Notes — Tristan de Montebello on Ultraspeaking, Speaking as Flow State, and the Accordion Method

Four questions [Adler frame]

Q1 — What is it about? A reframing of public speaking improvement from tactical advice (manage filler words, use confident posture) to flow-state cultivation through games-based deliberate practice. Includes two original preparation methods (Accordion Method, Bow and Arrow) for structured talks.

Q2 — How is it argued? From seven years of coaching experience and the World Championship of Public Speaking journey. The argument is largely phenomenological: Tristan describes what effective speaking feels like from the inside and designs practice to cultivate that feeling, not to impose external behaviours. The games are the evidence — they work because they surface specific failure modes under deliberate turbulence.

Q3 — Is it true? The flow-state framing is consistent with psychological research on expert performance (Csikszentmihalyi, deliberate practice literature). The “treat symptoms not root causes” critique of traditional advice is well-founded: counting filler words does not fix the anxiety that causes them. The games-as-practice methodology has a strong practical claim: the episode itself is a demonstration, with Lenny improving visibly across four reps. The Accordion Method is novel and practical; the underlying principle (prepare by speaking, not writing) is supported by cognitive science on the distinction between written and spoken language.

Q4 — What of it? For someone who dreads public speaking, the actionable path is: (1) change the meta-belief from “I am a bad public speaker” to “my software has bugs, the hardware is fine”; (2) practice the three games (Conductor, Triple Step, Conviction Prompts) at low stakes; (3) for any prepared talk, run the Accordion Method rather than writing a script. The Bow and Arrow (single-sentence arrow + supporting structure) is the final delivery mindset.

Glossary

Flow state — a condition of full present-moment absorption in an activity, where conscious self-monitoring is suspended and performance is driven by trained subconscious habit. Tristan’s central claim: speaking is naturally a flow-state activity; anxiety is the bug that prevents access.

Root cause vs. symptom — filler words, monotony, and racing are symptoms. The root causes are discomfort with pausing, distrust of one’s own intensity, and insufficient trust that the brain will deliver without conscious direction.

Ultraspeaking — the games-based speaking methodology Tristan co-created with Michael Gendler. Three published games: Conductor, Triple Step, Conviction Prompts. The course is available at ultraspeaking.com.

Conductor — a game in which the speaker receives a random speech title and then, while speaking, sees a sequence of numbers (1–10) representing energy levels they must immediately match. Trains access to different emotional states on demand. Tristan’s model: energy leads → emotion follows → words fill in.

Triple Step — a game in which the speaker receives a random title and must integrate random words mid-speech without breaking the through-line. Trains resilience and the ability to stay on a chosen direction despite interruption.

Conviction Prompts — a game that provides sentence-opening prompts (“I genuinely believe that…”, “It astonishes me when…”) that must be completed spontaneously. Trains executive presence by forcing the speaker into states of conviction.

Accordion Method — a preparation technique for structured talks. Compress the talk from full length down to 30 seconds (stripping to the essential core), then expand back up. The result is internalised, not memorised: the speaker knows the structure and can recover from any deviation.

Bow and Arrow — a mindset for prepared talks. Identify one arrow (the single sentence the audience should remember). Everything else is the bow — the support that gives the arrow impact. Shift attention from what you want to say to what you want the audience to retain.

The flow model of speaking [§ Core arguments]

Tristan’s central reframing: speaking is not a conscious skill requiring management but a subconscious skill requiring unblocking. The evidence: humans learn to speak without formal instruction; we speak effortlessly with people we are comfortable with; the bugginess appears only under social pressure.

This has practical consequences for how you practise. Adding tactics (pausing deliberately, making eye contact) is not wrong but operates at the symptom level. The goal is to reduce the noise — the self-consciousness, the leaking of insecurity, the premature disengagement before the end of a sentence — that prevents the natural subconscious process from running.

The three tactical principles

Think up (not down) when gathering thoughts. Looking down signals uncertainty. Looking up signals thought. Tactical, easy to implement with a post-it reminder.
End strong. The brain disengages before the finish line (as in freestyle rap fMRI research: the brain “blows up” before the final bars). Anticipate this and use summary prompts (“so to wrap up…”, “in summary…”) to prompt the brain to complete the sentence with conviction.
Stay in character. Never leak insecurities. The audience cannot see what you feel unless you tell them. Leaking forces your insecurity onto the audience’s interpretation of everything preceding it. Staying in character creates a reinforcing loop: confident behaviour generates confident audience reactions, which generate genuine confidence.

Games as flight simulators

The purpose of the games is not to make students better at the games — it is to create the turbulence that surfaces specific root causes. A game that goes perfectly teaches nothing. A game that causes the speaker to leak, rush, or break direction reveals the exact failure mode to address. The game is the mirror; the failure is the lesson.

Accordion Method in depth [§ Prepared talks]

The problem with traditional preparation:

Dump ideas onto paper
Edit and rearrange
Write final script
Memorise

The bottleneck is step 4: humans are poor memorisers. Memorisation is like a chain — miss one link and the speaker is lost. The script is also written language, not spoken language; delivering it sounds unnatural.

The Accordion Method:

(Optional) dump ideas on paper
Speak the talk for 3 minutes — any quality, but end strong and stay in character
Reflect: what did I like, what didn’t work?
Speak again for 2 minutes (must cut 1/3)
Speak again for 1 minute
Speak again for 30 seconds → the essential core (arrow + 2–3 pillars)
Expand back: 30 seconds → 1 minute → 2 minutes → 3 minutes

By the time the speaker returns to full length, the talk is internalised. They know the architecture deeply enough to deliver it at any time from 30 seconds to 5 minutes, and to recover from any deviation without losing their place.