Here’s a suggestion for rework.
Make top 25 or top 50 visible on first page?
The problem is you literally cannot
- stop denying you’ll use it to condition results to your silly ideas about what is “really competitive”
- immediately start talking about doing exactly that
to the point that it’s become formulaic for responses to anyone who calls it for what it is.
How it even came up in the PTR is beyond me. If you’re going to suggest new mechanics that are probably a waste of time to implement and have nothing to do with the PTR, at least suggest them outside of PTR feedback.
Yea, I cant do that, because it isn’t true. I’m not some big known D3 name obviously. But I am in TPA clan with Iria and DiE, dmkt, Idolis, and a few others that post regularly on the forums. And I think i could confidently say they’d vouch for me that isn’t my nature…In that, I push in a fairly competitive sense…But all the same, I’m not throwing 500 keys into a single GRift (most the time).
And even if others “use it to condition results”…Who the hell cares? If people want to stop pushing because they’re content with their results at whatever paragon, so be it, thats on them. But anyone truly trying to compete or not on the leaderboards is noticeable in most cases unless they’re just plain bad. In most cases anyway, people are going to be looking for the lowest paragon cases per GRift tier, not the highest.
More often than not, I don’t feel most people are going to try to manipulate the leaderboards. Honestly, I don’t really see a reason to. Not many people really care about Rank 1 at GR100, GR110, GR…Whatever. And usually that’s usually gonna be middle of the pack anyway so it doesn’t really give anyone bragging rights.
3 Likes
My alphabet army will take over the 1000-2000 and 2000-3000 paragon leaderboards, oh wait, they already donated all their good gear to a better cause…
Beyond items, augments, etc, you’re gonna have Skill Twinks. ie. I played 5000+ hours of DH on my main account. I’m signing into my friend’s account, which is 1800 paragon, and I’m going to push the DH boards.
There is no avoiding any of that: gear or skill twinks.
I’m gonna pull out the pragmatic card… ready…?
Evaluate the system we have now, where 1800 paragon players are no where near the top of ANY leaderboard. Paragon brackets will give them a much better chance to temporarily hold a top 10 rank in a subset of the board – twinks mixed in or not.
If you are holding back your paragon, you suffer on the unfiltered primary and set leaderboards, and to be honest, you may not even show up at all. I believe what Iria is saying mostly applies to the filtering of set leaderboards.
Let’s talk about that further.
Any player has the right to invest 100% of their time accumulating bounties, re-roll perfect gear and stay in the bracket they desire. Results don’t have to be organic, the filter parameter is paragon, not main stat, or gear rolls, or the player’s experience.
In low paragon brackets, how much better can the player’s gear really be?
150 vs 120 augments = 30 * 5 * 13 = 1950 main stat.
Primals vs Weakest Ancients = 10200 - 8525 = 1675 main stat.
That 1800 paragon player will have 25917 vs 22292 (~16% more damage) or 1GR more power from the main stat alone.
Missed or poor gear rolls, 8% CHC, 25% CHD will definitely swing the results by another 1-2 GRs. But that’s where key investment and RNG can really shake things up.
Most player will just want to play the game and keep empowering their character.
If I were drawing the line, I’d say the paragon brackets are very healthy for this game, where the average top 10 ranked player is holding at least 5k+ paragon.
I’m all for them.
3 Likes
a paragon filter would be great.
4 Likes
No one, which is why no one wants to see dev time wasted on that except for the people using it as such.
If you’re so carefree with the language that you can just define what “real” competition is, you don’t need the dev team to support it.
The only reason I haven’t “provided anything valuable” is because I don’t agree with it, which I guess is perfect form for people who want the developers to reinforce their ideas about proper play with some toggles - toggles which absolutely have no merit beyond creating a kiddie pool “cheaters” will still absolutely cannonball into.
The issue with the suggestion hides factors like augments, primals and better gear.
Having the current system would be more desirable. They only need better filtering for hybrid combinations.
