- NA Servers
- Player Behavior Design Values: Reward
- Patch 5.2 Jungle Changes
- Who Does Riot Balance For
- Single Posts
TL;DR: Weâ€™re actively partnering with ISPs across North America to build a dedicated highway for League traffic. So far weâ€™ve built what we call the â€œnorth bridgeâ€, which services connections in the northern US and Canada. Work is ongoing, and the goal is to create a unique network dedicated to transporting League traffic across the shortest, most stable route possible.
Riot Ahab here again with a mini-update on whatâ€™s happening with the new dedicated League network weâ€™re building in NA. In case you missed the original announcement and Q&A we posted earlier in January, you can check it out here.
These updates will serve as a peek behind the curtain as we continue to build out this network across NA. We canâ€™t share all the nitty-gritty details due to contractual complexities and security concerns, but we can provide a picture of what weâ€™re doing as we go!
For this first update, letâ€™s explore an overview of what weâ€™ve already built up to this point: the first half of the network, which weâ€™re calling the â€œnorth bridge.â€ Currently, this bridge is pathing League traffic for the northern US and Canada and is built out of a series of strategically-placed PoPs (points of presence) with direct connections between them.
Whatâ€™s the â€œnorth bridgeâ€, and what the hell is a PoP?
Think of it this way: if this direct network is a dedicated highway for League traffic between you and the servers, think of the PoPs as â€œon-rampsâ€; they get your data on the highway to the servers in the most direct route as physically possible.
These PoPs consist of two elements:
- Physical hardware: servers and switches placed in key cities
- Peering agreements: contractual agreements set up with ISPs to funnel traffic directly to the closest PoP
The first part (physical hardware) is pretty straightforward: buy some server space and make sure the on-ramp allows League traffic to hurry along onto the highway.
The bigger challenge, as is true with real highways, is making sure local traffic finds its way to the on-ramp–otherwise the highway does little good! This is where peering agreements come in: these partnership agreements mean ISPs now redirect League traffic off of their networks and toward the nearest PoP.
Once weâ€™ve made a peering agreement with a local ISP (which for big ISPs, like any contractual arrangement with a big company typically takes a fair amount of lawyering back and forth), thereâ€™s often a decent amount of necessary fine-tuning with that ISP.
This is important because different ISP algorithms can send traffic bouncing around before it loops into one of our PoPs, much like poorly laid-out street signs could send you driving several towns over and back again before finally finding a PoP on-ramp. Which means on our end we have to track down inefficient pathing routes and work with the ISP to redirect that traffic to make sure itâ€™s taking the most direct route to the closest PoP. It takes a little work, so bringing a PoP online and securing peering agreements doesnâ€™t always mean that everything is working as it should from the get-go. Plus, ISPs change– things break, they merge with other ISPs, split up, new ones may come, and old ones may go, which makes peering and fine-tuning a constant, ongoing process.
With that in mind, hereâ€™s a look at the direct network as it exists today:
Red dots represent live PoPs weâ€™ve set up so far. In those areas, weâ€™ve installed hardware and are actively working through agreements with local ISPs. Grey dots indicate PoPs that are still in-progress. Weâ€™ll post additional updates as those grey dots come online.
This map doesn’t show the direct paths connecting the PoPs, but if you connect the red dots you can see that the north bridge is largely complete. We still have plenty of peering contracts to negotiate with ISPs in these markets so we can make sure local traffic is taking the best possible path to its local PoP, but weâ€™re off to a good start. Typically, we first target the biggest ISPs in any given market because many major ISPs often provide traffic peering routes to smaller local ones (and youâ€™ll find a list of ISPs that weâ€™re currently partnering with at the bottom of this post). So if youâ€™re on a smaller ISP, your connection could likely peer through the larger ISP companies we have agreements with and already be on this new network!
So what does all of this mean?
Connections running through this network should experience a more stable, consistent experience while playing League, since itâ€™s taking the most direct route possible to the game servers. When Phase 3 of the NA Server Roadmap deploys, this network will provide as many players as possible a comparable ping when connecting to the more centralized server location. But we’ll speak more on that in the future!
