Jack Zimmerman

Previously Project Analyst @Deloitte | Amateur LoL Data Analyst 🎮📊 | Documentation & Systems Enthusiast 📝⚙️

[Jack Zimmerman @JackLGZimmerman]

@ipeirotis I get the feeling that summarised outputs vaguely follow constrained output formats like on X. Where people need to maximise impact of words, more implicit etc..

So they're optimising for this engagement bait type speech which gets old very quickly.

[Jack Zimmerman @JackLGZimmerman]

Research is pretty new in League of Legends, but if you're a coach and you're un-willing to:

- Send 3 recorded training sessions
- Have a 60 minute call

What holds you back? Is it embarrassment, procrastination, lack of faith in the research, or something else?

[Kabir Bubna @KabirBubna]

Calling League of Legends Coaches 🎮

Are you currently coaching a semi-professional League of Legends team?

I’m conducting a study examining how esports coaches design, deliver and review team practice. This is part of my PhD research at @QUT.

The project aims to better understand coaching in competitive League of Legends and contribute to evidence-informed approaches to coach development and practice design.

Participation involves:
🎤 Two online, audio-recorded interviews (30–60 minutes each): one before the training observation period and one after the recorded sessions have been analysed.
🎥 Recording and securely sharing approximately 3–5 upcoming team training sessions for research analysis.

The study focuses on the coaching environment rather than evaluating individual player performance.

By participating, coaches can contribute their experiences and perspectives to research that may inform future esports coach education and practice design. Participants will also receive a summary of the research findings upon completion.

Interested or have questions? Get in touch!
📧 [email protected] or drop me a DM here.

This study has received approval from the QUT Human Research Ethics Committee (10534) and is supervised by Adam Gorman and Craig McNulty, PhD

♥️ and 🔄 greatly appreciated!

[Jack Zimmerman @JackLGZimmerman]

@ninarin00 When you posting the paper?

[Jack Zimmerman @JackLGZimmerman]

@OpenAI Performative, you have a sandbox connected to a package manager with access to the internet, that would have been flagged. It was clearly a choice to see how it would break out of its containment, not an error.

They're performing ExploitGym tasks and shocked it found an exploit

[Jack Zimmerman @JackLGZimmerman]

What does perfect League of Legends drafting look like?

- Flex: All picks can be shared across multiple roles/builds within player pools. Deny information.
- Synergies & Counters: Draft balances maximising team identity, synergy and enemy counters for highest WR. Hardest part.
- Timing: Coach performs instant pick/bans, discussion happens on enemy time, with everything pre-confirmed in practice.
- Targeting: Ban weak player pools, pick to pinch them later in the series.

[Jack Zimmerman @JackLGZimmerman]

I think it's a tough comparison, when you arrive at the ring for boxing it's a ritual, out of your element, special. Player scrimming? They jump off SoloQ (with complete apathy) and into scrims.

Same environment, no profound change in mentality, not much changes and everything looks the same. There MUST be something to figure when it comes to having an abundance of access to the competitive environment unlike Basketball, Boxing, etc...

You stop caring, becomes a habit (addiction?), and I doubt you can develop. For coaching, worth researching how to control it

[Jack Zimmerman @JackLGZimmerman]

@Sheep_Esports Actually it's the money and branding, Faker can calmly lose in T1, it's familiar, he has support, and he knows the team. You get signed with a 30 million cheque you have massive expectations, in an unfamiliar environment with unknown teammates. Good way to kill his image.

[Jack Zimmerman @JackLGZimmerman]

@kimmonismus @synthwavedd I mean flash models are essentially 95+% of the full models these days anyway? Tt seems like it's becoming better to just focus on high efficiency flash models rather than then parent <n>T size models, I mean they're still useful but the price/performance ratio gets shot.

[Jack Zimmerman @JackLGZimmerman]

@Silverborn_ Doesn't the learning curriculum ("what content needs to be covered") need to be the foundation for identifying what "differently" means here?

[Jack Zimmerman @JackLGZimmerman]

Given how random the relationship looks here, the model's predicted G1 Draft WR% has little to no relationship with the outcome.

If this is where teams invest their planning, this should be the clearest signal.

If the signal isn't there, this suggests the drafting model is broken, rather than that stronger G1 drafting has no effect on winning?

[Jack Zimmerman @JackLGZimmerman]

@Dragdar_ Fair, weird responses from them though

[Jack Zimmerman @JackLGZimmerman]

@Dragdar_ Have you asked them why? Could be they didn't get value from the other tips, so they didn't bother changing the rest of their playstyle including their runes/spells

[Jack Zimmerman @JackLGZimmerman]

@Doublelift1 Start a sushi bar now.

[Jack Zimmerman @JackLGZimmerman]

@egodeath333z @DesLoverLoL @Veigar_v2 This doesn't really apply blooms, it's more rational than the norm though.

[Jack Zimmerman @JackLGZimmerman]

@DesLoverLoL @Veigar_v2 They don't do it because they think it's good, they let top-level teams to do the thinking for them and use that to spend time considering other ideas.

The main question is, is it worth assuming top-teams aren't optimal with items and spend time chasing that margin?

[Jack Zimmerman @JackLGZimmerman]

@SaturnEUW @KireiLoL More you play the more likely you get autofilled, but yeah it's different per role.