 |
|
 |
|
|
Author |
Message |
Nomad_Wanderer
Joined: 04 Apr 2006
Posts: 30
|
 Posted: Tue Jul 11, 2006 1:17 pm Post subject: Chance to get it right again. |
|
|
Okay... So, I've been trolling this forum for a few months now, and trying to get DKLP, CT_RT, and EQDKP to fit to our guilds existing points system/concept. It's been a difficult process (which I have not completed), as our current system is very people and labor intensive (lots of interesting exceptions for this and that), and all tracked on pencil and paper, and then post raid put in excel.
Our guild has decided to seperate out the 20 man raids from the 40 main raids into seperate systems.
Using the three established tools above (EQDKP, DKPLP, CT_RT (freddy)) how would you setup things to be:
1. Fair
2. Concise
3. Easy
It could be Time based, event(kill) based, or a combination of the two. Zero-sum, or whatever. I would like to state that we are interested in a fixed price system.
I'm currently thinking of a fixed price, zero-sum, non-capped, allow negative, Timebased only system.
Please reply with your ideas/comments |
|
Back to top |
|
 |
Lokorin
Site Admin
Joined: 03 Apr 2006
Posts: 697
|
| Posted: Tue Jul 11, 2006 3:28 pm Post subject: Re: Chance to get it right again. |
|
|
| Nomad_Wanderer wrote: |
| I'm currently thinking of a fixed price, zero-sum, non-capped, allow negative, Timebased only system. |
I'm of course biased towards time-based and zero-sum systems, so it sounds good to me.
It's also common to reward the first (and sometimes last) reward interval with a bit of extra DKP to encourage being on time. I would suggest doing it by having the first interval's weight set to something fitting such as 2 (there's an option to do that automatically), which preserves zero-sum.
Another thing that might make it more fair is some kind of guaranteed DKP per raid (there's an option for that too) so that you don't end up with raids that give next to no reward (e.g. progression raids).
The above option along with the natural flow of DKP will cause the system gradually become inflated (might take a few years). The point of zero-sum is for the system to be balanced, e.g. have a system total of 0 (or otherwise new members keep falling farther and farther behind etc). Therefor I would also recommend regularly calculating the system total by adding together everyones DKP to see how close to 0 you are (someone ought to write a plugin for this by the way, or make a standalone version of the DKP bank - LGPLed source code). If your total start speeding away from 0 then I would suggest doing something such as taking 5% away from the DKP in from every raid to slowly bring the system back to balance (which is done automatically with the reward modifier). Just be sure to get some paragraph about this into your loot rules from the beginning so that you don't have to try to sell it in later. Another option when dealing with this problem is to separate the balancing from the DKP rewarding with software such as the balancer plugin for EQDKP, this is far less likely to be accepted by your guild members though, so doing it slow and painless could be better (although balancing is also painless, but it may appear not to be). |
|
Back to top |
|
|
Nomad_Wanderer
Joined: 04 Apr 2006
Posts: 30
|
| Posted: Tue Jul 11, 2006 4:16 pm Post subject: |
|
|
Actually, that may be possible too.
I like the ideas behind the anti-hoarding code. Another aspect that I failed to mention is that the guild's current pen and paper based system is getting to cumbersome to maintain for both the 20 and 40 man raids. Currently in wow, the 20 main raids do not contain as much upgrade loot as the 40 man raids, so they aren't as concerned with them, and with the loss of an officer or two are having trouble keeping up.
IMO, this is an opportunity to demonstrate the power, flexibility, and stability of an application backed DKP system...
Our guild as talked about caps/hoarding in the past, as well as item tier skipping (someone who has tier 1 (low) items, deliberately saving/hoarding, hoping to skip tier 2, and go right to tier 3).
So the solutions to some of these problems that you already have found may actually be accepted.
A few questions:
I like the balancer with the anti-hoarding code, but my guildmates will say the following -- "I have been running this dungeon for weeks and have 90% of the items that are important for me to get from here... Why encourage me to buy something that I may only use part-time seldomly, taking it away from another class who would use it more? That's ANTI-guild progression, as I'm taking an item I will use a little from someone who would use it more, so that I can continue to ave a decent return on my points..."
I like the idea of the Anti-hoarding code, but want to hear how it works in practice some more, particularly around multi-class items....
For a guild that would be running these 20 man instances 2-3 month, when would you reccomend running the balancer? In your example it talks about every fortnight... What type of raiding schedule was that example built on? Does it matter if the balancer is run too often? (i.e. little changes in DKP between balancer runs), or too seldom?
In a 100% time based system, how do you prevent "time bloat", i.e. with an 100% event based system, people want to get in, complete the events, and get out, asap. But in a time based system, there's actually a reward for moving more slowly. Is there a way to add a DKP bonus for completing the night in a set amount of time or less?
Also in a Zero-sum system, how do you reward learning content? I.e. lots of deaths, with possibly no purchases? Is this DKP minimums? |
|
Back to top |
|
|
Lokorin
Site Admin
Joined: 03 Apr 2006
Posts: 697
|
| Posted: Tue Jul 11, 2006 6:22 pm Post subject: |
|
|
| Nomad_Wanderer wrote: |
| A few questions. I like the balancer with the anti-hoarding code, but my guildmates will say the following -- "I have been running this dungeon for weeks and have 90% of the items that are important for me to get from here... Why encourage me to buy something that I may only use part-time seldomly, taking it away from another class who would use it more? That's ANTI-guild progression, as I'm taking an item I will use a little from someone who would use it more, so that I can continue to ave a decent return on my points..." |
I should start by saying that I have never personally used the anti-hoarding part in a live environment because I couldn't convince my guild to use it.