Live nearby a red dot on the map above? Experienced any drastic fluctuations to your connection in the past few months? Let us know in the comments below. Knowing your local postal code as well as ISP helps us figure out how your data is being routed, but keep in mind if your ISP isnâ€™t yet on the list of current partners below, thereâ€™s a chance your data might not be pathing through our network yet.
Weâ€™ll post our next NA Server Roadmap update with what weâ€™re actively doing to deliver a consistent quality League connection as well as let you guys know once weâ€™ve started bringing the southern bridge online.
Until then, thanks for reading! Weâ€™ll stick around for the next 4 hours to answer whatever questions we can.
US & Canadian ISPs weâ€™re currently partnering with and still tuning how things work:
(note: this list is constantly growing as conversations and contracts develop, but at the time we can only list those weâ€™ve completed peering agreements with!)
- Atlas Networks Corporation
- Charter Communications
- Clear Wireless
- Cogent Communications
- Comcast Cable Communications
- Hurricane Electric
- Interconnected Associates
- LS Networks
- NTT America
- Pocketinet Communications
- Rogers Cable Communications
- Saskatchewan Telecommunications (SaskTel)
- Shaw Communications
- Syringa Networks
- TekSavvy Solutions
- TELUS Communications
- TeraGo Networks
- Threshold Communications
- Vision Net
Itâ€™s back! For a limited time, surprise yourself with a Mystery Skin!
Purchasing a Mystery Skin will unlock an unowned skin worth 520 RP or more (that could get you a discount of 84% off a skin)! As with the previous rounds of Mystery Skins, you can only unlock skins for champions you own.
Mystery Skins will be available in the Skins tab of the store forÂ 490 RPÂ and are limited to five per day. Grab your Mystery Skins fromÂ now until 23:59 on February 2!
TLDR:Â Our player behavior philosophies include punishment, reform and positive reinforcement. With rewards, we have to be careful to design systems that donâ€™t simply incentivize positive behavior for a small duration, but provide reasons to stay positive all the time.
Last time, I mentioned that Drevarius would go into a deeper dive on punishment in this blog, but with the surprise Mystery Gift for positive behavior going out to 95% of players in 2014, we wanted to take the opportunity to jump ahead and explore our philosophies around rewards and positive reinforcement.
One of our core philosophies is that there’s no silver bullet to improving player behavior in online games, and you always need a mix including punishment, reform, and positive reinforcement.
In 2014, we focused on the community and self-reflection. We ran a few experiments,Â including an exercise where players reflected on their last 10 gamesÂ and we all came to the conclusion that the one negative experience that happens occasionally should not define our community. With Snowdownâ€™s celebration of everyone coming together to be a part of the Legend of the Poro King, it was a great opportunity to deliver a positive behavior surprise.
One of the keys of positive reinforcement is the idea of â€œschedules,â€ or the expected frequency of a reinforcing event. Introducing surprise rewards unrelated to specific activities or durations is one of the most effective ways to encourage positive player behavior. The surprise element is crucial: imagine an achievement system where, if you are sportsmanlike for your next 10 games, you unlock a free skin. Players could simply behave for 10 games, unlock their gift and go back to playing the same way they were before (whether thatâ€™s positive, negative or neutral). So, instead, weâ€™ll continue to surprise players once in awhile for their positive behavior. Because players aren’t sure what the next reward is (or when it is), players will strive to be sportsmanlike in a larger range of games to try to get all the surprises.
For positive players in the game, this won’t really affect them and they’ll just get surprises every so often for being awesome. For neutral players, this effort might convince some of them to put in that extra effort in a few more games to get the next surprise. For negatively behaved players, this effort might also encourage a few to change their ways although we expect the biggest impact to be with the neutral players.
In future roll-outs itâ€™ll be possible for players to earn the next surprise so long as they’ve been positive since the last surprise was awarded. So, if you were chat restricted and missed a surprise, youÂ could still be eligible if you were positive in the time range between that one and the next.
