PCokey (NA)
: but what does 'high' mean? 10 casts per minute? 1? 2? 10000? Reallllly need some values on the y axis
We know exact values would be useful to y'all, but we made the call that we didn't want to risk the chance that they'd be useful to the no-goodniks out there.
: Is there any possible way to make it so I am alerted when Clairvoyance posts something? Math is my game and statistical analysis like this is the best stuff ever for me so I want to know how to get more!
Best I can offer is a [page](http://na.leagueoflegends.com/en/tag/insights) with all of our Insights posts.
: Just curious if someone doesnt have spell binded in normal qwer do you guys get the info still related to qwer or did you have to scrub data for instance where my q spell is binded on a and change it to a q. If the latter what was the craziest change from the traditional qwer did you see?
We get the data based on the order the abilities are defined in the game, not based on keybindings. QWER is just a close-to-universal shorthand for "spell1", "spell2", ...
: TURBO MODE: The Data Behind Your Spell Spamming
All right friends, since we love all champions equally and want all of you to have fun data about your favs, we dumped graphs for everybody into a big ol' [album](http://imgur.com/a/oXOSd). (there's a chance something got lost in the tubes during the upload, if you can identify any missing champs we'll make sure to get them added back in.)
Jeopardy (NA)
: Hey dude, that happens with me sometimes to - I guess it's inevitable when my animation cancels are just so buttery smooth that my Riven spells are always coming off of CD at the perfect times. Have you or have you not considered using your face?
I hear a shoe also works https://www.youtube.com/watch?v=mFOdSxc6cQk {{champion:107}}
: I play {{champion:28}} like a maniac, almost like her Q is always in URF mode. Or I'm raiding in WoW.
: Where is the {{champion:37}} data for URF from last year?
Haven't looked at URF, but here's Sona http://imgur.com/yLXS9at
: If it's not too much to ask, is it possible to get a chart measuring the usage of **every** champions' abilities? Knowing that Riven mains button mash even _more_ (that was possible?) in higher elos is all good and well. But what if I want to incorporate this into my gameplay by, say, finding the best usage of Galio's abilities to conserve his mana?
You had me at {{champion:3}} http://imgur.com/9odvuEe
: How many buttons do I have to mash at character select to make the cursor magically go from Iron Man to Vergil? I'm stuck in this never ending loop of picking Tony, but also wanting to win EVO. Thx bb! Actual question, why median instead of mean?
The median is generally a good measure of central tendency when you're worried about the influence of extreme values (outliers), or expect a heavily skewed distribution. I was nervous that someone like {{champion:28}} or {{champion:69}} would have means that didn't really represent the typical player of a given skill level due to the influence of a small number of super-spammers.
: Can we see graphs for Leona's spells? I bet higher MMR players use E far less often, because of how it affects your position.
Leona http://imgur.com/UXH5rsa. We almost included a section on how Leona apparently exists to go ham, since she doesn't seem to hold her engage tools as carefully as other supports (compare her E to the Thresh hook from the article).
: darn lol I wanted to see Karthus stats :P Granted the Q graph would be off the charts, BUT STILL
Karthus http://imgur.com/2XMWSRk. Luckily everything is on the charts, because we scaled the axes relative to each individual ability. TAKING THINGS LITERALLY **IS** FUNNY GUYS SHUT UP.
Asteck (NA)
: Can I ask the stats for Taliyah and Orianna or is that unavailable? Ori's stats specifically seem interesting to me as W is good for poke but uses so much mana and Q has such a low cooldown.
Orianna http://imgur.com/WEa8emV We didn't have data for Taliyah in 6.9 ;)
: {{champion:75}} No Chart could possibly measure how many Q's are used
OP delivers http://imgur.com/cjZ6xnz
: But... T is for trinkets...{{item:3340}}
: Can we see some of the actual number values instead of just graphs? And what about mouse-clicks? Some of my friends complain when I spam click back and forth, but most pros I've seen tend to spam their right-click as well.
We made a call early in the process that we didn't want this data to be useful to any bad guys, so we stripped the raw numbers in favor of the relative labels. Hopefully it still painted a fun picture for you!
