Case Study On HTT And URF: Problems And Solutions For All Future FGMs

Kholdstare13·May 31, 2015·2 votes·2,598 views

So, after seeing the biggest failure points on both URF and Hexakill TT, I couldn't help but think that both of these modes would be able to enjoy more longevity, or be able to fill in a more frequent rotation with just a few simple changes.

Problem: Staleness Big in URF, shows up a bit in HK. The game just gets stale. And why? champion variety. When you can expect a handful out of the same dozen champions in every single game, you get sick of it.

Solution: Auto-Ban Rotation The original URF almost had the right idea. Banning certain champions that were overused is moderately heavy-handed, but it works. It's not that they are OP, it's just that the staleness of seeing them every game, because all the little nooblets thing they are OP, gets frustrating fast. The solution: Every hour, the server updates their statistics for pick rate (over a large period, not just that hour). Any champions picked a certain percentage more than the median are put into the auto-ban list for the next hour. Custom games never use the auto-ban list. Basically this just forces people to keep things fresh. After a champion is banned and other champions are picked for a while instead, it will come back off the ban list, and only go back up if it is overplayed during its uptime. The more overplayed it is, the harder it will be for that pickrate to come back down below the threshold, so it may sit in the banlist for a few cycles.

The overall effect simply reduces stale picks.

Problem: Matchmaking This is the big one on HK TT. When you take every single player, and default them to the exact same elo, there is never a good outcome. Trolls, AFks, feeders, so much of it all mixed into 1, that it forces pure entropy, continuing to drag good players down, boost bad players up, and reset everyone back to that crab bucket default again. In a temporary mode, matchmaking will never have enough time to settle, and every game will be terrible.

Solution: Persistent MMR. I know featured game modes are all different, but there are general mechanics that players have that can translate to multiple things. My solution: provide a hidden "featured" MMR that persists on every account. During every FGM, matchmaking is done using the average between the player's featured MMR, and their normal MMR on the map being played. If the player is an expert on a map, they will be matched above their featured MMR to compensate for that expertise. If they are good at FGMs but bad at a map, they will be matched below their featured MMR to allow room for mistakes caused by not knowing the map well.

Problem: Snowballing What fun is a game when they enemy can win through sheer numbers, numbers usually caused by mistakes from your weakest teammate?

Solution: Dominionize Lower kill gold/exp, higher passive gain.

Games are meant to be a constant struggle against your opponents, relatively evenly matched at all stages of the game with skill as the deciding factor. Let skill be the deciding factor. "fed" is a concept that should not exist. Doing well is a reward unto itself, with the progress it provides towards the game's objective. There's no need for it to turn the remainder of the game into a free win/auto-loss. A lot of negative attitude comes from people giving up early on when the enemy gets a small lead. These are obviously SR players conditioned by the horrors of that map to expect their efforts to be meaningless from then on (and it's a self-fulfilling prophecy since they drop all their effort at that point), It's time we re-train them. Dominionize the power curve, and make it known publicly that these mechanics exist to make comebacks always a possibility, as long as you put in the effort.

These changes would help FGMs have a bit more longevity, and be more enjoyable for more people. Currently, people get frustrated with the problems of FGMs too quickly, and it leads them to have no possible fate beyond their temporary releases. With these changes, it would be possible to put in a permanent featured rotation, centered around selecting a new mode/map combination each week for a matchmade queue.

13 Comments

JedditeIsARAMJun 1, 20153 votes

There's that_** "WAH WAH WAH OTHER SUMMONERS ARENT PLAYING THE CHAMPIONS I WANT THEM TO PLAY"_ / _"WAH WAH WAH OTHER SUMMONERS ARE PLAYING CHAMPIONS I DONT WANT THEM TO PLAY"**_ non-problem in search of a non-solution, again.

RiotL4T3NCYJun 3, 20153 votes

This is good feedback, some interesting ideas (easy to read too). Let's go! ^_^

RE Staleness: We've discussed a rotating draft for certain modes before, and it's something we'd consider for future modes. It could be cool to refresh the shape of a FGM's puzzle mid-way through by rotating the champion pool. RE the bans during URF 2014, those actually were due to them being OP. Not just "regular OP", but a special toxic kind of OP that induced pretty un-fun gameplay patterns for everyone in those games. :/

RE Matchmaking & persistent MMR: Each FGM does actually have its own MMR right now. It's seeded from your Normal MMR initially (so if you have a high MMR you'll play other high MMR players) and then starts to solidify into its own separate MMR for that FGM. They also carry-over when we bring certain FGMs back.