Also, keep in mind that not every surprise will be a mystery gift. Every surprise will differ in magnitude, and be tailored to different players. For example, the last surprise before the end of Snowdown gift was an IP Boost, mainly beneficial for newer players still building out their champion pools. Other surprises may include collectibles like unique summoner icons (which some players will remember weâ€™ve tried before with the Santa Baron icon).
In the last blog in this series, weâ€™ll be back to discuss punishment as a deterrent for negative behavior and our philosophy around it. Thanks as always, and weâ€™ll see you in game!
TLDR: We did a pretty poor job ensuring you guys got the full reasoning for why we were doing this, and I really dropped the ball here. I failed to ensure that our patch notes reflected the intent of these changes, nor did I give Pwyff and crew nearly enough info to make that possible for them. I’m here to give the full context and also talk about how these changes fit into our strategy around the jungle right now.
So, the 5.1 “Smite with Charges!” change does a lot of cool stuff for the strategic freedom of the jungler, but we also knew that it would hit some of the same issues the Conservation mechanic had from season 4. Mainly, when you’re not ‘wasting’ a resource with overflow, it creates very clear windows for the jungler to go do something else with a very low cost. In other words, a jungler waiting for Smite to come off cooldown can gank or go kill wards without feeling like they’re making a deliberate trade for it.
Combined with a few junglers that could grab a camp or two and then bring out very potent ganks, this meant that 5.1 created a situation where the optimal strategy for some of our most powerful gankers was to go back to their old habits of deciding lanes before laners had even hit level 2 (in some cases). Worst of all was that failing these ganks didn’t really set these junglers behind, because many other junglers didn’t have the ability to invade and punish them after the failed gank.
The adjustments were made to hit this specifically, so that guys like Jarvan and Lee Sin had a bit more of a risky start if they tried this, as would Shaco, Xin, and Panth, even if they’re not currently centered in the spotlight.
We are aware that this change hurts some of the junglers with weak clears, leaving them with less gold for pots to stay healthy during their second clear. We know this is really painful when blind monks and dunking princes are taking over games. We know have a lot of work ahead of us. We’ve been making changes and adjustments to the systems around the jungle at a really high frequency, and that’s largely complicated the work that could have been done on the individual items or the champs. We’ve already been having a lot of internal discussions about what we need to do as the dust settles on the systematic changes. The 5.2 changes happened because we know early gankers needed to be much higher risk, and the extremely cheap jungle items were a main contributor to their reliability.
We still need to carve out more space for our tank junglers, and we are still trying to give junglers more reason to invade and counter jungle. These actions are one of the core reactions necessary to keep gankers in check, and currently aren’t functioning at the levels necessary for healthier play. Can’t say it enough, we have lots of work ahead of us still, and we’re committed to making this better. 5.3 is already very tight for changes, so 5.4 is much more likely for the next round of changes.
As an aside, we talk a lot about making precise changes that directly target the problem – I think for this one we took a more broad approach that ended up ‘hurting’ a lot of junglers, even if it was more directly aimed at aggressive level 2 gankers (Lee / J4 / Shaco / Panth / Xin). We’re aware, we’re looking into it.
We knew the items enabled a pattern, a pattern that JarvSin ( Lee Vin?) were able to opt into, but it’s not just them. The items created too low of a threshold for success for a jungler’s first few minutes. The item adjustment needed to fix that issue with the system, and would need to even if those dudes weren’t generally strong.
Also as I posted, it’s been very hard to make changes to individual champions while we’re also making so many large systematic changes. Trying to get the systems of the game stable gives us a lot more space to hit the champs we need to it, and more excitingly, start bumping up the champs that we think we’ve pushed down with the system changes.
The actual reason why Lee tends to due really well through many rounds of changes is because he’s incredibly flexible, very poorly stat bound (can do well with a huge range of builds), grants a lot of power that cannot be repressed through systematic changes. As stated above, he needs to be changed, and we know that many functions of the jungle changes won’t be realized without those changes.
