Replies: 5 comments 22 replies
-
The original plan was to make level actually be required to play certain mods (at all? for ranking?) to give a sense of unlocking in the game. I think it would be cool to outline other things we could potentially make level based unlocks to increase the excitement and goals of new players. |
Beta Was this translation helpful? Give feedback.
-
I like the idea of going more towards a "xp" system using multiple stats. Medal count is one of things i actually never thought about but it would probably make for a pretty good base stat to expand on instead of total score. |
Beta Was this translation helpful? Give feedback.
-
peppy's recent twitter post about making a new metric for related to dedication sparked my interest, so I would like to share a proposal for that metric, as well as a proposal to revamp the level system to use more statistics. New Statistic (dedication points)The calculation is the following : ClearValue = SSCount * 6 + FCCount * 2 + ClearCount
ScoreValue = RankedScore / 1_000_000_000
DedicationPoints = Sqrt(ClearValue * ScoreValue) (The above assumes that an SS will contribute to all 3 counts, meaning it actually increases clear value by 9, and likewise, an FC increases clear value by 3.) This gives the following leaderboard (made by @EngineerMark) : In my original twitter proposal, I inculded a bonus for medals to the total, which could be as simple as Level system revamp proposalAs proposed by OP, I included more stats to contribute to level.
The level-to-XP formula goes as follows :
This makes players reach level 100 at a similar point than the current system, while keeping levels going past that without feeling like the current system's wall. The highest level players with this formula sit at around Lvl150. You can play around with the values using this spreadsheet Notes : (1) Ideally, medals would be separated by ruleset and give different XP values depending on their rarity / difficulty. It doesn't feel right for the 1 star pass and 10 star FC medals to give the same amount of XP. (2) Performance is also a bit weird because PP reworks could cause levels to go down, but it can be removed if skill-based medals give good XP rewards instead. (3) Playtime is included instead of playcount because I don't see a valid reason why playing a 10 minute map should give you less XP than playing 10 1-minute maps. |
Beta Was this translation helpful? Give feedback.
-
I made a revision of my level rework proposal, removing the contribution from performance points, and adjusting the values to improve the balance and scaling. You can see a ranking of players by this level metric here, and the top 20 is on the image below. Here are the new values / level formula :
You can play around with the values using this spreadsheet Notes : |
Beta Was this translation helpful? Give feedback.
-
What do people think about purely, and unweighted summing pp. Potential issues I see: This also forgets about medals, or other things (example: interesting mod combinations) we want players to experience, they're essentially still going to pp farm just in a different way. There are many more issues, I will leave articulating those to people more knowledgeable. |
Beta Was this translation helpful? Give feedback.
-
Overview
Currently a user's level is determined purely based on their total score count from their profile, which arguably makes level a redundant and useless statistic. I've heard that there is a goal to change level in the future, but not really any set plans about how to do so (at least none that I've heard in public). I'd like to just throw out a few ideas for improving level even if some of this may have been discussed in dev channels or some other place prior.
Ideas
1. Create a new statistic to replace total score as a basis for levels
A similar (or even the same) formula to compute level can be used, but the idea is having a new value like "xp" to replace it. Certain activities can add to a user's "xp" total, such as gaining ranked score, setting a score on a new beatmap, and/or (most intuitively) each time a user unlocks a new medal. As a user does different activities like these their level goes up, much like it would in other games.
All of these events should be easily trackable to not add any additional overhead to the system. For the examples given profiles already access the ranked score, total SS/S/A counts, and total medal count respectively. This would just be a matter of multiplying the numbers by a predetermined value and summing them.
Events should be irreversible so that a user cannot randomly lose "xp". In every game the basic concept of experience and levels are a total which is continually rising, so we would need to sample stats which can only gain to achieve this. I am weary of including so-called "pp milestones" or similar events due to the fact that the user can overwrite the score with a lower value causing their "xp" and level to go down.
Here are a few of the user statistics which have potential to be used towards the xp count:
2. Revamp level formula to better suit players
It may sound strange at first but this is an issue that is made very apparent from the current way level is computed, which is as follows:
The obvious hole being that this piecewise function simply stops being exponential at level 100 and from that point on a user just needs to keep gaining ~100B score per level. Beyond this lower levels are extremely easy to achieve, especially considering that playing an Easy or Normal difficulty once and getting ~2,000,000 score will instantly bring you to almost all the way to level 7. Reworking this function to be more user-friendly at the lower end and more challenging at the upper end could greatly improve how level works (although finding this may be harder than I believe).
x^( 1 / e )
orln( x + 1 )
could be very simple candidates as a starting point to fit this function if you gave it some other coefficients.There is lots of potential here to make user level much more relevant and useful given these ideas, or at least I think so. Constructing the new system/formula will definitely take some tweaking of values and fitting of curves for this.
I hope I've outlined some good ideas to carry over into the next iteration of level, and that others will be able to build upon and think of an even better system.
Beta Was this translation helpful? Give feedback.
All reactions