Glaedr (NA)
: I think builds play a pretty big factor in this since obviously higher CDR means you get to mash out your abilities more frequently. Maybe some info on common first-buys for champs across all rankings would be something that could tie into this.
This is something we run into a lot when we want to look for patterns in the data: there's just so much to try to control for. When we're early in an exploratory analysis like this, we tend to look at aggregate data without too many breakouts. This does make certain assumptions. For example, this analysis basically assumes that there's one playstyle per champion. On a champ who can go to different roles, or one with a lot of variety to their build path or masteries, we could be washing out important differences in playstyles. Sometimes the data gives us hints (like the Trundle case), other times we don't know what we've missed until we dig deeper. If we were going to continue with this analysis, the next steps would depend on which team wanted to use the data. If the group working on items wanted to understand whether CDR is actually changing the rate of spell casts in a meaningful way, we'd probably break these charts down into different curves for end-game CDR. Alternatively, if we were working with a Champ team, we might group the data by class (Juggernaut, Marksman, Assassin, etc) to look for patterns on that dimension.
: The lower number of spell casted in lower ELOs with {{champion:92}} isn't accurate Low ELO {{champion:92}} 's are usually dead.
Good point-- I guess we didn't notice because @jeopardy's {{champion:92}} never dies. Jokes aside, it's not a bad idea to take a look at this analysis with > casts/(game_length - time_spent_dead) instead of just > casts/game_length to control for uptime. Thanks for the suggestion!
: First time I've seen a data scientist from Riot post in the forums. o: I've always been interested in it since all my internships/research had me working with big data. However, how much education is deemed acceptable to be considered a "data scientist"? I always see Masters+ for job applications whenever I see an opening. |:
If you'll still be in the market for internships next year, data science is likely to participate again. Please apply! [FAQ](http://www.riotgames.com/riot-internships/faq) If you're finishing up school and looking for a full time data science position with just an undergraduate degree, your mileage may vary. Our team ranges from B.S. to Ph.D. Independent research, work experience, and big data internships definitely do bolster an application from a more junior candidate. So it sounds like you've got some things going for you! If you think you're ready to throw your hat in the ring, [we're hiring](http://www.riotgames.com/careers/13287).
: Rioter possibly answer a couple questions? School project.
Wall of text. Sorry? 1. I work business-ish hours, about 9am - 6pm. This varies day to day, including start time, stop time, and number of hours at the office. Riot's very flexible this way. 2. My time is split between technical and organizational work. **As a data scientist, I spend time** * constructing datasets from the raw information we store, * doing statistical analysis and building machine learning models of player actions and other events in our ecosystem, * visualizing and presenting these results, * writing code to automate the above processes, and * talking with collaborators about how to interpret results and helping with strategy and decision making **As a manager, I spend time** * supporting the career development of other data scientists * recruiting and hiring * working with other folks to figure out vision, goals, and strategy for the Data Science discipline 3. Skills: * SQL * at least one programming language with extensive data processing / machine learning libraries (R and Python are most common, some SAS, recently some Scala) * at least one programming language for more general-purpose programming tasks (Python is most common, some Java and recently some Scala) * knowledge of statistics and machine learning * knowledge of big data processing frameworks and tools (for us this means mostly Hadoop, Hive, and Spark) * visualization tools (we use different combinations of R, Python, Javascript, Tableau, and Platfora) * ability to design experiments, break a complicated analysis down into smaller tasks, estimate effort and timeline for completion * communication (about progress, about problems, about results, about everything) * ability to take and give feedback (about everything) 4. I came from academics. After a post-doc I decided my favorite part of the work I was doing was solving problems with data, and that I thought I'd be happy solving interesting problems in lots of domains, not just my scientific subfield. The job title "Data Scientist" was (and, possibly, continues to be) pretty fuzzily defined, but I was vain enough to think it applied to me. A couple of places agreed with me, and I got my first gig. 5. Advice for people getting into data science (largely inspired by mistakes I've made): * Figure out what "data science" means to you, and make sure your definition is pretty close to the jobs you're applying for. For some people (like me), it means "solving practical problems using data and whatever methods will get the job done". For others, it means "advancing the state of the art in statistics and machine learning in areas that are important to my employer". Neither of these is right or wrong, but they lead to different kinds