I’ll throw out a few responses here.
On balance philosophy – we realize that our strategy for who we balance for isn’t well articulated. Do we balance for LCS or for Silver players? Or both? We’re trying to make sure we agree on what our actual goal is, because I think “We balance for everyone!” is a little too precious and unrealistic.
We’ll communicate our strategy when we’ve made sure we’re on the same page. Balance philosophy is a topic that greatly interests me. I don’t make the individual tuning changes personally (and you wouldn’t want me to) so I can’t provide as much context on specific patch notes, though I do help set the philosophy and direction for what kinds of changes we try to make.
Related, “win rate” gets used a lot as synonymous with balance, but that doesn’t tell the whole story. We need to communicate better about what game balance means to us, because there’s more to it than that. A lot of what you’re discussing in this thread is sort of “pre-balancing.” We’re making a change (that we have wanted to make for some time!) by removing DFG, and we know that change would have big ramifications if we didn’t adjust some champs to compensate. Those adjustments, while grounded in math, playtesting and player feedback, still often come down to educated guesses, because we don’t have the data from thousands of live games yet. A more accurate but painfully slow way to balance is to make one change, see how the live game reacts, then make another change in response to that. We think this would feel worse overall.
On design accountability — it totally exists. It’s part of my job to make sure it exists. Now, it’s not the kind of thing you’re likely to see a lot. We’re not going to publicly flog designers for making mistakes, though I can understand why that might be cathartic. On the other hand, designers (and anyone at Riot really) should be upfront and honest with you guys about when they think they’ve made a mistake.
On designers buffing champions that they like to play — this would be a big accountability issue. As someone pointed out, while we are all gamers here, this is also a place of business with a product (LoL) and customers (you guys). It’s royally unprofessional to try and tweak the game to benefit your own personal games. That’s a serious breach of trust (Rioter and players) that would come with serious consequences. We may make changes you don’t understand or don’t agree with, but it’s not because we’re trying to boost our own ELO.
On Lee and Thresh — they are champions we like, but more importantly, they also resonate with a lot of players. It’s not a popularity contest, but on the other hand, we believe a lot of why people play them is because they are fun to play, not just because they win a lot. Both champs have a lot of cool abilities, and arguably they both have so many that a) neither has a lot of weaknesses, and b) they compliment almost any comp you try to build. We’re trying to figure out ways to make them less awesome in every situation without stripping away what is fun about them. For example, last year or so we tried to tone down Lee’s mobility and ward hopping, but it felt terrible, and largely due to player feedback (intelligent, meaningful feedback, not whining and pitchforks) we reverted it.
As always, we appreciate the feedback. The more targeted and actionable it is, the more useful it is for us. I can go tell the balance team “Boards say you suck” and they would kind of look at me and say “Okay, what changes in how we adjust champions should we make?” and I would say “You just suck.” There’s not really a lot of direction for improvement there.
One of the options we’re looking at at the moment is whether we should play up Morde’s tankier side a bit more. Huge guy in a full suit of armor with a passive that generates a shield suggest tough front line dude. The removal of DFG already shifts him away from bursting down a target almost instantly, our thinking is that offering him a more sustained presence might be a good approach as a result. That’s very much an idea being explored right now though, not yet a proven direction.
Assassin Master Yi is still the most purchased skin overall due mostly to free RP. But the record for most skins purchased during release was set by a new skin in 2014. Any guesses? I’ll tell you if the most upvoted guess is right or not.
UPDATE: As of 11:45am PST, the most upvoted guesses are:
- 1) The ultimate skin that wasn’t
- 2) Project Yasuo
- 3) Dragonslayer Pantheon
and the answer is……..
- 1) Haha good one – maybe we’ll make up for it in 2015. Or maybe not.
- 2) Yes! Congrats Reddit.
- 3) Not the winner of this prize, but it WAS the most played 2014 skin in my personal quest for challenger (which sadly fizzled out somewhere in gold).
If you have any questions, feel free to ask me atÂ @NoL_ChefoÂ or e-mail me at email@example.com.