How much does rank really matter? An empirical investigation.

I definitely agree that there is a unique meta-game when you have a mixed-rank situation. It's a common topic in Factions — what do you do when you're a Bronze up against a Plat mid? What do you do when you're the lone Diamond on your team? Where do you send your best Summoners: do you spread out, or do you (e.g.) run Diamond/Plat bot and hope to snowball hard enough there to save your other laners?

Another topic you bring up which is much-discussed in Factions is the leadership angle. Some theorize that there's a reliable and consistent breakdown of skill progression, such as "you go from Bronze to Silver by improving mechanics, and from Silver to Gold by learning basic big-picture strategy, and from Gold to Plat by learning advanced strategy." If that's the case, then one might expect a big burst of utility from having at least one high-tier Summoner on the team. You can test this using the dataset we will (hopefully) gather: e.g., you might use a formula that gives a boost in predicted winrate based on the highest tier player on a team, modeling them as a "captain".

I think this is part of why the question of how much various rank differences matter is interesting: there are a lot of potential factors involved. The data I hope to gather here will, I think, shed some light on these questions.

As for the knowledge factor, with LoLNexus, your enemies' tiers are known to you when you head into game. And in modes like Factions, you generally will know the tier makeup of the enemy team as well.

Some ppl are silver 1 with 100 ranked games. Others are silver 1 with 1k. Does the two player have equal skill?

If the system places them both in Silver, it thinks so. But perhaps, for instance, players with more matches of experience, though they perform about the same in regular solo queue, will fare better or worse in mixed-rank matches. Maybe more games means more flexibility, and more familiarity playing against higher/lower-tier opponents.

If we see lots of variance — if, for instance, the performance of Silvers is really all over the place — this might be something to test later. It would require asking people to enter the number of ranked games for every Summoner in the match, though, which would be rather onerous.

Right now, we have a grand total of zero matches entered into the dataset, as the project was just launched last night and word is still getting around. To help ensure it gets off the ground, I've been careful not to place too many demands on the match-runners. I think entering everyone's tier is about as far as I can push people.

Main roles: I personally know some silver ranked players who can go toe to toe with a diamond ranked one if they get their main role/champ.

Well, that's part of what the game reflects: if I super-carry when I get my main (90% win rate) and feed with everyone else (20% win rate), then my overall win rate will reflect how often I tend to get my main.

Again, unfortunately, asking for everyone's mains/best roles would drastically increase the difficulty of reporting match results, and discourage data. But I do think this is also a potential issue for investigation that you raise: for instance, it might influence the application of the dataset to Factions, where the Champion lists are heavily limited. Maybe low-rank players do better when there's a more limited universe of Champions to learn to play as/with/against.

tl;dr: I do think it's a real question. I'm honestly not sure. If you held a gun to my head and made me guess at the relative magnitude of the rank gaps to save my life, I could make a guess, but I'm not at all confident I'd be right. There are lots of theories and hunches, but not much data (that we have access to).

This is one of the big questions that this data might allow us to investigate. It might show that there really isn't that big a difference, and it's really hard to predict whether a GGSSS team will beat an SSSSB team. Or it might show that even a small difference in rank significantly unbalances the match.

Why does this matter? Well:

• Pure curiosity.
• Factions and other custom game modes care about this a lot. Let's say Demacia gets DDSBB together and Ionia has PPPGS. Is that a fair match? Do we hit Start? Do we wait another 5 minutes to try to get a more even mix? If we play it as-is, how many points should Demacia gain if it wins: more or less than it would for a mirror match?
• Duos. Sometimes you find yourself in a Ranked game with a duo. Let's say that it's mostly a Gold game, but the enemy has a Bronze and a Diamond plus three Golds, while your team is all Gold. You end up laning against the Diamond. How cautious should you be?

I want to emphasize that this isn't about beating up on people or winning bragging rights for high-Tier Summoners, or for taking those bragging rights away. It's just empirical investigation.

Your comment is a good one in that it gets at another point: not everyone plays Ranked. If it turns out that the Unranked population is a total wildcard, then that's something to keep in mind when setting up Factions matches or whatnot.

Maybe it'll turn out that Rank is just a terrible way to balance matches for Factions. We might want to start mixing in number of wins or something else. Who knows.

The "Gold Standard" would be to give each Summoner an ID and track their Factions MMR. That would be a huge burden, though, and it'd be hard to disentangle their faction from their skill level.