Machine LoLing

I'm a league of legends data scientist!

check out my data blog:

[Machine LoLing @machineloling]

@2cara2louss4u Lmk when you feel the raw power of the AP kick in. I swear it’s so fun

[Machine LoLing @machineloling]

https://t.co/jyt1irizpo

This latest patch has hit Nami hard. Mandate is un-buildable and Moonstone, Helia, Aery and Dream Maker took hits as well. We are left with a 3-4% drop in Nami win rate.

As a Nami enjoyer myself, I decided to dig into her builds yesterday to see what makes the most sense with what we've got left. I got... carried away and now I have this nice build calculator for her to share. I tried to make decent assumptions about item passives.

All in all I found two builds that made the most sense. One will surprise you. Maybe you can find another. Feel free to explore.

[Machine LoLing @machineloling]

@mihawk013 @_Beora Should be fixed!

[Machine LoLing @machineloling]

Just a quick visualization of the change in ranked distribution this year. Read more here: https://t.co/kOR5jKWhoD thanks to @_Beora for historic data.

[Machine LoLing @machineloling]

@mihawk013 @_Beora Looks like a bug to me I’ll investigate

[Machine LoLing @machineloling]

@SheepeyDarkness Indeed. This is just an interactive blog post. A full site is coming.

[Machine LoLing @machineloling]

Now champion pool designer weights pool health based on champion play rates in different ranks! And we added a cool visualization for play rate by rank!
https://t.co/lqpyRL6iA0

[Machine LoLing @machineloling]

@coyoteirl Yeah they’re pretty aren’t they

[Machine LoLing @machineloling]

We weren't quite satisfied with our match-up and synergy derived classes, so we decided to step up our machine learning game even further.
This time, we cluster using embeddings from a deep-learning model trained on damage, scaling, match-ups, and synergies. The result is the most comprehensive and accurate champion classes currently available.
https://t.co/kvbpvSaxPG

[Machine LoLing @machineloling]

@ForceurUltime I’m limited in what I can pull right now sadly.

[Machine LoLing @machineloling]

Added an individual champion compare feature. Now you can easily see which champions are most similar or different from your main and why. See the "Individual Champ Compare" tab.

https://t.co/Z67XfQK97O

[Machine LoLing @machineloling]

@Gabrielsmurfacc X limited me to 4 pictures but you can see in the app

[Machine LoLing @machineloling]

There's a lot of content packed into the last post, but I think these plots are one of the cooler stand-alone bits.

https://t.co/ZzfKsMvUdT

[Machine LoLing @machineloling]

@wo_shi_ad_who Karma is flexible due to her multiple builds and early prio. Flexibility isn’t strength though. She’s not as good as specialists at poke or enchanting and her win rate reflects that.

[Machine LoLing @machineloling]

@erenjaegerSOS like designate X many as blindable and the rest can be whatever?

[Machine LoLing @machineloling]

Champion pools exist in an inherent tension between counter-potential and blindability. If a champ is more blindable its match-ups are less extreme, but that limits its ability to punish when given counters. It make designing a pool that can do both very challenging.
Come see how your pool fares. Make your own decisions on what matters for a good pool. Find replacements. Design from scratch. I've tried to make it as accessible conceptually as possible.

https://t.co/dmBtpSBIXw

[Machine LoLing @machineloling]

@erenjaegerSOS It's actually nearly impossible to have a strong pool because of the blindability/counter-strength tension

[Machine LoLing @machineloling]

@employedlol Senior Scientist Bioinformatician, PhD was a while ago

[Machine LoLing @machineloling]

@VERYRARE_0 You can lower the pick rate threshold to a certain extent but I just didn’t have the data for sub .1% play rate. Sorry :(

[Machine LoLing @machineloling]

@paradoxic_l Try the sliders and def check the candidate tab. There’s a ton of options for pool improvement!

[Machine LoLing @machineloling]

@Ralfh_zRose Should say Mel with Nami, but yeah Nami loves mage pairings!