What is everyone’s thoughts on which ranking system should be used?
Tiers if you want to be modern compared with other games, 1-50 if you want it to be “Halo” ranks.
I voted tiers just because people get too worked up and heated when they want to get a higher rank on the 1-50 scale.
Tiers, has less trolls and obsession.
Ok but what are your thoughts on the two top tiers? Pro and Semi-Pro? Being limited to only so many players?
> 2533274830324810;4:
> Ok but what are your thoughts on the two top tiers? Pro and Semi-Pro? Being limited to only so many players?
It’s better that way. 1-50 was more grinding than anything else h5 tiers are actual true skill ranking. why would we need 50 ranks for true skill? It’s unneeded. Tiers is simply better in every way possible.
The teirs while not perfect Is more competitive then halos 1-50, Player who make it to semi and Pro have to keep playing constantly to hold their ranks, halo’s 1-50 mean players can boost and trade 50’s which was a rampant thing in halo 2 and more so in halo 3.
Players have to keep playing and performing well to keep their ranks and they update in real time,
1-50
It’s just more steamlined. The tier system always comes off as obtuse IMO and doesn’t make you crave ranking up as badly.
> 2533274830324810;4:
> Ok but what are your thoughts on the two top tiers? Pro and Semi-Pro? Being limited to only so many players?
It’s not good. There will always be cheaters. It’s inevitable. Those players will have max rank. With limited space at the top cheaters now actually effect the entire ranking ecosystem by making it harder for legit players to get one of the now even more limited spaces. With 1-50 cheaters don’t effect you getting a 50 unless you physically play against them.
I personally like the 1-50 but i could be happy with the tiers if they were all even (as in no limited space tiers) they need to be obtainable by anyone with enough skill. If anything the Pro and semi pro could be a leaderboard within the game.
1-50 has the problem of when you hit fifty, you can find that there’s actually a pretty big skill gap between 50s. In the tier system, pros and semi pros are ranked directly against each other so you should know where you stand a bit better.
Realistically, the aesthetic form of a ranking system is completely irrelevant. Really, the only thing that matters is how well the rank corresponds to the player’s actual skill, and how well it matches the player with other players of similar skill.
That said, I think any finite, non-relative system (e.g. 1-50) is detrimental to the true purpose of the system. Unless the ranks are very wide apart, there are always going to be players whose skill is higher than “50” or whatever the highest rank possible is going to be. All those players will be stuck at that level, and the ranking shows that they are of equal skill, even though in reality there might be enormous skill differences between those players.
The positive aspect of the tier system is that at least the highest tiers can only have a limited amount of players. So that when sombeody comes in, someone else gets kicked out. That way the tiers are at least relative, because the “value” of that tier depends on how many good players the game has.
Ideally, the visual outfit of the ranking system would just show a number between minus infinity and infinity, depending on the player’s skill. Zero would be the mean, so anyone with negative rank would be below average, and anyone with a positive rank would be above average. The scale of the system could be such that 100 points would equal one standard deviation, so that someone with 200 points would be within the top 97.5%, and somebody with 260 would be in top 99.5% and so on. This way everybody would know their skill accurately, regardless of skill level.
> 2533274830324810;9:
> I personally like the 1-50 but i could be happy with the tiers if they were all even (as in no limited space tiers) they need to be obtainable by anyone with enough skill. If anything the Pro and semi pro could be a leaderboard within the game.
But they are, aren’t they? I don’t remember well, but wasn’t the semi-pro something like the best 200 players? Whatever it was, the point is, anyone who has enough skill to be within the top 200 can obtain that tier. It’s only the players who aren’t good enough who can only dream about those tiers.
1-50. It’s more familiar to those who are used to XP leveling, but it also tells them that they actually have to win to go up.
I would say tiers, but they seemed a bit broken during the beta. I mean, reaching something like onyx and skill being determined from 10 games seems messed up. The tiers also don’t go of your skill alone, more your skill with a random team. Your skill should be based of a multitude of things, such as kd ratio, win loss ratio, etc.
The way it is right now, little jimmy could be carried for 10 games and get a good skill rating for his team winning. Then bob could win some games and lose some games, while keeping a positive kd and scoring high/performing well on his team for every game but get a lower tier compared to little jimmy.