The leaderboards will show the current clear by set.and if you hover you can already check the gons. If you inspect you check the gear. So in the end it’s better rely on the current structure than from gimicks for ego. The leaderboards from set will help to proper gauge the set capabilities from folks who puts effort in there. That Will generate data that could later used to balance purposes. It’s like giving devs capabilities to check top 5 in each build from each class and have better gauge of the state of the game, same applies to players. Maybe that’s why the next season will be theme-less
Hovering to check paragon has already been touched on by me in this thread. It’s a bit time consuming and hard to read paragon with portraits overlaying the paragon to read it accurately. Checking hero details just to get a better read what their paragon is, is just more time consuming having to do that. You also commented about how this is affected by augments and gems, which also relates to your comment about checking items/gear to basically do the same thing…So it’s kinda a contradiction. And I agree, the set leaderboards will help filter out data you don’t want to see, and I approve of the set leaderboards. But regardless, probably 90-99% of the top 1000 for DH leaderboard is still going to be GoD6 (assuming).
2 Likes
Well, about the portrait overlaying gons I never had issues with that, even with several portraits. Some can have “similar” colors but never had issues. So if you had issues with that, I would suggest a feedback for a UI improvement to solve it because in that scenario most likely will cover the numbers for the player who uses it too.
About gons, augs and the primal filters. To gauge something, that like you said lacks differentiation if the player are inclined to check their gear either way.
One thing it’s press search by set and see like 10 rows looking for patterns without even clicking(the filter came from user side from 1k registers), you could check the their average gons. The other it’s fine tunning several filters to filter the resoults to a max 50-100 lines(filter will burden the server to process the filters selected). Because rare cases will be more than 10-100 folks at same gr clear. Most results will return 0 to (half) of registers already showed, meaning often will be less time consuming and less resource spent just making the user to check the GR level on those clears. They could as “client side” a “find me button” in the 1k list(as UI option) to help at least find the nick or the GR completed, that’s way easier to do because those 2 things are “almost” plain text on the user’s screen.
So if the “filter” thing really are needed, will not stop with those parameters you asked.
Gons, ancients, primals, augs, gem levels, primary stat, crit chance, crit damage, elemental damage,cooldown, resource cost, follower gear, gr clear level and stuff like that. It’s a feature that could take a while to implement in a proper way, and yet could still lack aspects. If the player have all those filters still will check the gear and most of the playerbase wouldn’t even bother with most of them. It’s a kind a feature that could take a while to put in the game while the “general value/usage” wouldn’t be that huge.
I’m didn’t even started to talk about data, that most likely they already have. I’m talking about filtering that data in some ways to generate proper information, even the d3 website don’t filter that well when you consider folks profiles. They even had issues and bugs just to starting implement the filter “Class sets” feature, shows a glimpse of how much work and resources they would need to spend for it and how trickier can be. The further specialization on the data will require even more time and work and could still be imprecise from several factors like active gons/inactive gons and the RNG the game has.
If you mean this post:https://us.forums.blizzard.com/en/d3/t/please-add-paragon-filter-to-leaderboard/29351/3.
A feature like that or you don’t go too deep or you will need to go really deep, any “half” path will be not precise to extract any accurate data with several false positives (like on PTR folks using sets bein on no sets category), even the most “precise data” could be have a huge margin of error. So rather rely on a huge margin error data with a ton of filters I would take my odds on minimal categories just splitting in things that will really matter, the sets with a reliable filter about it(like 6-piece bonus check).
Think GR 110 filter example: Consider that 50 folks solo cleared on GoD, those 50 folks, around 2 would be below 1k gons, 5 around 1.5k gons, 10 would be at 2k gons and the rest above 2.5k(33).
If the game was mainly solo, those info could be useful but often it’s not, will not matter much because you would need also check for augs(yes or no), then the values would be even worst to gauge making something like: 1-2 registers at (below 1.5k gons). What accurate data it’s that? What if those 1-2 registers are some random players who just took a break from some Exp runs as Zdh and tried to increase their bloodshards pool? Will mean those 1-2 players are better in a pool of 50 folks? It’s speculation because you don’t know nothing about the perks from it.
