Wolverous
Score: 22136490Posts: 39 Reputation: +12 Joined: 2018-11-11 |
Posted 2019-08-17 18:34
|
0
|
|
|
So I just finished Animality to 100% and really start to wonder what is wrong with the value of the points on MGS. I mean, I love MGS as the site itself, but points are totally worthless. Finising Animality, which is ONLY HITTING UP AND DOWN and finished in under 3 hours is worth 106936 points, but completing Dark Souls 3 which takes up to 70-80 hours and a lot of skill is only worth 22772 points? This is completely ridicoulos. Coloring Pixels, a game where I DRAW PIXELS... I mean, completing the first arts takes 45 seconds and is worth 1714 (!) points. ONE (!) Achievement. This is the SAME amount then completing Dark Souls 100%, more points worth then getting ALL Trophys in Borderlands 2 and WAY MORE points then completing the whole freakin Sphere Grids for all main characters in Final Fantasy X, which takes 100h+, only worth 1445 points. I mean this can't be real. The same for Zaccaria Pinball. A single Achievements is worth 1938 points which took me 20 seconds (!) to get. Otherwise getting 55 Exalted Reputations in World Of Warcraft which maybe takes you years is only worth 1077 points? This is totally embarrassing. Why this kind of point system? Why are espacially Free-2-Play Games so overrated? This must be so frustrating for people who try to complete great and long games. For me, I just want to understand why. Thank You. |
|
Marexai
Score: 8037983Posts: 115 Reputation: +46 Joined: 2018-10-08 |
Posted 2019-08-20 20:33
|
0
|
|
|
I believe the point system on MGS was designed fair and square but it does not probably factor that some games have an enormous number of achievements that don't require much or any effort. (It's actually the devs that are to blame.) I think it would be a good idea to include some no. of achievements per game / game length ratio... which could lower Pinball's points and increase those of Dark Souls or Skyrim. Another way would be to calculate the average hiatus between each achievement in a game: Skyrim's cheevos would be worth much more as they are a few hours apart as compared to say the Pixels. Although I have no idea how much pressure on the system all this would create... |
|
|
Marexai
Score: 8037983 Posts: 115 Reputation: +46 Joined: 2018-10-08 |
Posted 2019-08-20 20:46
|
0
|
|
| BTW, Animality is clearly a SPAM game – I don't understand why it hasn't been identified as such since an automated response system should have been in place. | |
primal_r
Score: 6766555Posts: 1497 Reputation: +259 Joined: 2011-09-09 |
Posted 2019-08-21 21:13
|
0
|
|
|
Well, one basic issue with the scores is that it does not take into account how fun a game is. It uses the available statistical data such as completion rate & completion time, to guess how hard an achievement is. But when a game is more fun, more people are inclined to grab an achievement. So the completion rate will get skewed. There is also the fact that in order to get outliers in score; a big statistical sample size is needed. So games with a lot of players get a bigger sample size and then they can theoretically get more score. And for Steam, where the games are generally cheaper than f.ex. PSN, people stock up on games that they just try once or twice and as such the statistics get biased. So, while I agree that there is certainly room for improvements, and some achievements are way off - there are also big chunks of it that seems to have gotten an appropriate relative valuation. As there are over a million achievements in the database; representing a wide variety of games, situations and play-patterns - the mileage will vary. |
|
|
Wolverous
Score: 22136490 Posts: 39 Reputation: +12 Joined: 2018-11-11 |
Posted 2019-10-22 04:31
|
0
|
|
|
Any chance we get improvement for this subject? Maybe we can help? I come to this because I just saw another ridiculous comparison. I finished Super Mario Land (Retro Achievements) which made a value of 38 155 points. Took me 4 hours. On the other hand I'm playing Judgment right now, and completing that one makes a score of 38 648 points but takes over 90 very exhausting hours. So humiliating in consideration of scores and how hard to get some achievements are. On the other hand I asked myself if it would be possible to separate the Retro Achievements, because RA has normal achievements and hardcore achievements (without saves) and atm the scores are the same, so it doesn't matter if played normal or in hardcore mode. Maybe there is a way to reward the players who play on hardcore?! |
|
.removedUser.
