Forums > News > Site Updates (Dec 10th)
Tracy Poff (2095) on 12/10/2015 8:45 PM · Permalink · Report
A quick update on some doings and undoings at MobyGames:
- The point change for reviews announced last time has been reverted, pending community discussion
- Company pages now display a list of games by the company
- Fixes and improvements for the new approver tool
Alaka (107782) on 12/10/2015 9:19 PM · Permalink · Report
-
About reviews, I would like to see them worth a flat amount of points. I mean, its not like game descriptions no matter how thorough are worth more than others. Approvers shouldn't have to worry about grading papers. Just my opinion on that situation.
-
Since it appears you're looking at company pages, are there any future plans on allowing for the editing of company trivia, similar to how game trivia is done?
Thanks for your work on the site,
Unicorn Lynx (181681) on 12/12/2015 3:52 AM · Permalink · Report
I agree. Just assign a flat rate of 10 or whatever points for review.
Cavalary (11608) on 12/13/2015 11:23 AM · Permalink · Report
If the bar for review approvals will rise very high and something will be done about poor already existing ones, sure. Otherwise, this will only mean that 3 sentences will be worth as much as a few pages of thorough, well-written and nicely formatted analysis... and up to two new games submitted. Oh yeah, very "fair".
MAT (241251) on 12/14/2015 12:19 AM · Permalink · Report
Well, badly scanned and losely rotated 200dpi scan with horizontal lines and whatnot is still worth 3 points just as a high quality one where user took time to fix and alight it and whatnot. Same for Sciere's 10000 words game descriptions as opposing to 2-sentence ones. Sciere's descriptions are longer and better than many of reviews I've seen, so there's no fair here.
As for silly 2 sentence reviews, those should be WIP-ed. We should have some 200 words or more minimum limit for reviews. Good and Bad sections aren't obligatory, but Overall one should have at least 200 words or so. Maybe more.
Alaka (107782) on 12/14/2015 3:58 AM · Permalink · Report
[Q --start Cavalary wrote--] And hey, I was asking for reward based on quality for descriptions too. [/Q --end Cavalary wrote--]
It's a bad idea to have things being graded on quality by an approver. It slows things down (some approvers won't even touch reviews because of this) and may cause accusations of bias and unfairness by a contributor. I get the feeling the site owners wants the site to allow both quick reviews and more detailed reviews in the future (similar to Gamefaqs system). Ideally the quick reviews would be worth something like half of the more detailed ones.
MAT (241251) on 12/14/2015 7:08 AM · Permalink · Report
Yes, quality grades shouldn't be included. Imagine having five levels of points for covers quality. Why should we only limit that to reviews or descriptions? Might as well for screenshot descriptions. No, it complicates things, and just because someone is more eloquent or has good writing skills s/he would be awarded more points. That's not a good policy, imho.
MAT (241251) on 12/14/2015 7:12 AM · edited · Permalink · Report
You say "heh...", but that review was helpful to 75% of people (80% now with my vote) according to readers' rating ;)) It's not the quantity, it's quality :) I personally would reward brief and concise reviews much more than some long dramas that spin tales. Of course, I would give 5 points for almost every of Oleg's reviews, but that's because his quality of writing and using really cool and exotic words ;P, not because it's interesting to read.
Having scores for reviews was bad to begin with mostly because the length was rewarded, not quality. And it was subjective. I personally gave 5 points to short but interesting and original reviews, while I didn't even bother to read through some very long ones to rate them.
Without grading system, I wouldn't approved the review in 5sec just at a glance that it's a valid review, and it would all go much faster. Approving is slow as it is, grading just makes it that much slower, and very subjective, based on approver in question.
Edit: Just read the "heh..." review, very cute. And very helpful. I'd give it 3-4 stars for sure if I was approving. It sounds like a very young gamer wrote it, and it's perfectly described as to how it affected him/her. 3-4 stars for cuteness :) Even got me to become curious about the game myself.
Martin Smith (81734) on 12/14/2015 10:13 PM · Permalink · Report
Perhaps the contribution points for reviews should be partly based on how the approver perceives the quality, and partly on how it is rated by users over time? Say, 2-4-6-8-10 from the approver and an extra point for every 2 helpful votes minus unhelpful votes, up to maybe 5 or 10?
MAT (241251) on 12/15/2015 12:03 AM · Permalink · Report
You got it wrong. People who want to find out more about the game will read reviews, it won't matter to them how many points certain reviewer got or how long the review is. It might matter to them how many other users found his/her review helpful. Although, from personal experience, people seem to mark reviews as helpful only if the reviewer praised the game they like as well. On IMDB those rating rarely affect me because they are inconsistent and something that certain reader didn't find helpful may be helpful to me and vice-versa. Rating should probably be removed, imho, as it may discourage people who get low ratings but like to write reviews just may not be too good at expressing themselves.