of projects and different day-to-day working styles and (often) different employers. My satisfaction comes from devising (hopefully?) creative applications of known techniques to new problem spaces. Other people get their satisfaction from doing research at the cutting edge of the field. * Find a version of the data science venn diagram and make sure you know how you fit into that space ([this](http://drewconway.com/zia/2013/3/26/the-data-science-venn-diagram) is the first one I saw). If you know you're weak in one of these areas, figure out how to get stronger or how to make the case that your other strengths make up for it. * Get used to working with messy data. If you're doing MOOCs or taking courses in a university setting, seek out additional datasets beyond the ones provided for your assignments. Datasets from homework problems are often sanitized and easy to work with in ways that real world data is not. * Get used to solving problems with your models, not just building and tuning them. The point of modeling in industry is to make better decisions. This (usually) entails building models either to gain deep understanding of something that happened in the past, or to predict how things will happen in the future. If you're used to ending a modeling project once you've minimized the test set error, you need to start focusing on how people will use the results of your model. * Think about how to explain the model building process, how to quantify the progress you made at different steps in model tuning, and how to present the parameters and predictions of the final model in a way that non-technical collaborators can understand. Develop a sense for which decisions and assumptions you've made are important to the end result, and which ones are just implementation details that can be postponed for subsequent technical discussions. * Be wary of becoming a "black box" data scientist. This means a lot of things. Don't use any dataset as if it was generated by a black box; work to find out where your data came from and how it was collected so that you can represent this understanding in your models. Don't treat any modeling technique you use as a black box; make sure you deeply understand the assumptions and constraints of the technique so that you can get the most out of your tools. Don't let your collaborators/customers think of you as a black box; help them understand your work so they can help make it better. * Learn to appreciate the contributions of your (technical and non-technical) collaborators. Doing data science well is hard, but that's true of any job worth doing-- we should recognize we're not special in this regard. There can be a tendency among folks who do complex technical work to overlook or even dismiss the work of collaborators who use less complex methods or whose contributions are very different from our own. This is shortsighted and definitely toxic to a team. Take the time to learn about the workflow of your collaborators in other roles and to appreciate the effort and expertise required by their craft. You'll build stronger relationships, you'll learn interesting and novel (to you) stuff, and you might even identify new ways to help each other out.
Khalibus (NA)
: New Ultra Rapid Fire Predictions
I just posted this in another thread, but I am really looking forward to playing my man {{champion:3}}. * His high base damage on Q means he's going to hit like a truck early on, and skill shots should be much easier to land on slower opponents. * I'm a little worried because healing for his shield is based on number of attacks so the delay on basic attacks hurts a bit, but worst case for the shield is that it adds a ton of resistances. Not sure if minions and monsters attack slower-- if not, we can still heal by walking into a minion wave with Bulwark up. * The recently reduced cooldowns on his E means he's going to be one of the best for buffing his team's move speed, and I think small edges in move speed are going to be super valuable given the more stately pace of the game. * The most important thing, though, is that slower opponents means easier multi-person ults. We just have to be very strategic about when we pull the trigger. I love the emphasis on well-considered strategic play in New Ultra Rapid Fire mode-- this looks to be my kind of game!
Crett (NA)
: I project that the most powerful Nurf champions will be {{champion:14}} {{champion:36}} {{champion:86}} {{champion:74}} {{champion:17}} {{champion:27}} {{champion:53}} {{champion:12}}
I think {{champion:3}} is looking like a great pick for New Ultra Rapid Fire. I anticipate that longer cooldowns will mean that high base damage is king, and each of his Qs will be that much more impactful.​ I love the idea of having time to carefully plan and aim each of my skillshots! This patch will really elevate the game to the highest level of competition, IMHO. Who are other people's power picks for New Ultra Rapid Fire mode?
Durzaka (NA)
: That first change is very reasonable. Although who besides Udyr has a loaded stun that would be a problem?
For a while if I was practicing CSing with {{champion:3}} against bots I would also practice tracking important cooldowns by picking an enemy team with lots of ways of interrupting my ult (something like {{champion:53}} {{champion:24}} {{champion:89}} {{champion:77}} {{champion:5}}). Also, {{champion:3}} buffs in 5.6. The kiting is gonna be more fun than ever with more Es to throw around. I wonder if there's a world where he builds {{item:3285}} ...
