So, it's been almost 4 years since I wrote this. I thought I would share an email I got recently about this analysis and my subsequent response:
I understand that that article was written along time ago i suppose towards the end of 2004, so i was wondering did ncsoft make any changes to the level grind dificulty? if so is there any info on it? could you give me a link to it?
Btw i thought the article was well written and v clear. I also agree with what you have to say about the need to ease up on the tediousness on the grind in order to help maintain the lineage 2 community. If you follow world of warcraft at all you'll see they just did this. they cut the exp needed between lvl 20-60 by 20% and made mobs easier and gear btr. for the reason to encourage players to level up alts. gus they took your advice to lineage 2 and implemented it, haha.
My response:
Well, as it stands now, I have no idea what NCSoft has done to address keeping the community thriving. They are pretty strict with their MMO model which is centered around this idea that character development should take a long time - but it can be more "fun" if you cooperate with others and share the burden of the "grind."
I haven't played L2 Retail in the longest time so I have no idea what changes, if any, they have made to exp/level requirements and if they modified MOB difficulty, exp values, etc., etc. From what I have seen since the last time I played L2 retail in 2006, they keep the game "alive" if you could even call it that, by slapping on content patches - new areas for all level groups to explore and utilize. I unfortunately do not see them changing their leveling/exp model because of the uproar the "veterans" would have over making the game easier for new comers. You have to understand, there are people who have been in that game since its inception have invested A LOT of time and personal resources into developing their characters.
But people by and large in that game, based on some data I gathered back in 2005, gave up trying to play legitimately, and instead either bought adena or used bot programs to automate game play. That should tell NCSoft something, but it doesn't seem to get through to them. People do not WANT to spend their lives in the game. They want to have fun, they like the PvP, but they don't like all the penalties for death, the pitiful adena drop rates, the material drop rates, and most of all they do not like having it have to take 2 months to reach 76-77.
What will the future be? WoW got a lot of things right because they understood people have a limited amount of time to spend playing video games. There is a very small percentage of the "Gaming" community that can play 8-10 hours a day, and essentially live in a virtual world. Knowing this, Blizzard adapted and tweaked their "progression" model to satisfy the majority of their playerbase who can only play 5-10 hours a week. And of course the question becomes, well, if I can only spend 5-10 hours playing a game, and my character doesn't really develop in that amount of time, why should I keep playing? That's why you see Blizzard making "life" in Azeroth easier and easier - they are catering to the majority that supports a large chunk of their revenue stream for WoW.
As far as NCSoft goes, I think they had more commercial success in the US with Guild Wars - not the kind that Blizz had with WoW, but by and large, more people played and stuck with Guild Wars than they did with Lineage 2. Lineage 2 is for people who have all kinds of disposable income AND time. But also remember, the majority of the community in Lineage 2 also bots heavily and buys adena - so they're really not enjoying the game as it stands on its own merits. I think if NCSoft has learned anything, their next real MMO will be softer on people's time commitment, but who knows really until they release it.
Glad you enjoyed the article.
And now...on to the original analysis:
Somehow I feel I will get a good deal of nasty comments thrown my way for pointing this out; i.e., I’m a whiner, I just suck at the game, all that usual stuff.
But before you sling such things my way, please do take the time to read this article carefully. I am hoping that if enough people read it (I am sending a copy to NCSoft, too) that perhaps some dynamics of the game will change so that Lineage II, which is an awesome game, will retain players.
Why is this important you ask? Blizzard Entertainment is soon to release World of Warcraft, another huge MMORPG. However, they have done thing the other games have not: they have scaled down the level grind. Unbalanced level grind has the potential (potential is a key word here) to lose players to games like World of Warcraft. Why? People who play games where leveling up is perhaps the most critical aspect of the game want to feel that they are making consistent forward progress; at various levels in Lineage II, leveling up becomes more of a chore than entertainment. RPG titles must retain their entertainment value through a properly balanced reward system in acquiring levels. As of now, Lineage II fails to do that.