> 2533274898131165;13:
> I would say tiers, but they seemed a bit broken during the beta. I mean, reaching something like onyx and skill being determined from 10 games seems messed up. The tiers also don’t go of your skill alone, more your skill with a random team. Your skill should be based of a multitude of things, such as kd ratio, win loss ratio, etc.
>
> The way it is right now, little jimmy could be carried for 10 games and get a good skill rating for his team winning. Then bob could win some games and lose some games, while keeping a positive kd and scoring high/performing well on his team for every game but get a lower tier compared to little jimmy.
If you think about it that way, 1-50 is a lot worse, since it only factors in wins and losses. With Tiers you at least get a bit more experience than your teammates if you’re the MVP, whereas you get all the same with 1-50.
> 2533274898131165;13:
> I would say tiers, but they seemed a bit broken during the beta. I mean, reaching something like onyx and skill being determined from 10 games seems messed up. The tiers also don’t go of your skill alone, more your skill with a random team. Your skill should be based of a multitude of things, such as kd ratio, win loss ratio, etc.
>
> The way it is right now, little jimmy could be carried for 10 games and get a good skill rating for his team winning. Then bob could win some games and lose some games, while keeping a positive kd and scoring high/performing well on his team for every game but get a lower tier compared to little jimmy.
No one is going to get carried ten games in a row. The probability of that happening is probably somewhere around one in ten thousand. If they’re one of those “lucky 500”, they’ll end up playing against players far above their skill level, and sooner or later their luck will end, and the system will drop them to the appropriate level. Likewise, Bob isn’t going to be constantly getting matched with teammates below his team level, unless the matchmaking is very bad, and will eventually reach their skill level as well.
People never understand how ranking systems work. All a skill based rank is, is an estimate of a player’s probability of winning. Roughly speaking, if a player’s probability of winning is higher than 50%, then it means that, on average, the player will win most of their games, which means that, on averge, their rank will go up. When their rank goes up, they will end up playing against better players, and their probability of winning will go down. If the probability of winning drops below 50%, then, on average, they will lose more than win, and their rank will go down. Their rank will go down until their probability of winning goes above 50%, and so on.
The player’s rank will fluctuate around their actual skill level. The system carries some “inertia”, which is why they will always overshoot a bit when they reach their actual skill level. But over time, the overshoots will get smaller and smaller, and their rank will become more certain.
The important thing here is that, regardless of the details of the system, as long as the rank reflect’s the player’s estimated probability of winning, this will happen. No matter how the estimation is done, as long as it’s carried out in some systematic fashion (and not by dice roll), sooner or later, the right skill level will be found. The important thing here is that the system works over time. It’s not going to give a perfect estimate in the first few matches, but play a few dozen matches, and it’ll be a lot closer.
Worries like someone getting carried or someone getting bad teammates are meaningless, because while that may happens sometimes, the matchmaking system chooses which players go to which teams either at random, or more likely to balance the ranks. Sometimes bad players get carried, but sometimes they also end up losing because of their team. Likewise, sometimes good players end up with a bad team, but other times they get carried. Most of the time, however, the player will end up with a team that is approximately as good as they are.
All suggestions like “K/D and assists should be taken into account” are pointless. The reality of it is, the systems we have employed in games right now, are extremely fast already. For example, MIcrosoft’s original paper on TrueSkill tells that in 8 player FFA, the rank converges close to the final value in only 10 games. It’s very unlikely that complicating the system with extra variables would bring any more than marginal improvements to the convergence properties, and that is if they were introduced properly. There’s also the possibility that introducing them in the wrong way would make the system worse.
I haven’t seen a good argument for why calculating win probabilities would be an incomplete way of predicting player skill. They say individual skill doesn’t matter in team games because of it, but what they don’t realize is that unless a player plays with the same team every time, they will get equally matched with teammates worse than them, and teammates better than them, so, while one game doesn’t reflect their individual performance, the average of multiple games will always reflect their performance.
The reason players often perceive ranks in games to be bad is because the games often have sloppy matchmaking systems. If you get matched with players who are much better or worse than you, chances are it’s because the matchmaking system prioritizes search times over match quality. If you don’t want to see players get carried, or carry other players, you should ask for a more patient matchmaking system. But the flip side of that is that with good match quality comes long search times, which most players also don’t want. It should be up to the player whether they want good matches, or fast searches.
I’d prefer tiers providing they sort it from the beta. It was so very broken. I figured in the beta that you’d need to have 4 wins for every loss to make any progression at all.