: Galio? Who is this Galio you speak of?
: 1. see fun thread 2. "control-f galio" 3. {{item:3070}} {{item:3070}} {{item:3070}}
: Your favorite champion gets elected to President of the United States. What happens?
1. see fun thread 2. "control-f galio" 3. {{item:3070}} {{item:3070}} {{item:3070}}
: Most ranged champions have better tools available to them for taking towers than Zz'Rot Portal already. If it becomes a must pick on all champions, then the base stats of the voidspawn can be dropped even further and more of it shifted over to how your defenses contribute to its damage. That is, pushing it even further into the realms "tank only-ness" Damage output counters Zz'Rot Portal innately in that you can clear the voidspawn and the void gates. If your opposing ranged character purchases an early Zz'Rot Portal, then you can build clear and sustain (Tiamat) to deal with it or you can threaten pressure elsewhere. By putting their value into pushing power, they sacrificed stats that could've helped them with Dragon.
Speaking as a player and fan, not a Rioter involved with design, I'm eager to try this on Galio. In my experience, he's an example of a mostly-ranged character who doesn't have great tools for taking towers (aside from his waveclear), plus he loves getting magic resist and armor on the same item.
: Patch 4.21 notes
In b4 I play {{champion:3}} in every role except ADC. Apologies in advance to my teambuilder teammates.
: "I didn't see that VU coming." or Nobody Expects the ChampUP Inquisition!
What's that you say? Send me a long email about {{champion:3}} ? {{champion:33}}
: The new Hunter's Machete is actually really close to Quill Coat, not to old Machete or Spirit Stone. It gives you 40 health per 5 while you're fighting monsters, 15 mana per 5 (that number could change, and there are more mana sources in the jungle than before), and damage that doesn't scale with any of your combat power, just with how long you're in combat with monsters. We are concerned about keeping tanks relevant, as long as they're relevant for the right reasons and aren't just overtuned chain-gankers. I will say that I don't think they're markedly worse off than before - even Nautilus can clear pretty reasonably! and Amumu/Nunu/Skarner/Maokai all seem quite good - but I could be wrong about that and if I am, we'll be following up on them.
As another example, from my limited experience Volibear is also in a really interesting place with the new jungle. He likes a lot of the items and a lot of the enchantments, and his sustain lets him stay out for a reasonably long time on his first clear. I feel like I have a lot of useful and interesting choices to make in itemization with the Thunder's Roar. For me, he's really fun at the moment. (Disclaimer: not a designer, but I have played some internal playtests.)
Lizdeath (NA)
: On Reports, Tickets, Illegal Activity, and Calling Riot Out for Bullshit
Sorry you went through that, sounds absolutely miserable. I'm going to use this as an opportunity to learn some more about our support systems. Can't promise any particular action, but I'll follow up if possible. Sorry for the ambiguity, just don't want to overpromise.