MAT (241251) on 12/16/2015 4:45 AM · edited · Permalink · Report
Here's my proposal for point about reviews.
Every review should be awarded 8 points as a base, just on approval, no matter the quality or length.
There should only be positive voting available, meaning that the review was helpful to anyone, and each positive vote would add additional 1 point until it would overall reach 15 points (max points per any review).
Negative votes should not exist, because that would discourage reviewers and contributors, and it really isn't friendly community when someone sees that many people rather his/her review as bad. Instead, such reviews simply wouldn't have positive votes for being helpful, so noone would know if that's because nobody read them yet, or just didn't rate them as helpful.
That's my proposal for new review scoring system.
Alaka (107782) on 12/16/2015 5:54 AM · Permalink · Report
You can still have negative votes, just don't have them penalize the contributor by reducing their point totals. However, some reviews on this site obviously took so little time to write, points should be deducted on general principle. I mean seriously? Of course, if anyone reads the canned response for reviews, its says one line reviews aren't allowed, but nobody cares about the site standards in this category it seems so whatever. Sow the wind, reap the whirlwind.
Pseudo_Intellectual (66801) on 12/16/2015 4:15 PM · Permalink · Report
As someone who's been here for over a decade and never yet written a single review, it's amazing to see just how many cycles debating them can consume. So much time and effort spent sweating over an aspect of the site that almost literally doesn't exist for me!
Victor Vance (18202) on 12/16/2015 9:46 AM · Permalink · Report
[Q --start MAT wrote--]Negative votes should not exist, because that would discourage reviewers and contributors, and it really isn't friendly community when someone sees that many people rather his/her review as bad. Instead, such reviews simply wouldn't have positive votes for being helpful, so noone would know if that's because nobody read them yet, or just didn't rate them as helpful.[/Q --end MAT wrote--]
I really like the idea!
MAT (241251) on 12/11/2015 12:06 AM · edited · Permalink · Report
I like how company pages look now, but are these games developed, published, or in any way related to a company (i.e. if a company is marked as Additional Art under release info)?
I love this list, hope you won't remove it, but the links on the right seem irrelevant now, at least the "All Games" ones.
I see you're only displaying latest 50 games. Too bad you don't display all like on developer's credits page you display all games someone worked on. I wouldn't mind the scrolling, though. At least it should be openable to view all in a same format than to open it in a new table presented way.
Great work, always love the aesthetic and useful things, always had to click on All Games to see the list in the past, good that it's no longer needed.
<hr />Edit: There seems to be a bug in the listing. Check Blizzard Entertainment, Inc. company here...
http://www.mobygames.com/company/blizzard-entertainment-inc
...you will notice they have SWAT 4 (2005) listed, yet they never worked on that game and aren't listed under any release info we have on file. How did that game get connected to Blizzard? It's not the only game, and it doesn't appear on Browse All or List All game links which list 67 games in total, but your lists says 71 games, but when you click more to view all it goes to a list with 67 games.
MAT (241251) on 12/11/2015 1:06 AM · Permalink · Report
Owning doesn't mean much since we don't have ownership relation in our DB. Westwood was owned by Virgin until 1998 then by EA, but I don't think that's visible unless a company is in the release info. There are also many subsidiaries, but each have their own listings.
Either way, even if there is a relation, we don't see it, and the list says to have 71 game but then leads to an actual one with 67 games listed. Something's not right there.
Tracy Poff (2095) on 12/11/2015 2:22 AM · Permalink · Report
I like how company pages look now, but are these games developed, published, or in any way related to a company (i.e. if a company is marked as Additional Art under release info)?
Any game for which the company appears in any release info in any capacity.
I love this list, hope you won't remove it, but the links on the right seem irrelevant now, at least the "All Games" ones.
Well, the game browser lets you sort alphabetically and such, so they're not quite irrelevant. Anyway, they don't take up much space.
I see you're only displaying latest 50 games. Too bad you don't display all like on developer's credits page you display all games someone worked on.
It'd be neat, yeah, but for companies with lots of games (e.g. Nintendo with over 700) it was rather unwieldy.
...you will notice they have SWAT 4 (2005) listed, yet they never worked on that game and aren't listed under any release info we have on file. How did that game get connected to Blizzard? It's not the only game, and it doesn't appear on Browse All or List All game links which list 67 games in total, but your lists says 71 games, but when you click more to view all it goes to a list with 67 games.
That is confusing. I checked, and there's a rejected release with Blizzard on it for that game. Showing those releases is quite an unexpected behavior. I'll have to do something about it.