Many of the faithful of Lineage II will stick with the game no matter what, and that is fine. But remember Blizzard Entertainment: they have a track record for selling multi-platinum titles that crush competitors’ games. Still, there is a healthy and thriving community for Diablo II, which has been out for more than four years. If you don’t believe me, check out the number of Diablo II clans and Ebay auctions that sell items for the game. That is some serious staying power.
I’m certainly not advocating changing the game so radically that everyone can have a level 70 character in a matter of days. That would be ridiculous in my opinion. However, given the game dynamics and balance of Lineage II (C1) as it is now, the rewards reaped for leveling are not compelling enough at appropriate levels to sustain its popularity in the wake of World of Warcraft and Everquest II, which are soon to be released, and address the level grind issues.
This article basically details the mathematical trends between character leveling and experience points of MOBS. Once you read it, you might say, “so what?” or “damn, that sucks.” Either way, it is meant to be informative.
A note about the data: The data collected for this analysis was mined from http://lin.warcry.com. MOBS that had no values for experience points and dual MOB entries were removed from the data set as to ensure accurate analysis; therefore, every single MOB in Lineage II is not accounted for. Furthermore, many of the experience point values for these MOBS have changed since Chronicle I, however, the change has been so slight that in the end, the analysis would reveal the same overall trends.
A word about outliers and why they are important in analysis.
What is an outlier? An outlier is a data point in a data set that does not follow the trend of the rest of the data points, collectively. The diagram below shows a data point that is circled, which is considered an outlier. As you can see, the circled data point is far off from the rest of the data points in the data set. This particular data point can severely influence the analysis of the data set. At times, outliers are removed in order to clarify the underlying mathematical patterns that a data set is exhibiting. Outliers are also identified to understand why that particular data point does not fit in with the rest of the data points. In the graph below, the particular “outlier” is the Queen Ant, who at level 43 has over 18,000 experience points. The outlier is therefore understandable – The Queen Ant is a unique MOB, who obviously is going to have higher experience points than of other MOBS of the same level.
Mini Outliers
For purposes of this analysis, I have located and isolated “mini outliers”: these are data values for MOBS for a particular level of MOBS whose experience point value varies greatly from the rest of the MOBS in that level. Mini outliers were identified by MOBS whom had 50% or more experience points than the majority of MOBS who have roughly the same amount of experience points. Mini outliers are important to identify because they greatly influence the average experience point value per MOB per level. This will become important later in the analysis, as you will see. For now, below is an example of how mini outliers were found for a particular level-group of MOBS.
The rows that are highlighted in light purple identify that these three MOBS are outliers in the level 45 group of MOBS. It does not take complex mathematical equations to show that these three MOBS values vary greatly compared to the rest of the MOBS. However, we will take the Dread Panther as an example of why it is a mini outlier.
First, we see that the majority of the MOBS in this level-group have an experience point value of 2430. The Dread Panther has 4519 experience points. Doing a few calculations we find:
Dread Panther Exp./Majority Exp. Value = (4519 - 2430)/2430= 2089/2430 = 86% Exp. Pt. Difference
Quite plainly, the Dread Panther goes beyond the 50% threshold set by the analysis.
(Credit must be given to Kwood who pointed out a slight computational error from before. Thank you.)
Average Experience Points Per MOB per Level
This is an important statistic to derive from the data because this value has a direct influence on how many MOBS of a particular level are needed to be killed in order to attain the next level, assuming a character kills these MOBS on their OWN. Below is a chart showing the average experience points per MOB per level without any mini outliers removed:
Doing some simple math, we can show that the average experience points per MOB per Level is 19% higher than the majority of the MOBS at level 45.
19% = (2902-2430) / 2430
If we look at the same average experience points per MOB per Level with the mini outliers removed, we have an average experience points per MOB per Level of 2430.
Number of MOBS killed needed to level
This derived statistic tells the tale of problem with level grind in Lineage II. The number of MOBS needed to be killed in order to attain a level was derived by taking the amount of experience points needed to acquire that level divided by the average experience points per MOB per Level.