: Hello summoner! My name is Anbringehr, I'm a candidate for a position within Riot Games ( player support, I have my second interview this upcoming Tuesday). I just want to say that there is in no way one specific path into Riot Games. The paths into this culture, are as diverse as the people who work here. I want to give you some resources that could help you on your quest. **Websites** 1. Linkedin.com ( **http://goo.gl/cJOeKx**) *This website has been a tremendous help on my journey. This professional portfolio website allows you to paste yourself as a candidate directly in view of other rioters, and potential applicants. It offers a unique insight into the culture of Riot Games, as well as the experiences, thoughts, and ideas of other candidates that could provide you with inspiration.* *There are those that list their interview experiences from their first unsuccessful attempt into Riot Games (myself included, look up **Cody Bradley** on the Riot Linkedin page) on this page as well, as a means to help our fellow summoners try to ascend into the position they want.* (Aaron Li's Experience)**http://goo.gl/bpylZk** (He made it!) *This is truly a great blog done by Aaron on his experiences. It explains everything you could possibly want from an interviewee's perspective. I suggest looking it over once, twice, five times to gain a better perspective and insight on the player side of things.* **http://goo.gl/aEwD12**(Riot Games) *This is a great tool to research the company, you are going to want to read it top to bottom, so when you do come for an interview, you are able to come into the interview ready to hit the ground running, Watch Riot Conferences, GDC Vault Talks, make sure you read Developer Trackers and catch up on your Red Trackers.* Be yourself! (**http://goo.gl/ZunW3V**) *The key is to be yourself, Riot can generally smell bullshit a mile away. The objective isn't to find a Job at Riot Games. Jon Pan, made a reddit post about the hard realities of working for Riot Games **(http://goo.gl/Zj74Wn)**, I suggest you read that as well, provides a unique and fresh insight into that aspect of company culture* **OTHER HINTS AND TIPS** * Research League of Legends, Your knowledge and passion will be tested (some teams require more, some, less) * Understand **"Player Focused"** * Riot games is a Player Focused company, so, it helps having gaming experience and knowledge of other titles and Genres. Note : Your "gamerness" will be assessed. Understand that players are not customers. Realize that they are people, that they are gamers, but more importantly, that they are summoners. Just like you and me. Establishing a personal connection with a summoner you are trying to assist is the best way to provide the best first hand player experience, because we understand as players as well the problems they are experiencing. We are able to provide an honest and empathetic response because we know. **WE KNOW** I hope this helps. Good luck! If you have any other questions, feel free to ask!
Good luck on Tuesday! Player support is super important and I'm glad you've got a shot at it =) My one tip: make everything a conversation. These are your potential peers who have been in your interviewee shoes, share many of your interests, and are hoping you succeed. Conversations aren't usually one sided, and (in my opinion) asking questions of your interviewer is a good thing. Of course don't be squirrelly by avoiding straightforward answers, but consider asking clarification questions up front and maybe even ask for feedback after your answers ("Can you think of anything I missed?").
swarleh (NA)
: If I wanted to work for Riot..
We have a ton of people doing a ton of different things, so there's no one way in. What's important is to be really good at what you do, really passionate about games and the folks who play them, and really excited about being the best teammate and collaborator you can be. Some general advice: get some breadth in your first couple of years. Keep an eye out for an intersection between what you enjoy doing, what you're good at, and what you don't mind working really hard on. Try to get an internship ([maybe at Riot!](http://www.riotgames.com/NAinternships)) because learning how to work on a team, seeing how people manage projects, and building something for a customer are all super important experiences that are hard to replicate in the classroom. Some specific advice totally biased by my own experience: try to get your hands dirty with some data analysis projects. Look for interesting courses having to do with statistics, machine learning, natural language processing, or data mining. It's a good kind of thinking to bring to a problem.
: So i was in a ranked que
But then you picked {{champion:35}} and everything was awesome, right?
Prami (NA)
: Thoughts. W's cooldown is too low. The armor and magic resist channeling into AP makes Galio want to cry. Soraka needs reduced healing from all sources except for what is triggered by Starfall, to prevent abuse. W doesn't need a mana cost if it's already constrained by her health. The AOE silence is a good change, but the root sounds out of character. People are going to really really miss the mana sustain, and they will resent that it's been taken away. Mana sustain is part of Soraka's character. Increase the cooldown on W, add back on temp armor, increase the health % cost, and change the % health cost to be off of current health instead of maxiumum. Put MR shredding back on Starfall & have the effects & damage decrease the further away from the center the enemy is. Change the healing off of Starcall, on W, from flat healing to health regeneration. Remove the AP clause from her passive, and replace it with a clause reducing the healing from all sources (except for Starcall + W) by 40%. Remove the root from Equinox, and replace the effect with a boon to allies standing in the AOE (How about Mana Font?). Here's a thought. Why not make her W a channel? That way the player can decide how much health they want to sacrifice for their ally.
My first thought was also about {{champion:3}}. Then again, like 80% of my thoughts are about {{champion:3}} so I guess it's not that surprising.
Infero (NA)
: Female Onboarding: An analysis of the first time female League of Legends experience.
Thanks for sharing this, and for your hard work on it! When I see interesting stuff like this I can't help but ask for more =) So, two questions: 1. Given your design and results, what followup would you be most excited to run? 2. If you were setting the strategy for a team trying to improve onboarding, what would you prioritize highest based on these results? Thanks again!