So now take this example: In the DH leaderboards you have 90-99% using the same build, means 1-10% uses the other builds. That’s a plain and simple data right there. If you split those 1-10% in categories like (sets/non sets), you will get right signals already about gaps between builds, with a base of 1000 being less than 1-10% in a overall base of 1000. That’s right there it’s actual e far more precise data generating way more information, while the first case it’s just a wild speculation.
For UI/UX, less it’s better. That’s why putting too much stuff could confuse or be just a huge amount of time spent in a feature that will be not that heavily used.
That’s why DH needs some changes, it’s not healthy to have top 1k overall roughly equals to a single build or set. Because classes have around 5 sets and LoD builds. If in 1k rows you have 10-100 rows without it, means bad balance and changes needed. That’s why I upvote for nerfs or buffs if it helps to reduce gap between most used build to other ones(apples with apples), builds needs to become more a preference not about overall power, so quoting devs:
Developer’s Note: In their current form, the Solo Leaderboards are great at showcasing the strongest builds for each class but tend to only highlight a small percentage of the wide variety of builds that players have available to them. We want to empower the community to play their favorite set and be represented on the leaderboard.
I don’t think their goal, at least right now, would be that but even if they consider that kind of filter will be painsome and still not cover all needs. Not sure how much burden could put on the current structure doing so, because any query on database with ton of filters means more impact on performance to do that query, because when you put more “filters” you put more heavy lifting on the server to categorize that data.
About the burden to do that kind of trick:
Think a huge Library(books unsorted), each book has a genre, title, author, publisher, pages and format.
- If you need the number of the books, you just count the number. (all players)
- If you need the number of books by genre, you will go through all books then split those from that genre. (each class) - Current Leaderboards
- If you need to know the books from of certain author in certain genre, you would go through all books, split the ones from certain author, then split from certain genre. (each build) - PTR Leaderboards
- If you need the number of pages from a books of certain author in certain genre, you go trough all books, split the ones from certain genre, then split from certain author, then count all pages from those books. (GR level) - One filter enabled in your suggestion
Each “filter” means more heavy lifting, not always the heavy lifting will be the same, because each “parameter” will be run in loop in each register. A query on the database to search (GoD at certain gon, with/without augs at certain gr level) would search something like this:
Select * FROM Players where Class = “DH” AND GoD_Set = “true” AND Gon <= 1500 AND Aug = “No” AND GR = “110”
That will go at least through all players from the game, then get the ones from certain class, then go for certain set, then go for certain Gon, then go for if they have aug(not implicit just true or false), then go for certain GR.
That will become a massive pain if the database wasn’t implemented for it or they didn’t created flags on it to improve the search or the data handling.
Would take some seconds to return the data, but the burden in the database to do that search will be way more higher than just using:
Select * FROM Players where Class = “DH” AND GoD_Set = “true”
Then consider multiple folks doing several searchs with filters at same time, when at the beginning of each season often we have several performance/disconnect issues.
Dude it is entirely too early in the morning for SQL. Also I am p sure they will use the set id as another field rather than making it a bool.
Select * FROM Players where Class = “4” AND Set = "69"
1 Like
A simple website to be able to filter the leaderboard would simply suffice…Too many people are overthinking a measely filter to a leaderboard. Claiming leaderboard manipulation; Waste of Bliz/Devs time to implement; needing multiple leaderboards to make it work (I think Iria claimed like 100-180 or somethin, I dunno; and now apparently not enough bovines to spin the hamster wheel to generate power to keep up with the load of thousands of people filtering the leaderboard at once…
Send an email to the maxroll.gg people. They already have a “ranks” section on their website which has all the world leaderboards; they would just need to add a filter by paragon option.
As for overthinking, I’m a thinker by default; I put all my IRL stat points into INT and this is the monster I have become.
For the record, I wasn’t opposed to the idea of a paragon filters or brackets, just the side-effects from naively implementing a system. With more thinking about how to deal with the unwanted issues, a genuinely great idea could surface.
1 Like
I said that at the outset. It’s far preferable to build-a-rank-one being implemented in the game.
It’s not “overthinking” anything to understand how the thing being asked for works and point out potential problems with it.
