Mathematik und Collect Missions
Verfasst: Mo 27 Okt, 2003 10:56Mathematischer Beweis dafuer, dass Collect Missions einfach nur Glueckssache sind:
( http://pub36.ezboard.com/ffreenorrath29 ... ID=1.topic )
There's a post on the druid's grove LDoN board that describes the fact that some dungeons contain more mobs than others. This post expands on that in some detail to explain why, for example, collects are shunned.
First, success or failure often is NOT determined by skill though it obviously has an impact in some cases. DPS does matter a LOT in some dungeons and little in others, as should be clear after the following analysis.
I'll start with collect missions because they are the easiest to demonstrate mathematically that dungeon size matters dramatically. People have hypothesized that "collect size" (the number of items you need to collect) determines the difficulty of the mission. This does not appear to be the case (though see the conclusion).
As anyone who has played many dungeons knows, they vary a lot in layout and size. It turns out, based on logging by people on the Druid's Grove, that it's not just the layout by the sheer number of mobs varies (Panamah, for example, uses EQWatcher to keep track of these things; I also use EQWatcher and hear how many kills we've done per hour).
Dungeon Size
Some dungeons have ~90 mobs, some have more like ~150. You don't know which you have until you zone in. This explains a lot, for me. There are some dungeons that my group can completely clear, every mob dead (we can't quite do that in 90 minutes, but we can within the 2 hours). Other dungeons, we can't clear more than about 80% of the dungeon. This is with the same group and the same number of kills per hour (we get between 50 and 55 kills per hour, depending on factors like whether someone goes LD, needs to go afk, etc.).
Of course we had the subjective impression of "wow, we must have really kicked butt" for the dungeons we cleared, but it turns out that our skills, tactics, etc. had little or nothing to do with that - dungeon size/population was the key factor.
We took a collect mission the other night where we needed 28 items. Not only did we succede, we cleared the dungeon (I believe we found all 42 that the dungeon was seeded with, actually). We've had other missions with 30+ items that we didn't win (actually we've won them ALL in "overtime").
Collection Missions
Here's the simple math. Say your collect is for 20 items. They seed, according to "official word" (apparently) and people logging, 1.5 times the collect items into the dungeon. So, in this case, 1.5*20 or 30 items into the dungeon.
Case A. Small Dungeon, Small Collect. The dungeon has 90 mobs. This means that 30/90 or ONE THIRD of the mobs have the item you need. ON AVERAGE, you will need to kill 60 ((90*20)/30) mobs to collect 20 items. You can still get unlucky and have to kill every mob (and, apparently, containers), or you can get REALLY lucky and kill only 20. Best case 20, worst cast 90. On average, 60.
Case B. The dungeon has 150 mobs. This means that ONE FIFTH (30/150) of the mobs will have the item you need. ON AVERAGE, you will need to kill 100 mobs to collect 20 items ((150*20)/30). You can get unlucky and need to kill 150 or you might get lucky and need to kill only 20. Best case 20, worst case 150. On average, 100.
In other words, if you get a big dungeon in a theme, you are going to have trouble doing a collect mission - if you can kill 100 mobs in 90 minutes, you have a good chance at winning, but if you can't kill 150 in 90 minutes, you still have a chance of losing. If you can't kill 100 mobs in 90 minutes (groups without really high DPS), then you are going to lose most of the time. My group can kill more like 80 maximum, for example.
Collect Size
Does collect size matter, as people have guessed? For case C & D, make the number of items 30 instead 20. In this case, 45 are seeded into the dungeon.
Case C. 90 mobs. ONE HALF of the mobs will have the item. This means you need to kill 60 mobs on average to win (every other mob, on average, will have the item you need). That is, (90*30)/45. Best case 30, worst case 90, average 60.
Case D. 150 mobs. Almost ONE THIRD of the mobs will have the item (30% of them). This means you need to kill 100 mobs, on average, to win. That is, (150*30)/45. Best case 30, worst case 150, average 100.
Summary of Collect Missions
Case A: Small Dungeon, 20 Collect. 60 mobs needed.
Case B: Large Dungeon, 20 Collect. 100 mobs needed.
Case C: Small Dungeon, 30 Collect. 60 mobs needed.
Case D: Large Dungeon, 20 Collect. 100 mobs needed.
In a way, SOE did the math right in terms of balancing the number of collect items you have to find. What they didn't balance was the "difficulty" of large dungeons and small dungeons. A group with high DPS that can kill 100-120 mobs in 90 minutes won't have a lot of trouble BUT they still can get unlucky despite their very high DPS and/or skills.