# of MOBS Killed needed to level = Exp. Pts. For level (x) / Average Exp. Pts. Per MOB per Level.
Below shows an example of how many MOBS need to be killed for level 45, with and without mini outliers:
With mini outliers: 808 MOBS killed
Without mini outliers: 965 MOBS killed
Differential: 157 MOBS killed (19% more MOBS!)
As you can see, the differential between averages with mini outliers and without mini outliers is considerable. Why?
As mentioned before, these mini outliers are important because they greatly influence the average experience points per MOB per Level. Many of these mini outliers are over 100% larger in experience value than the majority of MOBS in that level-group. Furthermore, the majority of the mini outliers do not have a high spawn frequency.
Let me illustrate an example of this spawning frequency, using the Orc Barracks as a hunting ground. As many people who have hunted there, Turek Orcs populate this area. Groups are often found (you were probably in one of them) hanging out in a particular camp, waiting for Turek Orc Warlords to spawn. You might also have noticed that competition for these camps where these Warlords spawn is fierce. Why? Well, the Turek Orc Warlord, a level 30 MOB, has over 2000 experience points (and gives 110 Skill Points, too). Turek Orc Footmen, Suppliers, Orc Elders, Sentinels and Archers have anywhere from 800-350 experience points less than the Warlord, though they only vary about 4-5 in levels. However, the Warlord spawning frequency is anywhere from 3-4 times less compared to these other types of Turek Orcs. This makes sense. The Warlord has a higher defense rating, is faster, harder to kill and therefore gives more experience points and skill points.
However, what is interesting is that the Turek Orc Elder is a level 34 MOB that has roughly 800 less experience points than the Turek Orc Warlord, but 300 experience points less than the lowest level (25) Turek Orc, the Sentinel. The Turek Orc Elder has a lower spawn frequency than the Turek Orc Sentinel, has a higher attack rating, a higher defense rating, and casts Drain Life to take anywhere from 75-100 hit points from you. Yet, the reward for killing a MOB that is 9 levels higher than the lowest Turek Orc is much lower. Furthermore, the Turek Orc Elder is not a mini outlier in the level 34 group of MOBS.
Below is a graph showing the differentials between number of MOBS needed to be killed to level, with (Before) and without (After) the mini outliers in the data set.
The following is the table that shows individual MOB kill differentials needed to level.
Explanations and reasons behind the trends in this data.
The rows highlighted in light purple show in some case how there is a huge differential in MOBS needed to be killed in order to level. There are reasons for this trend.
First, the mini outliers greatly affected the average experience points per MOB per Level. The mini outliers exhibited properties such as:
- They were a unique MOB
- They had a low spawning frequency
- They had a experience point differential greater than 50%
These mini outliers, when removed, show that a much greater number of MOBS must be slain in order to level up.
Second, this data reflects the number of MOBS needed to be killed, if one kills these MOBS on their OWN. Unless you are an archer of some sorts, past level 20, you are in a group with 3-5 other people. That being the case, you can take these values of MOBS needed to be killed in order to level and multiply them by 3 or 5! That will give you an idea of how many MOBS you realistically need to kill in order to level up. And with the outliers removed, you see how many more MOBS are needed to be slain in order to level up.
For example, if your character was at level 24, and if you were soloing, you would need to kill 406 level 25 MOBS in order to level up. But, if you are in a group with 3 other people, that number goes up to 1218 MOBS!
Why I left out the 7% experience point bonus
People wonder why I left out the 7% group experience point bonus per character. I have left it out because it consistently does not function properly.
For example: A friend and I were hunting Turek Orc Warlords a while ago, just to test it out, to see if it REALLY worked.Now, Turek Orc Warlords have 2018 experience points right? And with two people in a group, there should be an extra 141.26 experience points per player in the group. Right?
Unfortunately, it does not seem to work that way in the game. The experience points are split evenly down the middle. We each got 1009 experience points and 55 skill points.
Time Considerations
Ok, you just realized you had to kill a lot of MOBS at level 25 if you were in a group. However, groups, if built right, have the potential to kill more MOBS in the same amount of time versus if you were just soloing on your own. Furthermore, there are other time factors to consider when in a group: the amount of time you need to rest your hit points and mana points. When in a group, this amount of time may go up or down depending on the overall quality and skill of the group you are in.