Score: 0Posts: 39 Reputation: +8 Joined: 2013-02-14 |
Posted 2019-10-23 17:56
|
0
|
|
| Isn't the points curve just like all the other trackers in that it's based on how many have an achievement out of how many players for a given game? Cause if you understand what that means, it makes sense why games like Zaccaria Pinball, Animality & Super Mario Land (games only a relatively small modicum of players are tracking and even less are completing) have more expensive achievements than Dark Souls, Skyrim & Judgement (games that are owned by piles of people with a generally high completion rate). | |
|
Marexai
Score: 8037983 Posts: 115 Reputation: +46 Joined: 2018-10-08 |
Posted 2019-10-23 21:23
|
0
|
|
|
JSoup, you are right, but the OP meant the difference between two games that yield the same amount of points, though the first one takes 4 hours to complete while the latter 90. This is not the problem of this site but of the devs who designed the achievements – some games have simply far too many achievements while others too few for their lengths. I think it could be approached by implementing a modifying formula that could take the game length into account: 1. either by calculating the completists' average time span between the first and the last achievement, and the resulting figure would be naturally applicable to the overall score – also for players who haven't completed the game yet 2. or by using the game length data from howlongtobeat.com to modify the formula This is just my theorycrafting. I am not good at math or programming so I have no idea how difficult this might be. |
|
|
Wolverous
Score: 22136490 Posts: 39 Reputation: +12 Joined: 2018-11-11 |
Posted 2019-10-24 04:39
|
0
|
|
|
EXACTLY! The Achievements needs to be calculated with Game Length. I also thought about howlongtobeat.com, it's technially possible to implement this on this site. Like: Playerbase, Game Length, Completion Rate. That would automatically result in a way more realistical point system :) |
|
|
primal_r
Score: 6766555 Posts: 1497 Reputation: +259 Joined: 2011-09-09 |
Posted 2019-10-24 22:02
|
0
|
|
|
There are multiple factors that decide the score of an achievement, but yeah - rarity is the most dominant one; and rarity will be skewed based on how fun a game is to play in general; which is currently not accounted for. The site already factors in time of completion and completion rate. |
|
|
Wolverous
Score: 22136490 Posts: 39 Reputation: +12 Joined: 2018-11-11 |
Posted 2019-10-25 05:49
|
0
|
|
|
I understand this rarity thing, but then the "amount of players" value has to become less?! I mean, Zaccaria Pinball, Soccer Kings - '11' Lamp Achievement, 7,1% completion rate worth 1542 (!) points and takes like two or three minutes and Judgment, The Greatest Detective (Obtained ALL (!) other trophies), 0,6% completion rate worth 1600 points? which takes 90+ hours. One factor kills the scores. The 0,6% achievement should give a LOT more points if rartiy is so important. I mean even playing through the WHOLE Judgment Story which takes nearly 40 hours is only worth 422 points. Don't get me wrong, but killing the third Boss in Super Mario World is worth WAY MORE points on MGS. LOL. The point value doesn't have to based on how fun a game is to play, but if you say time of completion is already in use, then something goes really wrong as you can see with those examples. It would be so great if you would improve those factors. |
|
|
Wolverous
Score: 22136490 Posts: 39 Reputation: +12 Joined: 2018-11-11 |
Posted 2019-10-31 06:28
|
0
|
|
|
Zup! Zero 2 has a value of 186149 points. A spam game with 17% completion rate :D Now I have seen everything. |
|
Krazitrain
Score: 27733829Posts: 3475 Reputation: +1694 Joined: 2014-10-09 |
Posted 2019-11-01 02:39
|
0
|
|
|
In the case of the Zup style games, I'd suggest (if possible) ignoring any duplicated achievement names/descriptions, and count each one only once. For example, there are probably 10+ achievements for "Reach Level 2", all named the same with the same description. Have something that ignores any achievements in this game with the same name, and only count it once. This would resolve the high scores, without flagging them all as spam. |
|
anonymous1
Score: 28439041Posts: 32 Reputation: +6 Joined: 2018-09-21 |
Posted 2019-11-09 23:04
|
0
|
|
| I agree. With exophase I love the new website design. But the new exophase point system changed many of my leaderboard stats. At least meta game score stays consistant, | |
|
primal_r
Score: 6766555 Posts: 1497 Reputation: +259 Joined: 2011-09-09 |
Posted 2019-11-10 20:14