A group with lower DPS that kills 50 mobs per hour is almost always going to lose the B & D (large dungeon) scenarios. They just can't kill the average number of mobs needed. They can still get lucky and get all the pieces within the ~75 mobs they DO kill, but it's unlikely.
Rescue Missions
This one isn't AS mathematical because the layout and location fo the target matters more than the sum total of the mobs. On average, though, the more mobs there are in a dungeon, the more you will need to clear to get to your target. If, on average, you need to clear 70% of the mobs in a dungeon, you're looking at clearing 68 mobs versus 105 mobs. However, layout of the dungeon impacts this by changing the percentage of mobs that need to be cleared. If the dungeon is very linear with few side branches, you may need to clear 90% of the mobs instead; if the dungeon has many side branches, you may need to clear only 50% (or whatever).
Size of the dungeon still has an impact - there are small linear dungeons and large linear dungeons, for example. A small linear dungeon could be harder to complete than a large non-linear one.
Assassinate Missions
This is very similar to a rescue mission. Apparently during beta they were quite different because you could invis or pacify your way to the boss, thus circumventing the number of mobs in the dungeon. Now, however, you have to kill some percentage (I believe it's 60%) of the mobs to SPAWN the boss mob.
If there are a large number of mobs, it takes longer to spawn the boss mob. If there are 90 mobs, you need to kill only 54 mobs to spawn the boss. If there are 150 mobs, you need to kill 90. This has lead to problems for some people - in a large dungeon you can reach a "split" in the layout of the dungeon BEFORE you've killed enough mobs to spawn the mob.
Of course you've only SPAWNED the boss at that point - you still need to clear the rest of the way in some cases, which will TEND to take more time in a large dungeon. As with rescues, the layout of the dungeon matters - with a linear dungeon, you may still need to kill practically every mob.
Both rescues and assassinates add the difficulty of finding the target, though a tracker + a map can mitigate that difficulty (it can still be tough sometimes to determine where the target is, just depending on the layout of a dungeon; however, a tracker and a map allow you to correct your errors fairly quickly).
Slaughter
These, the highly preferred mission type, differ from those above in a simple detail: they have no relation to the size of the dungeon. If you need to kill 58 mobs, it doesn't matter whether there are 90 total or 150 total. (The layout can still matter some - the time it takes to move and/or pull can factor in; long empty corridors may slow down the kill rate; large rooms that "require" pacify or crowd control my compound that.)
Also, these missions have little or no luck involved. There are no "wrong turns." There is no worry that you'll get unlucky and just not find enough mobs, etc.
Conclusion
Skill and, once you've reached some minimum of skill, DPS matter but it's RANDOM how much it will matter. My group, for example, with it's generally low DPS can do just fine in Slaughter missions, Collect missions in small dungeons and Rescue missions in small, non-linear dungeons. We don't do well in large dungeons simply because our DPS is on the low end (but probably not on the low end for casual players).
The fundamental problems are:
1. For collect missions, the seeding of items must take into account the number of mobs not just the number of collect items. Basically, they need to scale the number of items so that you need to kill, on average, 60 mobs to win. People will STILL be more likely to lose large dungeons (many can't clear a large dungeon but CAN clear a small dungeon), but the risk is reduced.
2. Extra reward needs to be provided for Rescue and Assassinate missions. There's no good way to control for dungeon layout and dungeon size in this case, so increase the "reward" in the risk/reward ratio. Make these missions yield 25% (or whatever) more AP's on a win or win in overtime. These require killing AT LEAST as many mobs as the slaughter mission IN ADDITION to either leading someone out or killing a boss mob.
3. Slaughter mission difficulty is about right IF reward is boosted for rescue and assassinate. Many groups are still going to have trouble with these - they are a guaranteed win for SOME groups since it's just a matter of DPS, not doing anything stupid, and not having any "outside" factors occurring (e.g., Linkdeaths).
The bottom line is that if you care about APs per unit time, you're crazy to take anything other than Slaughter missions. Who likes to have their success or failure determined by a flip of the coin? Currently the coin flip that determines which dungeon you get will determine, for many people, whether you win or lose. Some people are beginning to use the "leave adventure" function when they see that they are in an "unwinnable" dungeon (i.e., one that is going to require more kills than their group can produce).
Challenge is good and I will say that LDoN has challenged my group and made it MUCH more fun to play the game. Dealing with frequent multiple pulls, managing mana over the course of an adventure, etc. is fun. If we screw up, that's fine and good - we actually enjoy the post game analysis of figuring out what we did wrong and how we can improve in the future.