However, let me illustrate an example where groups are inefficient in comparison to the efficient solo player.
Consider a group of 5 players in the Orc Barracks, hunting Turek Orc Warlords (TOWs from now on). On average, let us assume (I have actually timed groups killing Warlords) that a group of 5 players takes 15 seconds to kill a TOW; it takes the solo player 45 seconds to kill that same TOW. Those 5 players each roughly get 400 experience points every 15 seconds. The solo player, who killed the TOW in 45 seconds, gets all 2000+ experience points (and all the Skill Points, too) – per 15 seconds, that solo player gets 667+ experience points. The solo player is clearly 50% more time efficient than the 5 player group at leveling.
It is also important to mention that as MOBS increase in level, so do their difficulty, due to their increased attack ratings and defense ratings. Therefore, it takes a player longer to kill these higher level MOBS, even as they earn the privilege to use higher grade equipment. These factors make the level grind higher.
Another important trend to note is that rarely do players hunt MOBS that are their own level. More often than not, they hunt MOBS are higher in level. The psychology is relatively easy to understand – higher level MOBS are supposed to have more experience points, hence why they are hunted. However, as it has been shown, that is not always the case.
(This was added due to comments about Group MOBS)
Many have pointed out to me that my mini-outliers are intended as "Group MOBS" - meaning they are not meant to be solo hunted. This being the case, a group of two people hunting "Group MOBS" will split the experience and skill points evenly. In this case, they no longer become mini-outliers, but only if the "Solo MOBS" are hunted in groups, too. Differentiating between these two types of MOBS is unnecessary because functionally, when in a group of (x) number of people, the total experience point reward will be the experience point value of the MOB divided by (x). Unless you are performing "ladders of powers" data reexpressions to the data set, these mini-outliers, or "Group MOBS" would still have the same experience point ratios to "Solo MOBS."
Perhaps it would be of some help to actually define a quantitative value for “Level Grind.” However the problem with defining Level Grind as such requires that all factors of a characters skills, abilities, mana points, attack rating be factored into an equation that takes into account the MOB’s defense rating and attack rating. This could definitely be done, and perhaps one day I will come up with a dynamic formula for determining Level Grind.
Problems that Level Grind Produces in Lineage II.
I. Group Saturation
As a result of the Level Grind in Lineage II, one will typically see large numbers of groups and/or soloists concentrated in hunting grounds where low frequency MOBS spawn and have higher experience point values than MOBS in other hunting grounds who are higher in level. Classic examples of this phenomenon can be found in the Orc Barracks and the Ruins of Despair.
II. Lack of incentive to explore and hunt in other areas in the rest Lineage II
Due to the high concentration of groups in these areas described above, many players will forego or skip altogether other hunting grounds in Lineage II where MOBS are of higher level because the experience point reward does not justify the time or the resources needed to kill these higher level MOBS.
In relation to hunting in these higher level MOB areas, why spend 10 arrows and 10 D grade Soulshots on a level 35 MOB and get 1500 experience points when you can spend 5 arrows and 5 D grade Soulshots and get 300-500 more experience points by killing a lesser level MOB? Furthermore, the lower level MOBS are just easier to kill, thereby making the time and resource investment into level much cheaper. The risks to the self are also much less when hunting lower level MOBS that have higher experience point values than that of MOBS that are of higher levels.
III. Entertainment “Value” is weakened by the diminishing returns in increases of average experience points per MOB per Level.
In short, after a certain level, depending on your race, class and playing style, advancing in the game becomes more of a chore than entertainment. For a game whose business model depends entirely on recurring revenue through monthly account fees, a game can see a decline in the number of players renewing their monthly access due to the time/reward ratios being so low. As mentioned at the beginning of this paper, with the expected release of World of Warcraft, which has addressed the issues with Level Grind, games like Lineage II have the potential to lose players to these games, thereby weakening the overall online gaming community in the game itself. Because there is heavy reliance between races in the game, a steady exodus of players who keep the gaming community going can seriously jeopardize the balance of Lineage II.