|
+3
|
|
|
Well I have modified the system many times over the year... Or, I think refined is a better wording for most of the changes. Added new metrics that have changed things around a bit... Usually the leaderboard stays about the same though; except for people who have focused almost exclusively on overvalued achievements. |
|
turbodream
Score: 2032826Posts: 6 Reputation: +3 Joined: 2019-11-08 |
Posted 2020-01-12 22:30
|
0
|
|
|
My understanding is that most of these games with really big points value will get devalued over time. So the system is pretty fair. Anyways, I don't think many players are deciding what to play based on points only and if they are they are missing on a lot of fun titles (and wasting their money). IMHO, basing points on time (how long to beat) is a bad idea because a lot of older games are short to beat when you are good at them but you need to hone your skills first. Nobody beat Mega Man 2 in 1 hour the first time they played it. |
|
|
Marexai
Score: 8037983 Posts: 115 Reputation: +46 Joined: 2018-10-08 |
Posted 2020-01-13 00:48
|
0
|
|
| howlongtobeat.com was a bad example on my part, I admit... but it was to show that the time it takes to complete a game should be considered somehow. I'm now having in mind those games that take 2 minutes to complete and have a few dozen achievements. Calculating the timestamp of the first and the last achievement would be a great metrics, though it's probably already implemented somehow... | |
|
Wolverous
Score: 22136490 Posts: 39 Reputation: +12 Joined: 2018-11-11 |
Posted 2020-01-13 16:31
|
0
|
|
|
Another big problem is: "So games with a lot of players get a bigger sample size and then they can theoretically get more score." I recognized that this only counts for MGS itself. Cause of newest example "AO Tennis 2" I saw that only 2 people are playing it (here on MGS), so the score of those achievement will probably never go higher, because their are just no more people here on this site how play that specific game. This game has a bunch of difficult Achievements (still 0% on Steam) and it is and I think will ever be, worth nothing 🤔 And I don't think Howlongtobeat is such a bad example. I always compare my gametime with HLTB and 95% of the time I'm faster then the expected Average time. Mega Man 2 in 3 hours is, in the age of internet, absolutly doable. ¯\_(ツ)_/¯ |
|
|
primal_r
Score: 6766555 Posts: 1497 Reputation: +259 Joined: 2011-09-09 |
Posted 2020-01-14 01:23
|
+2
|
|
|
Yeah in terms of sample size, some platforms do provide completion metrics; so these metrics could be used partially instead of the mgs-user-average. Or possibly to fill out when there are not enough sample size (otherwise we get bias based on a high percentage of the members of this site being achievement hunters, and platforms reporting for everyone - possibly even every owner as opposed to every player). I do think factoring in time-to-complete-achievement to a higher degree than currently is a good idea. For the 1-hour mega-man runs, a high percentage of the people who do complete it in 1-hour will have completed the game in "normal" time first, at which point they will have earned their achievements. Also, using the media value of stuff usually filters out most of those edge-cases. The current time-to-complete factor is based only on game-completion time. This value becomes quite strange for games like Zaccaria Pinball, where some achievements are very difficult but most are extremely easy; and the layout of the game allows you to choose any order of achievement earning. Meaning "time to complete the whole game" is extremely long, but effort to earn most achievements is easy. I mean, there are some achievements in Zaccaria that should have their current score, but there are loads that should be worth a tenth of their current. |
|
|
Wolverous
Score: 22136490 Posts: 39 Reputation: +12 Joined: 2018-11-11 |
Posted 2020-01-14 02:34
|
0
|
|
|
Ahhh this Zaccaria example makes things clearer. Here you would have to look again what percentage the respective achievement has, clearly bringing the game to 100% takes years, but if 7.0% have the achievement, then somehow it must be possible that it is not worth 1600 points!? I think the idea is good to increase the value of achievements that take longer, which then values games like Yakuza and Co. that last an unbelievably long time, but just come across too worthless here. But would the increase automatically give Zaccaria Pinball more value, or can you do something about it? Really cool that you try to improve things with us :) |