What is NOT good and fun is having the RNG determine whether you're going to lose, despite using the best tactics you can use. There are many times where we can look at a map 20 minutes into the adventure and KNOW whether we're going to lose. Doing everything right and slaying as quickly as possible and then realizing that you can't possibly get to where you need to be in the alotted time is frustrating. No, the time isn't completely wasted - you still got experience and loot, but that is beside the point.
Frustration involves some factor getting between you and your GOAL. Having that be a random factor is bad. This is the reason that many of us refuse to do long camps for an item - despite the fact that you might be getting good experience and loot while doing the camp. LDoN amounts to a long camp even if you succede every time, but it's finite and controllable, making it more pallatable to the casual player. Haivng the dungeon size/layout determine your success changes that though - the natural response is to minimize random chance and simply take slaughter missions.
My group WANTS to do the other mission types to keep things fun. However, the "fun" of saving that high elf woman for the 10th time wears thin. The fun of slaying Bob, who is unifying the mobs in the dungeon, tapers off pretty quickly. If the difficulty of the missions were equivalent, then you would still just take what's offered and go do it - variety at least is built in. When some missions are much more prone to failure and there's NO reward associated with the additional risk of failure, people are just going to avoid those missions. Either reduce the risk (e.g., by seeding more items based on the number of mobs) or increase the reward (e.g., by increasing AP's for some mission types).
One other note. From my limited experience with some themes and from looking at maps of the different dungeons, it appears that some themes are worse than others in terms of the dungeon layouts. For example, Guk and Miragul both seem to tend to have very large dungeons which make anything but a slaughter mission in those themes very risky for the average player.
I was shocked to see my rating in Guk jump to about 600 upon winning ONE mission, which implies that either few people are fighting there or few people are winning there. Mistmoore, for example, isn't "better" just because of the number of undead - you seem more likely to get a "good" dungeon layout and population. We've won many rescue missions there and the few we lost are because of number of mobs we had to kill was too great for our DPS - for some themes, I'd never take a rescue mission.
Well, one other note. The analysis above assumes that there is no correlation between what type of mission you get and the number of mobs in a dungeon. It is possible that you are more likely to get small dungeons for rescues, for example - but that does not SEEM to be the case.
( http://pub36.ezboard.com/ffreenorrath29 ... ID=1.topic )
There's a post on the druid's grove LDoN board that describes the fact that some dungeons contain more mobs than others. This post expands on that in some detail to explain why, for example, collects are shunned.
First, success or failure often is NOT determined by skill though it obviously has an impact in some cases. DPS does matter a LOT in some dungeons and little in others, as should be clear after the following analysis.
I'll start with collect missions because they are the easiest to demonstrate mathematically that dungeon size matters dramatically. People have hypothesized that "collect size" (the number of items you need to collect) determines the difficulty of the mission. This does not appear to be the case (though see the conclusion).
As anyone who has played many dungeons knows, they vary a lot in layout and size. It turns out, based on logging by people on the Druid's Grove, that it's not just the layout by the sheer number of mobs varies (Panamah, for example, uses EQWatcher to keep track of these things; I also use EQWatcher and hear how many kills we've done per hour).
Dungeon Size
Some dungeons have ~90 mobs, some have more like ~150. You don't know which you have until you zone in. This explains a lot, for me. There are some dungeons that my group can completely clear, every mob dead (we can't quite do that in 90 minutes, but we can within the 2 hours). Other dungeons, we can't clear more than about 80% of the dungeon. This is with the same group and the same number of kills per hour (we get between 50 and 55 kills per hour, depending on factors like whether someone goes LD, needs to go afk, etc.).
Of course we had the subjective impression of "wow, we must have really kicked butt" for the dungeons we cleared, but it turns out that our skills, tactics, etc. had little or nothing to do with that - dungeon size/population was the key factor.
We took a collect mission the other night where we needed 28 items. Not only did we succede, we cleared the dungeon (I believe we found all 42 that the dungeon was seeded with, actually). We've had other missions with 30+ items that we didn't win (actually we've won them ALL in "overtime").
Collection Missions
Here's the simple math. Say your collect is for 20 items. They seed, according to "official word" (apparently) and people logging, 1.5 times the collect items into the dungeon. So, in this case, 1.5*20 or 30 items into the dungeon.
