Battlegrounds ELO ranking project
Archit_1812
Member Posts: 609 ★★
@Kabam Miike, @Kabam Crashed, @DNA3000,
Account name: Archit_1812
Name: Archit
Github: GitHub
Progression: Valiant (almost)
I'm working on a project to create an ELO rating system specifically designed for Battlegrounds matches. The system will take into account not just player performance but also the deck composition during matches. Here's how it works:
What is ELO?
ELO (or Elo rating system) is a method used to calculate the relative skill levels of players in two-player games. For Battlegrounds, my aim is to extend this to multi-player, team-based matches by considering factors like last season's ranking and deck composition.
What I will be doing:
I will be developing an algorithm that evaluates player performance based on their chosen champions (out of 30 available) in a match. If a player selects mostly champions from the top half of the roster, they will gain fewer ELO points. Conversely, if more champions from the bottom half (lower champion ranked to be considered, same rank not considered) are selected, the ELO gain will be greater. Other factors may include match outcomes and player rankings.
My wish is that this ELO rating be incorporated into the game, visibly to the players, and as a medium of matchmaking as well. Players within 30 point ratings can be matched (due to my suggested K-factor being 32)
To make this project a reality, I need the following data:
Player Rankings: The last season's rankings of players to help baseline performance expectations.
Champion Selections: The specific champions selected by players during matches.
Match Outcomes: The results of matches (win/loss/draw) for accurate ELO adjustments.
If you are a developer or have access to this data, I would greatly appreciate your support. Let’s collaborate to make Battlegrounds more engaging and competitive for all!
Devs, I'd greatly appreciate your assistance.
Thank you for your time and support! MCOC community, please help this message reach the devs!
Account name: Archit_1812
Name: Archit
Github: GitHub
Progression: Valiant (almost)
Project Description
Hello fellow Battlegrounds enthusiasts, devs and GuardiansI'm working on a project to create an ELO rating system specifically designed for Battlegrounds matches. The system will take into account not just player performance but also the deck composition during matches. Here's how it works:
What is ELO?
ELO (or Elo rating system) is a method used to calculate the relative skill levels of players in two-player games. For Battlegrounds, my aim is to extend this to multi-player, team-based matches by considering factors like last season's ranking and deck composition.
What I will be doing:
I will be developing an algorithm that evaluates player performance based on their chosen champions (out of 30 available) in a match. If a player selects mostly champions from the top half of the roster, they will gain fewer ELO points. Conversely, if more champions from the bottom half (lower champion ranked to be considered, same rank not considered) are selected, the ELO gain will be greater. Other factors may include match outcomes and player rankings.
My wish is that this ELO rating be incorporated into the game, visibly to the players, and as a medium of matchmaking as well. Players within 30 point ratings can be matched (due to my suggested K-factor being 32)
My Requirements:
To make this project a reality, I need the following data:
Player Rankings: The last season's rankings of players to help baseline performance expectations.
Champion Selections: The specific champions selected by players during matches.
Match Outcomes: The results of matches (win/loss/draw) for accurate ELO adjustments.
If you are a developer or have access to this data, I would greatly appreciate your support. Let’s collaborate to make Battlegrounds more engaging and competitive for all!
Devs, I'd greatly appreciate your assistance.
Thank you for your time and support! MCOC community, please help this message reach the devs!
8
Comments
Almost Valiant? You mean Paragon?
Edit: That's seriously ALL you could focus on?!
1. What you're asking for is borderline outrageous. You're asking for the complete match history of a season of Battlegrounds. I'm pretty sure they'd laugh at me if I asked for it. I'm not entirely convinced if Crashed himself asked for that the data analysts wouldn't just kill him, eat him, and wear his skin around the office so no one would notice his absence. That would probably take an enormous about of effort to collect, and even if they did it they would not likely let it leave the building. There are privacy issues, there are game integrity issues, there are even competitive issues with allowing such data to walk out the door.
2. What you're attempting to do is incredibly ambitious. You're talking to someone who once did a phase space analysis of NPC abilities of an MMO to calculate the best fit algorithm to vectorize a custom set of abilities into a difficulty metric to directly implement into an actual live game system. Trying to factor roster strength into ELO sounds significantly harder, not in the sense of it taking more work, but in terms of finding a way to do so that would not be mathematically nonsensical.
3. Even if you overcome both of those challenges, there's the last one, which is even if you figure out how to do this magically, and even if you brainwash the developers to hand over all their data, what you're asking for is insufficient on its face. There's no way to incorporate roster strength into ELO from match data alone, because champion options are constantly changing (champions keep getting added to the game) and the meta keeps shifting. The data you get in one season that says these match ups are better or worse will not correlate to how the following season will work. A player can have a strong deck in one season, and that same exact deck could be a train wreck in the next season. How do you say "well this player had a stronger deck so his ELO margin should be smaller" when that data comes from a season where his deck strength was completely different?