EG: If you're playing Hexakill: TT right now, it's picked up from where you left off last time (which was also seeded from your original Normal MMR). This is also separate to your URF MMR, One For All MMR, etc.

RE Snowballing: Twisted Treeline has some inherent problems that Hexakill isn't going to be able to solve (in some cases potentially exacerbate) and you're right, snowballing is one of them. I personally feel Ascension was one of the better solutions to this with its completely uncompromising 'anti-snowball mechanic', but obviously we can't just shoehorn that into Hexakill:TT. I think you still need to be rewarded for performance, which in the context of Twisted Treeline currently = gold. Flattening the delta produced (as you suggest) might be something we could look at though. I'm also interested to see what the new TT changes coming in 5.11 do to snowballing on the map. If we do Hexakill on TT again, there'll probably be some nuanced differences.

BADxW0LFJun 4, 20151 votes

Idea to help fix the problem of staleness...make champ select random (like in aram). This forces people to use different champs all the time and not constantly use the same op champs all the time.

SachabotJun 8, 20151 votes

This would actually lead to the same problem ARAM has: creating accounts with only the champions that are strong there, so you're never at a weak point. You would have to give access to EVERY champion in order to stop this. Which may not be a bad idea, as it would allow people to try new things.

Kholdstare13Jun 8, 20151 votes

{quoted}

This would actually lead to the same problem ARAM has: creating accounts with only the champions that are strong there, so you're never at a weak point. You would have to give access to EVERY champion in order to stop this. Which may not be a bad idea, as it would allow people to try new things.

It wouldn't lead to ARAM problems since the rotating banlist would be global, but I can see a technical possibility that it could ban out every champion a limited account owns plus all the free week ones. Some extra code might be needed to allow a bonus rental in that case, or restrict queuing if the account doesn't have enough currently available champions when matched against the banlist.

MinardeJun 1, 20151 votes

All three ideas sound pretty nice. Maybe change the ban rotation to a day or something, since rotating hourly might be a bit too rapid. I'm not sure how well the anti-snowballing proposal would be taken though, since SR players seem to enjoy snowballing, despite disliking being snowballed against.

Kholdstare13Jun 2, 20153 votes

{quoted}

All three ideas sound pretty nice. Maybe change the ban rotation to a day or something, since rotating hourly might be a bit too rapid. I'm not sure how well the anti-snowballing proposal would be taken though, since SR players seem to enjoy snowballing, despite disliking being snowballed against.

Per day would be too infrequent, especially if the mode changes out every week (which would need to reset the auto-ban list). Hourly means you're getting pretty close to live updates on the pick rates, without overstressing the server. Overpicked champions get quickly removed from the selection, and should quickly return to it due to not having time to go over the limit by too much. Again, this is just a staleness prevention method, not a balancing factor. Spam pickers' perception of OP champions is almost always wrong, but it doesn't make it any less annoying to face the exact same opponents every match.

Champion SkinJun 4, 20151 votes

Something that could be fun would be to possibly get random champions... though of course it would also cause problems like getting champions you don't want or feeling like your team is doomed or some crap like this (of course swaping and re-rolling like in aram would of course help to some degree)

Kholdstare13Jun 7, 20151 votes

I'm going to say again, the snowball RUINS hexakill. It's 100% broken. People shouldn't get a free win through numbers when they're THAT BAD at fighting.

Unless you want to reduce the surrender time to 0, FIX THIS PROBLEM. Kills should not be worth triple, or passive worth 1/3 normal. 1 bad teammate shouldn't give your opponent free 1v1 wins against YOU when you're better than them. There is zero reason their champion should be doing the work for them through numbers you have no control over.

Kholdstare13Jun 8, 20151 votes

If you insist on snowball as a mechanic, make gold a win condition. If any team gets a 20% gold advantage over the opposing team, end the game immediately. Fighting against that is the most toxic experience imaginable, and playing with it is the most soul-drainingly hollow experience you could ever have. People want to win because they are good, not because the base numbers say they get a freebie.