Case A. Small Dungeon, Small Collect. The dungeon has 90 mobs. This means that 30/90 or ONE THIRD of the mobs have the item you need. ON AVERAGE, you will need to kill 60 ((90*20)/30) mobs to collect 20 items. You can still get unlucky and have to kill every mob (and, apparently, containers), or you can get REALLY lucky and kill only 20. Best case 20, worst cast 90. On average, 60.
Case B. The dungeon has 150 mobs. This means that ONE FIFTH (30/150) of the mobs will have the item you need. ON AVERAGE, you will need to kill 100 mobs to collect 20 items ((150*20)/30). You can get unlucky and need to kill 150 or you might get lucky and need to kill only 20. Best case 20, worst case 150. On average, 100.
In other words, if you get a big dungeon in a theme, you are going to have trouble doing a collect mission - if you can kill 100 mobs in 90 minutes, you have a good chance at winning, but if you can't kill 150 in 90 minutes, you still have a chance of losing. If you can't kill 100 mobs in 90 minutes (groups without really high DPS), then you are going to lose most of the time. My group can kill more like 80 maximum, for example.
Collect Size
Does collect size matter, as people have guessed? For case C & D, make the number of items 30 instead 20. In this case, 45 are seeded into the dungeon.
Case C. 90 mobs. ONE HALF of the mobs will have the item. This means you need to kill 60 mobs on average to win (every other mob, on average, will have the item you need). That is, (90*30)/45. Best case 30, worst case 90, average 60.
Case D. 150 mobs. Almost ONE THIRD of the mobs will have the item (30% of them). This means you need to kill 100 mobs, on average, to win. That is, (150*30)/45. Best case 30, worst case 150, average 100.
Summary of Collect Missions
Case A: Small Dungeon, 20 Collect. 60 mobs needed.
Case B: Large Dungeon, 20 Collect. 100 mobs needed.
Case C: Small Dungeon, 30 Collect. 60 mobs needed.
Case D: Large Dungeon, 20 Collect. 100 mobs needed.
In a way, SOE did the math right in terms of balancing the number of collect items you have to find. What they didn't balance was the "difficulty" of large dungeons and small dungeons. A group with high DPS that can kill 100-120 mobs in 90 minutes won't have a lot of trouble BUT they still can get unlucky despite their very high DPS and/or skills.
A group with lower DPS that kills 50 mobs per hour is almost always going to lose the B & D (large dungeon) scenarios. They just can't kill the average number of mobs needed. They can still get lucky and get all the pieces within the ~75 mobs they DO kill, but it's unlikely.
Rescue Missions
This one isn't AS mathematical because the layout and location fo the target matters more than the sum total of the mobs. On average, though, the more mobs there are in a dungeon, the more you will need to clear to get to your target. If, on average, you need to clear 70% of the mobs in a dungeon, you're looking at clearing 68 mobs versus 105 mobs. However, layout of the dungeon impacts this by changing the percentage of mobs that need to be cleared. If the dungeon is very linear with few side branches, you may need to clear 90% of the mobs instead; if the dungeon has many side branches, you may need to clear only 50% (or whatever).
Size of the dungeon still has an impact - there are small linear dungeons and large linear dungeons, for example. A small linear dungeon could be harder to complete than a large non-linear one.
Assassinate Missions
This is very similar to a rescue mission. Apparently during beta they were quite different because you could invis or pacify your way to the boss, thus circumventing the number of mobs in the dungeon. Now, however, you have to kill some percentage (I believe it's 60%) of the mobs to SPAWN the boss mob.
If there are a large number of mobs, it takes longer to spawn the boss mob. If there are 90 mobs, you need to kill only 54 mobs to spawn the boss. If there are 150 mobs, you need to kill 90. This has lead to problems for some people - in a large dungeon you can reach a "split" in the layout of the dungeon BEFORE you've killed enough mobs to spawn the mob.
Of course you've only SPAWNED the boss at that point - you still need to clear the rest of the way in some cases, which will TEND to take more time in a large dungeon. As with rescues, the layout of the dungeon matters - with a linear dungeon, you may still need to kill practically every mob.
Both rescues and assassinates add the difficulty of finding the target, though a tracker + a map can mitigate that difficulty (it can still be tough sometimes to determine where the target is, just depending on the layout of a dungeon; however, a tracker and a map allow you to correct your errors fairly quickly).
Slaughter
These, the highly preferred mission type, differ from those above in a simple detail: they have no relation to the size of the dungeon. If you need to kill 58 mobs, it doesn't matter whether there are 90 total or 150 total. (The layout can still matter some - the time it takes to move and/or pull can factor in; long empty corridors may slow down the kill rate; large rooms that "require" pacify or crowd control my compound that.)