Buying an item results in the person having less points after the balance than if the person wouldn't have bought it, hence if the person is only interested in having a high amount of spendable DKP then the person should not but the item. What is more important is that if the person is saving up for some rare item (i.e. wants to be in the top x) then the (less) DKP earned by already having a lot of DKP will still represent the same amount of time relative everyone that is close to that person in the standings. So if one is in the top and has attended raids for one week longer than the person in second place then it will still take the second person a week to reach the first person, even if the DKP difference might be minimal.
Hence the time earned relative everyone close is never altered. If the person doesn't buy anything for a week then that person will have earned a week of time relative those who are close. The difference comes when someone high in the ranking buys an item, as a person climbs higher in the standings the actual time investment needed to buy an item increases. A person with low DKP might only need to spend DKP equal to a couple of days of raiding while a person with high DKP might have to spent a weeks worth of DKP. If the persons in the top are aiming for a rare item then the time investment spent by buying the item gradually increases until a point where it becomes so high that it is no longer interesting.
So what I'm trying to say is that as long as any item in the game is interesting to the person (who is feeling the AH effects) and that item is heavily contested then buying an uninteresting item would come at a high price as the person would be sacrificing more invested time (loosing much time in the race for heavily contested item) than a person with less DKP. However if there are items which are interesting and not heavily contested then the person would probably get more loot quality for the time invested by buying those items instead of continuing to race for the heavily contested one.
Hence the systems promotes buying items that are uncontested and interesting when possible, but also makes it possible to "race" for an item, but at a higher cost (i.e. probably only those who already got everything will be interested). Lets not forget that the DKP "lost" by the people high in the rankings is redistributed to those with less DKP (who probably need it more), hence promoting progress.
| Quote: |
For a guild that would be running these 20 man instances 2-3 month, when would you reccomend running the balancer? In your example it talks about every fortnight... What type of raiding schedule was that example built on? Does it matter if the balancer is run too often? (i.e. little changes in DKP between balancer runs), or too seldom?
|
The only negative part about running it too often would be lots of individual adjustments. I have not reviewed the math recently, but it feels like the more often the program is run the greater the resolution and the less chance there is for some quirk. The algorithm makes a linear approximation based on the current DKP so running it too seldom could create situation when one gets more DKP after the balance by buying an item than if one would not have bought it. It would require that much DKP has entered the system since the last run combined with the person being at the inflection point of the penalty curve. I have not tried to calculate situations when that could occur, but a way to make sure that it never happens is to run the balancer so often that the maximum DKP earned during the time period is less than the lowest price for an item dropped during the day (as that would mean that one ends up at less DKP even if 100% was "lost" before buying the item and 0% afterwards). I would guess that a fortnight satisfies that (the balancer was built in a guild that raided approximately 4*6 hours a week).
| Quote: |
| Finally, in a 100% time based system, how do you prevent "time bloat", i.e. with an 100% event based system, people want to get in, complete the events, and get out, asap. But in a time based system, there's actually a reward for moving more slowly. Is there a way to add a DKP bonus for completing the night in a set amount of time or less? |
With zero-sum you can never earn more from a night by staying longer, you can only earn DKP equal to the distributed value of the loot dropped (assuming that you don't get more loot by staying longer, but I'm assuming you meant a constant amount of loot).
*edit*
Missed one question:
| Quote: |
| Also in a Zero-sum system, how do you reward learning content? I.e. lots of deaths, with possibly no purchases? Is this DKP minimums? |
Yes, this is DKP minimums or bonuses. |
|
Back to top |
|
|
Nomad_Wanderer
Joined: 04 Apr 2006
Posts: 30
|
| Posted: Tue Jul 11, 2006 6:42 pm Post subject: |
|
|
Thanks.. as always, you have some great insight into this stuff.
Zero sum Time based solves alot of problems.... I think that's the real winner.
As you can see from my "time bloat" comment, I'm still getting my head around it 
Thanks again! |
|
Back to top |
|
|
Nomad_Wanderer
Joined: 04 Apr 2006
Posts: 30
|
| Posted: Tue Jul 11, 2006 8:21 pm Post subject: |
|
|
When talking about Time based Plus Zero sum, how does DKP_LP handle this?
For example, does the zero sum calculation happen in realtime, with realtime stats, or does it use aggregated post raid stats?
I.e. 8 items at 20 a peice, Total loot dropped equals 160 DKP. There were a total of 44 people in the raid, but never more than 40 at a time, (4 people eleft and 4 different people joined).
is it 20 / 40, and the people present at that interval get it, and the next one happens to have 2 drops, 40/ 40, and the people present get it.. etc.
is it 160 / 44? and then divided into the number of intervals, and awarded to those present at those intervals?
The value of a given raid will be determined via Zero sum. You will recieve a percentage of that value based on the time you spent with the raid. If you spent 10% of the time in the raid, you will recieve 10% of the points. If you spent 100%, you will recieve a "Full share" of the points, and also will recieve bonuses(below) |
|
Back to top |
|
|
Lokorin
Site Admin
Joined: 03 Apr 2006
Posts: 697
|
| Posted: Tue Jul 11, 2006 9:09 pm Post subject: |
|
|
It will be 160/44 divided by the number of reward intervals if the zero-sum distribution mode is set to per interval. If it set to per minute then it will be 160/44 divided by the number of started minutes and then distributed by each minute participated in a reward interval. A more detailed explaination of the algorithm is also available. |
|
Back to top |
|
|
|
|