Tahalden (EUW)
: Right on it! :-D
That kind of tool would be a great contribution-- keep us posted if it's something you decide to take on in earnest =)
Daen (NA)
: I would **love** to include patch changes in the analysis. When setting up the data I was unable to find hard dates on when patches changed in LCS, so I didn't include it.
I'll talk to the folks who put that content together and make that feature request. No guarantees if or when it'll happen, but I think it's a good idea to have the info on the site.
Daen (NA)
: An Analysis of Champion Diversity in NA LCS 2014 Summer Split
This is great stuff! Since you've obviously put a lot of time and thought into this topic, let me pose a question to you: Using your plots (or some other analysis of the data you collected) how would you quantify the "health" of the meta? There are a lot of ways to think about this, and I'd love to hear your take on it. Thanks for sharing your work!
: Why is Alistar support not more common?
I'm pretty new to the game, and until recently I hadn't really been able to find a tanky support I liked playing. But I picked up {{champion:12}} about a week ago and I'm smitten. I don't know how good he is, but he FEELS like he really has the potential to impact the game. For whatever reason, he makes me want to master him in a way that {{champion:201}} and {{champion:89}} don't. He's also forcing me to improve my map awareness, since his roams can be so good. I picked him up after seeing him played top in the EU LCS. It's pretty cool that, at least in the EU LCS playoffs, he has become sort of a mind-game pick (the way {{champion:117}} sometimes is, too). When a team picks him early, you never know if he's going to go top or support. I don't want to spoil anything, but there are probably about 3 games of Alistar support in the EU LCS playoffs, if you want to see him played at a high level. There are more games of him played top, and I still learned a lot about positioning, mechanics, and builds even when he wasn't played as support. EDIT: Alistar support games from EU LCS playoffs (no team names to avoid spoilers) [a game](https://www.youtube.com/watch?v=yFNIuh1D0r8) [another game](https://www.youtube.com/watch?v=m5k37geEggU) [a third game](https://www.youtube.com/watch?v=GlN6Y2U0v6Q)
: Riot VS Users
OP says no back-to-back posts, but I believe I'm allowed to post multiple times? If so, 191.
: Riot VS Users
Please don't fail, please don't fail, please don't fail... 247?
: Common misconception - Blitzcrank's Q is not a grab, it is actually a long-ranged fistbump. When an enemy is touched by the fistbump, the power of friendship compels them to go give Blitzcrank a big hug. At this point he generally punches them in the face.
: Ooh are you a new rioter? If so, what do you do?
So far I've proven to be really good at eating pistachios and talking about {{champion:3}}. Contrary to popular belief, the two are not mutually exclusive. Officially, I'm a data scientist. That means I science **ALL** the data. I science it _good_. Our team tries to use statistics and machine learning to help our designers and content creators understand how players use the stuff they make (LoL-the-game, champs, skins, eSports, fantasy, boards, ...). If we're doing our job well, the data, analysis, and insights we provide will help Rioters update their already-substantial domain expertise and make even better stuff for players. But, like I said, I'm new. So far it's mostly pistachios.
: I'd say you like to do damage while still helping out your team with cc and defensive abilities :)
It's like you're staring straight into my soul =)
: You're a loner with a heart of gold. Your loyalty and love for others makes you powerful; you have a purpose and you will not accept defeat. While your strengths include putting others first, you often forget to take care of yourself. You internalize problems and you are constantly searching for peace.
Thanks so much for weighing in! Great job starting this thread and keeping up with it, it's a lot of fun to read.
: You seem righteous and diligent, and while that may end up to be your shortcomings, it is also your greatest strength. Wow, the more I write these the more they start sounding like horoscopes :P
Well it looks like you've got a future in the horoscope business if you want-- these are great! Thanks again =)
: What do your mains say about you?
For my first post as a Rioter, I'd love to hear what you have to say about: {{champion:3}} {{champion:99}} {{champion:10}} Thanks for doing this!

Riot Nitsuabo

Level 39 (NA)
Lifetime Upvotes
Create a Discussion