Also, these missions have little or no luck involved. There are no "wrong turns." There is no worry that you'll get unlucky and just not find enough mobs, etc.
Conclusion
Skill and, once you've reached some minimum of skill, DPS matter but it's RANDOM how much it will matter. My group, for example, with it's generally low DPS can do just fine in Slaughter missions, Collect missions in small dungeons and Rescue missions in small, non-linear dungeons. We don't do well in large dungeons simply because our DPS is on the low end (but probably not on the low end for casual players).
The fundamental problems are:
1. For collect missions, the seeding of items must take into account the number of mobs not just the number of collect items. Basically, they need to scale the number of items so that you need to kill, on average, 60 mobs to win. People will STILL be more likely to lose large dungeons (many can't clear a large dungeon but CAN clear a small dungeon), but the risk is reduced.
2. Extra reward needs to be provided for Rescue and Assassinate missions. There's no good way to control for dungeon layout and dungeon size in this case, so increase the "reward" in the risk/reward ratio. Make these missions yield 25% (or whatever) more AP's on a win or win in overtime. These require killing AT LEAST as many mobs as the slaughter mission IN ADDITION to either leading someone out or killing a boss mob.
3. Slaughter mission difficulty is about right IF reward is boosted for rescue and assassinate. Many groups are still going to have trouble with these - they are a guaranteed win for SOME groups since it's just a matter of DPS, not doing anything stupid, and not having any "outside" factors occurring (e.g., Linkdeaths).
The bottom line is that if you care about APs per unit time, you're crazy to take anything other than Slaughter missions. Who likes to have their success or failure determined by a flip of the coin? Currently the coin flip that determines which dungeon you get will determine, for many people, whether you win or lose. Some people are beginning to use the "leave adventure" function when they see that they are in an "unwinnable" dungeon (i.e., one that is going to require more kills than their group can produce).
Challenge is good and I will say that LDoN has challenged my group and made it MUCH more fun to play the game. Dealing with frequent multiple pulls, managing mana over the course of an adventure, etc. is fun. If we screw up, that's fine and good - we actually enjoy the post game analysis of figuring out what we did wrong and how we can improve in the future.
What is NOT good and fun is having the RNG determine whether you're going to lose, despite using the best tactics you can use. There are many times where we can look at a map 20 minutes into the adventure and KNOW whether we're going to lose. Doing everything right and slaying as quickly as possible and then realizing that you can't possibly get to where you need to be in the alotted time is frustrating. No, the time isn't completely wasted - you still got experience and loot, but that is beside the point.
Frustration involves some factor getting between you and your GOAL. Having that be a random factor is bad. This is the reason that many of us refuse to do long camps for an item - despite the fact that you might be getting good experience and loot while doing the camp. LDoN amounts to a long camp even if you succede every time, but it's finite and controllable, making it more pallatable to the casual player. Haivng the dungeon size/layout determine your success changes that though - the natural response is to minimize random chance and simply take slaughter missions.
My group WANTS to do the other mission types to keep things fun. However, the "fun" of saving that high elf woman for the 10th time wears thin. The fun of slaying Bob, who is unifying the mobs in the dungeon, tapers off pretty quickly. If the difficulty of the missions were equivalent, then you would still just take what's offered and go do it - variety at least is built in. When some missions are much more prone to failure and there's NO reward associated with the additional risk of failure, people are just going to avoid those missions. Either reduce the risk (e.g., by seeding more items based on the number of mobs) or increase the reward (e.g., by increasing AP's for some mission types).
One other note. From my limited experience with some themes and from looking at maps of the different dungeons, it appears that some themes are worse than others in terms of the dungeon layouts. For example, Guk and Miragul both seem to tend to have very large dungeons which make anything but a slaughter mission in those themes very risky for the average player.
I was shocked to see my rating in Guk jump to about 600 upon winning ONE mission, which implies that either few people are fighting there or few people are winning there. Mistmoore, for example, isn't "better" just because of the number of undead - you seem more likely to get a "good" dungeon layout and population. We've won many rescue missions there and the few we lost are because of number of mobs we had to kill was too great for our DPS - for some themes, I'd never take a rescue mission.
Well, one other note. The analysis above assumes that there is no correlation between what type of mission you get and the number of mobs in a dungeon. It is possible that you are more likely to get small dungeons for rescues, for example - but that does not SEEM to be the case.