Tracy Poff (2095) on 12/11/2015 7:01 AM · Permalink · Report
That is confusing. I checked, and there's a rejected release with Blizzard on it for that game. Showing those releases is quite an unexpected behavior. I'll have to do something about it.
I've fixed this for the next update.
MAT (241251) on 12/11/2015 2:59 PM · Permalink · Report
And another thing, non-approved games seem to be listed as well.
See this game...
http://www.mobygames.com/game/75794
...you cannot access nor find it via search engine, but it appears on Blizzard's list of games even though it's apparently not yet approved, seeing how it has ID in the URL.
Tracy Poff (2095) on 12/11/2015 7:49 PM · Permalink · Report
Yes, that's right. My fix restricts the list to only approved releases of approved games, so it'll be taken care of in the next update.
MAT (241251) on 12/13/2015 2:56 AM · edited · Permalink · Report
There are still some inconsistencies...
...you can see Back to the Future listed on number 50 under 2010 even though PS4 release has 2015 as an earliest date.
On the other hand, here...
http://www.mobygames.com/company/cinemaware-inc
...you have Wings mentioned second under 2014 because Windows platform has it under that year, even though original release is in 1990.
The way those two games are listed on those two lists is contradictory with one another. I personally prefer the upper version, but I think you're trying to achieve the later version since I've seen it on many other pages as well.
Tracy Poff (2095) on 12/13/2015 3:41 AM · Permalink · Report
The date shown is the earliest date of any release of the game on which the company is credited. So Telltale is listed on BttF in 2010 (which makes sense, given that they developed the thing) and Cinemaware, Inc. is listed on Wings in 2014 because the 1990 release is credited to Cinemaware Corporation (which is credited with that game for 1990).
Does this not seem sensible?
MAT (241251) on 12/11/2015 1:32 AM · edited · Permalink · Report
How are games listed? Years and platforms seem to be groupped into a single entry, but using the latest year instead of the first year it got released.
Let's say a company has these games listed:
Tracy Poff (2095) on 12/11/2015 2:16 AM · Permalink · Report
It's ordering by and displaying the latest release date that the company worked on, rather than the earliest. This was intended sort on and display games by earliest release date the company worked on, limited to the most recent 50 such games. When developing, I was originally displaying the oldest 50 games instead, and I mixed it up when I swapped for the current behavior.
I've already fixed this in the development version, so it'll be fixed next update.
Simon Carless (1834) on 12/11/2015 2:37 AM · edited · Permalink · Report
Yeah, I already spotted this one, MAT, and it's fixed in dev version :) Keep the comments coming tho - they are useful.
Tracy Poff (2095) on 12/15/2015 1:49 AM · edited · Permalink · Report
Review point value proposal:
Each review has a base value of 1 point. 'Helpful' votes add a half point. 'Not Helpful' votes subtract a half point. Reviews can be worth as few as 0 points or as much as 10 points each.
For reference, out of 16470 reviews, only 51 reviews would have 10 points. 35 of these are by Oleg! The average review would be worth 1.5 points. 1404 reviews would be worth zero points.
The next step after this would be making reviews auto-approve.
Edited to fix the bogus statistics I gave originally. Forgot to recount the stats after I changed my proposal...
Simon Carless (1834) on 12/15/2015 4:48 AM · Permalink · Report
That's different! I could go with that. But I'm easy. And I do think reviews should autoapprove, yes.
MAT (241251) on 12/16/2015 2:51 AM · Permalink · Report
Then we no longer need ratings during approval from 1-5.
Also, you should reset all the votes if someone edits his/her own review. We cannot tell if someone rewrote entire review, or just deleted it and wrote "blah blah blah" literally and still kept high points.
Auto-approving shouldn't be accepted because users can write any kind of non-review crap in it and it gets approved. As for editing, approver should check every edit (see comparison to old writing), and if there were only some typo changes, approver should choose to leave or reset votes on that review, because if review was written anew, keeping old votes would be wrong and misleading.
Also, managing to earn 0 points for review is not good. That is discouraging people from writing and you can never know if the bad votes are simply personal or objective, so 1 point should still remain as minimum.
BTW, seeing how my review points could be affected by people's votes, that would certainly discourage me to ever write a review ever again, and would probably go back and edit all my reviews to blank and delete them rather than having my points fluctuate up and down all the time.
Evgenii Andzhe (18261) on 12/15/2015 5:03 AM · Permalink · Report
I can't upload new covers to several games. I have "Internal Server Error" anytime then I try to upload something.
Tracy Poff (2095) on 12/15/2015 5:09 AM · Permalink · Report
I understand that these things usually fix themselves, but does anyone have any idea what causes this? There's nothing in the error log.