Take dwarves for example. Dwarves are responsible for crafting materials, such as Soulshots, which are needed to level up more efficiently. Imagine if 25% of these dwarves who craft these items left the game. What would happen to the economy of Soulshots? Because there are less dwarves, less Soulshots would be made – hence a decline in supply. Soulshots would then become more expensive and demand would remain the same, or worse increase, requiring players to have to kill more MOBS just to be able to afford the supplies to level somewhat efficiently. In a word, playing the game would be economically inefficient. There is no telling what kind of ripple effects would take place throughout the game were this to happen.
What is the solution to fix all these problems/trends?
Fortunately, only one thing needs to be changed. MOB difficulty, spawn frequency, weapon damage, spell damage, defense ratings, item modifications and character leveling modifications do not need to be implemented.
If one looks at the following graph, one will see that the MOB data set exhibits a low logarithmic experience point change per level increase. Eyeballing it, I would estimate that it is to the ¼th power, perhaps even lower.
Essentially what needs to be done is that the higher level MOBS need to have higher experience point values that reflect their difficulty and level. At current levels, the average experience points per MOB per Level growth is too low when considering how many experience points are needed to be earned in order to increase one’s level.
I do not suggest that they should be modified so high that the growth pattern of average experience points per MOB per Level is similar to Diablo II. However, I do suggest that the mathematical trend exhibit a higher logarithmic/exponential function; y = x^1/2 would be appropriate equation for the MOB data set to roughly fit - any higher than this is unnecessary.
Why HP/EXP is the wrong metric to use in determining trends associated with level grind in Lineage II
Performing a regression analysis, or more plainly stated, the relationship between Hit Points and Experience Points for MOBS, one will see that looking at the relationship between the is only good for one thing:
Using Hit Points to predict what a MOB's experience point value is and vice versa.
It is not good for showing trends of MOB's increase in experience points' relationship with an increase in level and how that relates to the overall level grind.
Look at the following charts and graphs. The first chart and graph shows the strength of the relationship between Hit Points and Experience Points for MOBS through level 60. As you can clearly see, there are two outliers - The Queen Ant and Core MOBs.
The second chart and graph show the strength of the relationship between Hit Points and Experience Points for MOBS through level 60 with those two outliers removed. As you can see, the strength of the relationship is much stronger now (p value of .972 vs. .757 from before.)
If one performs multi-variate regression analysis with Experience Points as the dependent variable and Hit Points and Level as the independent variables, you will find that the data set will once again reflect a logarithmic function versus the linear function that HP/EXP reflects. What this shows is that using a MOBS level is not as effective in determining a MOB's experience point value than using Hit Points. This is what is known as "incremental validity" - Adding Hit Points to the data model does not tell us anything "new" about the data itself.
With that in hand consider this.
This would never happen, but it illustrates why HP/EXP is a bad example. Imagine a Brown Keltir had 2000 hit points and had 2400 experience points. But the level of the Brown Keltir is 1. Now if this were the case, level grind for many characters would not be an issue up to a certain level. Brown Keltirs spawn frequently, have a low defense rating, therefore making them easier, much easier to kill, than a hihger level MOB with the same amount of hit points and experience points.
I understand that on the surface, to many it seems intuitive to look at Hit Points and its relationship with Experience Points - just by eyeballing a set of 20 MOBS, you can see there is a clear and distinct linear relationship. But as mentioned before, Hit Points are only good for determining the experience point value of the MOB. Hit points does not take into consideration the difficulty of the MOB itself, which is related to the level grind. Higher level MOBS are harder to kill, and therefore, should have a higher experience point reward for killing them.
However, if you all go back to the original analysis and look at the scatterplot of level vs. experience points, you see that the data set exhibits a low logarithmic function. Intuition should also tell you that the higher the level the MOB, regardless of what its hit points are, you should be seeing a greater increase in experience point value - however, as the data set exhibits, that is not the case.
