - 3 - 4 - 5
5. <Deep Fried Mon Calamari> 1 - 2 - 3 - 4 - 5

It'd take a bit to set it all up. But once it is set up it wouldn't be that hard to maintain such a list.

Also, I love your made up guild names. They are awesome.

I'll see your and raise you a
It ain't about how hard ya hit. It's about how hard you can get hit and keep moving forward. How much you can take and keep moving forward. That's how winning is done!

Hotcakes's Avatar


Hotcakes
06.26.2013 , 10:08 AM | #468
After taking 2 weeks off and finding a new person for our progression team, Inconceivable is finally 5/7 S&V HM. ^^ Here's hoping the second night goes well.
Inconceivable! Spaceballs-the Guild!
*~ Ebon Hawk Server ~*
Anaid - Lvl 55 Sentinel Kaleena - Lvl 55 Scoundrel Healer
Jurie
- Lvl 55 Sorcerer Healer
SŰnia - Lvl 55 Shadow Tank

Bombbuster's Avatar


Bombbuster
06.26.2013 , 11:45 AM | #469
Quote: Originally Posted by Hotcakes View Post
After taking 2 weeks off and finding a new person for our progression team, Inconceivable is finally 5/7 S&V HM. ^^ Here's hoping the second night goes well.
Gratz and GL.
Quote: Originally Posted by Ordo View Post
Interesting thought. As an extension, what if we didn't worry about rankings for this boss or that boss and instead had some sort of marker per-boss?
Would something like this work?:

(Second and third posts down on this thread) http://www.swtor.com/community/showthread.php?t=635285

I like the idea, but at the end of the day we still have to come up with a progression ranking system that determines which order the progression guilds will be listed on whatever type of list we end up using.
Mal: Appears we got here just in the nick of time. What does that make us?
ZoŽ: Big damn heroes, sir.
Mal: Ain't we just.

KeyboardNinja's Avatar


KeyboardNinja
06.26.2013 , 01:04 PM | #470
Quote: Originally Posted by Bombbuster View Post
I like the idea, but at the end of the day we still have to come up with a progression ranking system that determines which order the progression guilds will be listed on whatever type of list we end up using.
And that's definitely the concern that I'm trying to address. Fundamentally, we need a linear system. That means assigning value to one achievement relative to another achievement (e.g. defeating NiM TfB vs NiM S&V). We don't want a ranking system that equally rewards clearing Writhing Horror and Dread Guard (since they are an order of magnitude separated in difficulty), which leaves us with the need to assign point values.

Leaving the system "as is", where it simply rewards having cleared more bosses highly incentivizes non-linear boss farming. Additionally, it provides the same reward (1 point) for clearing Nightmare Dread Guard as for clearing any other boss. I am strongly, strongly opposed to any system which does not address this issue.
Computer Programmer. Theory Crafter. Streaming Dilettante on The Ebon Hawk.
Tam (shadow tank) Tov-ren (commando healer) Aveo (retired sentinel) Nimri (ruffian scoundrel)
Averith (marksman sniper) Alish (lightning sorcerer) Aresham (vengeance jugg) Effek (pyro pt)