There’s a big difference between a simple query that can be cached for all players in a region with a fixed number of results and the thing you are asking for.
We can assume that everything is kept forever but only the top 1000 results are displayed.
It’s also reasonable to assume that this is not a single table holding hundreds of millions of records since records started being kept. We’ll assume for now that every list of records we can select via the dropdowns available to us is a separate table and not something generated by operations on the entire dataset.
We can assume for purposes of example that there are not separate tables holding details on the character that cleared the run and this is all in one long row.
Finally, we’ll assume they aren’t going to drastically reorganize nearly seven years of data.
SELECT * FROM specified_record_table ORDER BY clear_time ASC LIMIT 1000
This is a pretty efficient operation that doesn’t really process the entire table of data to return the maximum 1k results we see as players for a given leaderboard.
Adding a paragon filter to that requires the entire dataset for the specified leaderboard to be processed to return the 1000 results that meet the conditions specified - trivial at the outset but increasingly demanding as the record period goes on, potentially millions of rows being processed. In the time I was writing this post alone, likely hundreds (if not thousands) of records were added in the relative end-of-season inactivity.
There’s a reason people get payed to “overthink” things.
1 Like
So you’re saying there isn’t a way to only search / find the top 1000 (or less, cause nobody really wants to see 1000 entries if they’re looking for something more specific), and stop there…? The search doesn’t need to query up the hundred of million entries from the leaderboard as long as it automatically stops after finding (up to) 1000. This could further be narrowed down if you added more filters. Paragon: 2600-3000, GRift Tier: 130-135. Allowing far less info to be searched, reducing strain on the system in order to not search through all 100 million possible saved entries within the whole leaderboard, and beyond.
The idea that people would try to alter their playstyle because of a paragon filter is akin to the idea that having separate leaderboards for each class set means players are going to abandon the meta and compete for the top spots on underperforming sets. It’s just not likely. Most people will ultimately want the best build, the highest paragon and the fastest overall clear time.
I think a paragon filter is a beautiful idea that lets you see how you stack up against players around your level.
1 Like
It needs to parse all records until it reaches 1k, and filters are just more operations for it to run against the data.
From an end-user perspective, it’s less valid results, but from the system’s perspective it’s just more work.
An easy way to explain it is as shown above, the system doesn’t do much actual work. It just grabs the first 1k rows available ordered by the parameter given. It’s not exactly 0 overhead, but it’s effectively no different than grabbing the first 1k rows you see.
It’s kind of like “grab one thousand candies from this bowl” vs “grab 1000 candies from this bowl that are 1-2 grams in weight”.
tl;dr adding more filters does more work, not less. The system has to parse more records to find those that meet the criteria of the filters, it just shows less.
The really have backups. That’s why sometimes some progress we do are lost and they revert a previous state. For me happened a few times on those years. They have some kind of update checksum and some kind of previous state. Not sure how much states they save, but for sure they have more than one for each player. The “most older” stated I already had was around 38 hours before.
There is a way to do that, if your query have those info. But doesn’t mean they need to do that. Third parties could. There is a difference in client side search to server side search. If the filter it’s really needed the most wise decision would be client side. If you check the current leaderboard in game you already see the name, the gr and the time. So if you want filter gr as plain text on client side wil be like (CTRL+F) on browser. If you hover on the player name you receive details like visual and their gon. That not sure if would be implicit or via “load on hover”. If was implicit the same CTRL+F system could work.
Client side only works with the “received” data and the burden will be on client side. The “load on hover” would request that info on servers again, so if you implement multiple requests from it will burden the server.
That’s why I said what I said. So if you can ask all data once in a interval like 1-3 hours you could work with that, D3 planner already gets your profile, some outside ranking systems too. Because they ask raw data and filter themselves, both systems aren’t free to maintain and isn’t free to implement either. If you really want d3 devs spending some time doing this, new seasons would be even worst than the previous ones in terms of instability, we have at least like 4h-48h of instability issues each new season, with filters like that could increase that period at least about 50%. That’s why I’m against it, if you need a system like that you can check third parties for that, they do things like that already.
Then your friend’s account should get banned for account sharing 
1 Like