Join us with one click:

We will never post without your permission!
OR
Join us using e-mail:
Desired password:
Community -> Message Center
Message Center
Here you can see changes to forum topics that you are following. Replying to a topic makes you automatically follow it. If you want to follow a topic without replying, you can click the Follow link in topic.

ForumsMGS Discord ChannelAchievement/Game Comments


My Unread threads
All my followed threads
Site Newsblog (5)
Site Newsblog

Optimizations
11:05 January 02, 2024 by primal_r
I'm currently making a bunch of optimizations to the site, focusing on top-lists and also on pagination.
If the site uses less resources, we can have more scanning threads running in the background.

Let me know in the forums if you find anything off!

Roblox scanning update, dates + skips
14:24 September 02, 2023 by primal_r

* Roblox scans will now include the timestamp for their earns, instead of being listed as "earned offline" (thanks to Adrmcr for letting us know the proper endpoint!)

* Roblox achievements named [ Content Deleted ] or Content Deleted will now be skipped. Previously they would trigger the automatic invalidation filter, which is not desired.
These achievements were (potentially) real achievements before their deletion from Roblox. Anyone who earns those achievements and scans them on Metagamerscore before they are changed to those names will be able to keep them (unless the game they belong to are later marked as spam through some other mean). They will also keep their original name on Metagamerscore. ( unless those names are against the terms of service ).



MetaScore V5.3.1 deployed
10:26 August 12, 2023 by primal_r
MetaScore V5.3.1 deployed

MetaScore V5.3.1 changes the two metrics related to popularity that metagamerscore calculates.

TLDR
Scores went down for a week, should be ballpark back soon.

Background / problem
In the previous version (5.2), all games were ranked in comparison to all other games in order to know how popular the game was. Popular games then got a small bonus to their achievement values, while unpopular games got a small penalty.

As we saw a huge influx of hundreds of thousands of tiny 1-player Roblox games, two things happened:
* The calculation time of this metric suddenly became a bottle-neck for quickly going through the scanning-queue.
* As all the tiny games were added to the bottom of the list, all other games were pushed towards the top, being spread over only the tiniest of percentiles; potentially overvaluating all games.
Importantly, this shift had already been slowly happening for a few years, but at a much smaller scale so it was less noticeable.

Initial fix problem
I tried to do a fix to both of these by introducing a minimum treshold, below which it would just always set to minimum value. The result on calculation time was very successful,
and the low/high games were unchanged in popularity. Due to the immediate result seeming to yield unchanged result but significantly faster calculation, I released this silently a week or so ago.
But after further analysis it turned out that the result on games in the medium size bracket became devalued. (Partly because of the fact that this was not purely a Roblox-game thing).

Another problem with my initial implementation was using the same treshold check for both popularity metrics. As such, many players saw a 10 to 15 percent drop in score over the past week.
Part of it could (and still can be) explained as normal fluctuation accumulated over the past 4-5 weeks where the queue was stalled and updates didn't happen correctlly, but most of it was due to the above change.

New solution
Version 5.3.1 aims to fix these things in a more solid way by having that minimum clamp but also having a linear interpolation within the bottom bracket as a smooth "bridge" between the two modes.
The two metrics now also does their entire calculation with separate tresholds.
Baseline for the multipliers have also shifted so that the new average should be closer to the previous values for most cases.

Additionally
Version 5.3.1 also changes the value for games marked as spam from 100 to 1.
Only those who have a huge percentage of spam-games will see any tangible difference from this. This action was done as a precaution to disincentivize creators from trying to circumvent the calculation logic.

I will continue to monitor the situation to make sure the actual result is within acceptable bounds.

Note: Metagamerscore calculates everything very asynchronously, so it may take about a week for everything to take effect.

Metagamerscore and spam games and roblox
23:00 July 17, 2023 by primal_r
I want to reiterate the metagamerscore policy on spam games.

The Idea of the achievement valuation on metagamerscore is to have some kind of correlation with difficulty.

Games that require no effort/are made just to give free badges will be marked as spam and give next to nothing in score.

Our current spam detection algorithms are tuned toward Steam spam games, and with the addition of the roblox platform we have seen a huge influx of spam games from there which do not get stuck in the automatic filter.

I will make adjustments to the detection algorithms so it detects the roblox spam games better, but this will take a few weeks.

In other words: dont waste time playing spam games, play fun games instead. Even if some of the spam games seems valuable now, it is only a matter of time before they are marked.

New platform support: Roblox badges
21:53 July 12, 2023 by primal_r
Today we have a new platform on metagamerscore; Roblox!

You can now get Roblox badges and have them tracked on Metagamerscore.

Sadly I haven't found any way to get the timestamp for when the Roblox badges was earned, so they will show as "Scanned offline" and not count towards the yearly scores.

They will still be fully tracked and scored for other types of scores though.

Proud sponsor:
© MetaGamerScore.com | info@metagamerscore.com Link Exchange | About/FAQ
Donate | Terms of Service | Privacy Policy
The company names, product names, service names and logotypes used on this web site are for identification purposes only.
All trademarks and registered trademarks are the property of their respective owners.