Manavis - Experience gain charting

[Jaxad0127]

Apologies for the double post, but I believe this is worth drawing attention to.

In the dataset, there are 1102 experience gain instances of any type for level 1. That's a lot of instances, given how you will only need one or two instances to level up at this stage. However, there are only 32 instances for level 2.

What is happening that makes ~97.18% of characters leave before level 3?

Almost all instances are killxp, of which a little under a half use a short sword or a bone knife. So, a lot of this will surely be alts. But is it the alts that people get bored with, or the alts that stick around? Are new players really that dissatisfied with what they find? What is going on here?


---Freeyorp

Looking at it over a longer time range will likely help. This is just a single day's worth of logs.

[Freeyorp101]

Apologies for the double post, but I believe this is worth drawing attention to.

In the dataset, there are 1102 experience gain instances of any type for level 1. That's a lot of instances, given how you will only need one or two instances to level up at this stage. However, there are only 32 instances for level 2.

What is happening that makes ~97.18% of characters leave before level 3?

Almost all instances are killxp, of which a little under a half use a short sword or a bone knife. So, a lot of this will surely be alts. But is it the alts that people get bored with, or the alts that stick around? Are new players really that dissatisfied with what they find? What is going on here?

Looking at it over a longer time range will likely help. This is just a single day's worth of logs.

People shouldn't be taking more than a day to reach level 3, and the level 1 instances stick out regardless of the time slice chosen. That's a lot of people giving up and leaving right away.


---Freeyorp

[tux9th]

What is happening that makes ~97.18% of characters leave before level 3?

Well you reach level 3 when you kill the scorpion who attacks hassan. So either people don't get how to pass hassan and quit or they do pass hassan and quit afterwards.

[Crush]

What is happening that makes ~97.18% of characters leave before level 3?

Almost all instances are killxp, of which a little under a half use a short sword or a bone knife. So, a lot of this will surely be alts. But is it the alts that people get bored with, or the alts that stick around?

People often create alts solely as mules or for chatting. There is no reason to level these characters.

Unfortunately it's hard to say how many of these early quitters are really new players, because TMW doesn't prevent players from creating multiple accounts. It does in fact encourage it by allowing someone to log in with multiple characters at once when they are on different accounts.

And when it would be possible to isolate the early-quitters, it would be even harder to find out why they quit, because you can hardly ask them.

[Nard]

Apologies for the double post, but I believe this is worth drawing attention to.

In the dataset, there are 1102 experience gain instances of any type for level 1. That's a lot of instances, given how you will only need one or two instances to level up at this stage. However, there are only 32 instances for level 2.

What is happening that makes ~97.18% of characters leave before level 3?

Almost all instances are killxp, of which a little under a half use a short sword or a bone knife. So, a lot of this will surely be alts. But is it the alts that people get bored with, or the alts that stick around? Are new players really that dissatisfied with what they find? What is going on here?


---Freeyorp

I am not sure that chars leave before level3: If I take my own example, I have an account that I actually use for storage, I wanted a consistent set of names, so I reserved the whole set of the 9 names by creating the 9 chars. At the moment 2 of them are level 42 the 7 others are level 1. As I see it (some players have different habits), I level chars in 4 "steps": 1 to 30-40, 30-40 to 60, 60 to 70, 70 and upper each step in a different place.. It can differ a bit if I play in party, with a partner. Players have to share their time between grinding, levelling their main char (or actual main)., and levelling alts to break monotony. Of course, periodical quests are an occasion where many alts (old and new) go outside to see the sun, thus we should see peek xp gain at low levels in these periods. The age requirements on seasonal quests is also a reason why players create chars by advance without playing them(it was not enough last eater ).

When I suggested force or agil on Y axis, I didn't think as a player but as pure statistician and seeked for a variable that could dicriminate the players. The densities in the status plot str-str or agil-agil seem rather equally scattered. Thus my suggestion. There are probably other (better?) choices, but they may require to record different data.
Edit,: keeping X axis as is will show the same map discrimination except that they will not only move along the staight line.

[yourmistakes]

in regards to 97% of players leaving or not levelling more, here are my ideas from several years ago: http://forums.themanaworld.org/viewtopi ... 12&t=14859
quoted for ease of reading

i've split this post from http://forums.themanaworld.org/viewtopi ... 12&t=14858 to avoid getting off topic
imo, the starting area of tulim doesn't really show the game off. everything is drab browns and tans, the tile sets out there badly need updated.
i propose a solution that could be accomplished in a number of ways: start the player off as a bad ass.
either have the player start as a high level char, or have them start fighting nerfed versions of high level mobs.
either way, throw them into the **** after the start, with a huge battle raging around them in a colorful and varied map. they should inevitably lose the battle and awaken with the current starting gear and level 1 at sorfina's.

if starting them off as a high level char is possible, you should give them the option to choose class and have the server adjust stats+gear+skills accordingly.

to reiterate: the game should sell itself.

this is just one of many possible solutions, of course, but the fact remains that we have concrete data showing the regular departure of characters at an early level.

[Freeyorp101]

Have three prototypes with the map chart y axis showing experience gain instances, average str, and average agi, respectively. Bubble size has been changed to be constant for each.

I think I actually like the total level xp versus xp instances more than the current one. It makes the outliers that much more apparent compared to when the latter relied on bubble size. With the y axis using XP gain instances, it's also consistent with the other charts.

map-xp-instance-y.png (74.74 KiB) Viewed 2980 times

When compared to the average attribute, I've changed the scale to be linear. But I'm not sure that this is as useful since it is redundant when you can already select maps, or ranges of attributes with the stat chart, and a lot of information is lost when we only consider the average. That said, it does spread things out quite nicely, which I think was what the main aim was.

map-avg-str-y.png (70.03 KiB) Viewed 2980 times

map-avg-agi-y.png (71.05 KiB) Viewed 2980 times

---Freeyorp