AI summary by TwinMind
New Recipe for Pi - Numberphile
Numberphile
14m 28s
Historical Pi Series Representations
- Pi series expansion history dates to late 14th/early 15th century Indian mathematician Madhava from Kerala
- Founded Kerala school of mathematics
- Discovered Pi/4 = 1 - 1/3 + 1/5 - 1/7 + ... (alternating series with odd fractions)
- Wrote formula in verse format: "diameter multiplied by four and divided by unity is found and stored again..."
- Sometimes attributed to Leibniz but Madhava came first
Ramanujan's Advanced Pi Formulas
- Ramanujan developed ~17 different pi series (exact number uncertain)
- One formula for 1/π: 2√2/99² × Σ(k=0 to ∞) [(4k)!(k!)⁴ × (26390k + 1103)] / [396⁴ᵏ]
- Dramatically more accurate than Madhava series
- Based on "Ramanujan-type formulas" structure
Modern State-of-the-Art: Chudnovsky Formula
- Since 1998, Chudnovsky formula is state-of-the-art for pi calculations
- Formula for 1/π: 12 × Σ(k=0 to ∞) [(-1)ᵏ × 6k! × (545140134k + 13591409)] / [(3k)!(k!)³ × 640320^(3k+3/2)]
- Based on Ramanujan's insights but much more rapidly converging
- Used for record-breaking calculations: 105 trillion decimal places (last digit was 6)
Performance Comparison (First 4 Terms)
- Madhava series: within ~0.2 of π
- Ramanujan series: within 10⁻⁴⁰ of π
- Chudnovsky formula: within 10⁻⁷¹ of π
- Shows exponential improvement in convergence rates
New Pi Formula by Sena and Sahar
- Two Indian string theorists discovered new pi representation
- Formula: π = 4 × Σ(n=1 to ∞) [1/n! × (1/4)ₙ₊λ₋₄ × 4/(2n+1) × (2n+1)²/(4n+λ-1)]
- Uses Pochhammer symbol notation: (a)ₙ = a(a+1)(a+2)...(a+n-1)
- Contains parameter λ that can vary, creating infinite family of series
- Different λ values give different convergence rates
- As λ → ∞, reduces to Madhava series
- Optimal λ values appear to be between 0-10 for first few terms
Media Reception and Criticism
- Generated controversy online due to media overhype
- Authors never claimed to revolutionize mathematics
- Formula performs better than Madhava but nowhere near Chudnovsky performance
- Discovery was side result from string theory research, not main focus
- Paper actually about string theory, not pi calculations
- Authors felt misrepresented by media coverage
Technical Performance Analysis
- λ = 3: within 0.001 of π (first 4 terms)
- λ = 1,000,000: within ~0.2 of π (equivalent to Madhava)
- Convergence rate depends on truncation point and λ value
- No universal optimal λ value identified
- Performance varies with different λ choices but doesn't approach modern methods