Your ambition is admirable, but your are reaching for something vastly outside your grasp, and I think not being aware of how far out of grasp this is, is in part a measure of how far out of grasp it is. However, you can certainly attempt to prove me wrong. All the data would tell you is who wins against who and at what statistical rate. Okay, make something up. Make up match data based on some reasonable guess, and then design your system. Make the data inputs into the system, such that if you ever were to get such data, you could plug it into your system and it would then adjust to that changing data. You'd have to be able to do that anyway, given the issue of changing metas and champion availability. So build it first, and then see if you can convince anyone to give you the accurate data to refine it. Maybe if you could present something to the devs, they'd listen. I still think you're not going to get the data, but at least you could have the conversation. Right now you're coming to that conversation empty handed.
I tried to make this super elaborate game with pieces and rules and complex interactions, and of course this was a complete disaster. It was incomplete and unplayable. The guy in my group with the best game made something super simple. You could describe the game in two minutes. You could play it with a simple board he sketched out on cardboard and simple pieces. And it had a very simple premise. I forget all the details now, but the concept was a game where you attempted to hijack planes flying about the map. By mutual agreement, we all played his game, and only his game, for the entire period.
Everyone has to make this mistake at least once; many have to make it multiple times until it gets drilled in. You cannot make the big thing until you've made the little thing, at least a hundred times. We all want to make masterpieces, but no one is ever going to care about your failed masterpieces. They care about your successful doorstops.
The first thing I ever did that caught an actual developer's attention in a game I was playing was when I spent a week figuring out that one little ability in the game was incorrectly set to the wrong duration, and I could demonstrate this both with testing and with comparisons to other similar abilities. Who cares, right? Well that led to me looking at other abilities, and looking at how to properly measure timing in the game precisely, and eventually becoming the expert at precise measurements for how fast things to emit damage, and that led to working with the devs on optimizing offensive balance, and combined with my work on calculating defensive abilities led to me being asked to work under contract to design a way to measure how much rewards a particular game system should give to players based on how difficult the content was.
There's no way I or anyone else can just walk into a dev studio and say "hey, give me all your data and I will fix your reward system for you." Never going to happen. And yet, and I'm still slightly amazed about it, that's kinda the opportunity I got, but it took several years of work to get there.
I did not get the attention of the MCOC devs by asking them for stuff. I did it by presenting them with the work I had done, given what I had available to me, and what I could do with the time I could spend. I invested the time before they had to risk investing theirs. And I started small, because my goal was to demonstrate my judgment on the small stuff was valid, so maybe my judgment on the medium stuff might be worth something, and then perhaps it might one day be worth spending time on the big stuff. Or at least, moderate larger stuff.
I think there are a lot of people who play games, who have aspirations to make them. But I think most of them want to jump straight to making whole games, or whole systems within existing games, without refining their knowledge and skills on the little stuff first. That stuff comes with a lot of valuable lessons: how to look, how to analyze, how to do the work and be complete, and even how to talk to actual game developers. Do something to catch a developer's attention, however small, and it might provide you with an opportunity to talk to one, to see things from their perspective, to learn how they see what you were looking at. In my opinion, this is invaluable experience if this is something you have any desire to pursue one day.
It would seem far less far fetched if someone like you who has built a reputation.
I've seen some crazy requests in my life and this one is up there. Right up there with the customers who think they can talk to Warren Buffett because they're mad at our company. Like if hes just sitting in a office behind customer service with unfettered access.
But I guess 60% of the time, works every time- Brian Fantana.
*it is Elo not ELO - it’s the man’s name not an initialism.
**we got our data through people submitting screenshots and it was only Elo among those who chose to participate - still has serious flaws (low reporting of losses for one thing)
I took the time to explain it all, precisely because from my perspective this is likely a non-starter, but it actually takes some knowledge and experience to know what is and is not feasible. It is the lack of that knowledge and experience that makes this sort of thing sometimes unavoidable. I don’t fault people for reaching for the stars, they just have to be realistic when reality catches up with them.
I will say however, and this is just my opinion, it is probably not the best thing in the world to say “hey devs, I need your assistance to make a change to your game.” There’s a bit of unintended presumption in that specific wording. I would only say that to a developer I knew well enough to get away with it, like that time Crashed assisted me with the redesign of the Battlegrounds system.
I think you may post on another subforums the Marvel Contest of Champions Content Creator Program or even try to get some Kabam emails and send to them.
Here on Forums you can see, there's a lot of kids with the only thing in mind, to disagree.
Gee, I imagine what that would be like. The world is governed on data. The fact that DS Analysts, Machine Learning Architects and Artificial Intelligence Programmers have JOBS is the fact that data is accessible. Take the Elo system I made for Formula 1 (on my GitHub page - link in the post) as an example: 74 YEARS WORTH OF DATA is accessible because of the fact that someone did in fact, to paraphrase your own joke of sorts, walk into the FIA store and ask for copyrighted material, credit, of course, given where credit is due.
As DNA said what you are proposing in terms of deck strength and draft choices isn't practical, nor do I think it would make sense for us to pursue from a design standpoint.