op819_sqrt_mul_div_sqrt_sub_one_tendsto
The sequence sqrt(R) * (sqrt(R) - 1)⁻¹ converges to 1 as R tends to infinity.
Statement
theorem op819_sqrt_mul_div_sqrt_sub_one_tendsto : Filter.Tendsto (fun R : ℕ => Real.sqrt (R : ℝ) * (Real.sqrt (R : ℝ) - 1)⁻¹) Filter.atTop (nhds 1)
- Source
- step towards op819_online_single_high_efx_sharp_sqrt_law
- Verified
- 04 Nov 2026
- Axioms
Classical.choiceQuot.soundpropext
The proof, explained
The proof decomposes $\sqrt{R} \cdot (\sqrt{R} - 1)^{-1}$ into $1 + (\sqrt{R} - 1)^{-1}$ for all $R \ge 2$, where the denominator is non-zero. Since $R \to \infty$, the square root $\sqrt{R} \to \infty$, which implies $\sqrt{R} - 1 \to \infty$ and hence $(\sqrt{R} - 1)^{-1} \to 0$. Adding 1 gives that $1 + (\sqrt{R} - 1)^{-1}$ converges to 1, and since the original sequence is eventually equal to this expression, it also converges to 1.
The sequence $(\sqrt{R} - 1)^{-1}$ tends to 0 as $R \to \infty$.
As the natural number $R \to \infty$, its real cast also tends to $\infty$. Composing with the real square root shows that $\sqrt{R} \to \infty$. Adding the constant $-1$ preserves divergence to $\infty$, giving $\sqrt{R} - 1 \to \infty$. Finally, composing with the reciprocal function maps $\infty$ to 0, establishing that $(\sqrt{R} - 1)^{-1} \to 0$.
uses
Real.tendsto_sqrt_atTop,tendsto_natCast_atTop_atTop,Filter.tendsto_atTop_add_const_right,tendsto_inv_atTop_zerohave h1 : Filter.Tendsto (fun R : ℕ => Real.sqrt (R : ℝ)) Filter.atTop Filter.atTop := Real.tendsto_sqrt_atTop.comp tendsto_natCast_atTop_atTop have h2 : Filter.Tendsto (fun R : ℕ => Real.sqrt (R : ℝ) - 1) Filter.atTop Filter.atTop := Filter.tendsto_atTop_add_const_right Filter.atTop (-1 : ℝ) h1 have h3 : Filter.Tendsto (fun R : ℕ => (Real.sqrt (R : ℝ) - 1)⁻¹) Filter.atTop (nhds 0) := tendsto_inv_atTop_zero.comp h2
The shifted sequence $1 + (\sqrt{R} - 1)^{-1}$ tends to 1 as $R \to \infty$.
The constant sequence with value 1 tends to 1. Adding this to the previous result that $(\sqrt{R} - 1)^{-1} \to 0$ yields a limit of $1 + 0 = 1$.
uses
tendsto_const_nhds,add_zerohave h4 : Filter.Tendsto (fun R : ℕ => 1 + (Real.sqrt (R : ℝ) - 1)⁻¹) Filter.atTop (nhds 1) := by have h := (tendsto_const_nhds (x := (1 : ℝ))).add h3 rw [add_zero] at h exact h
The sequence $\sqrt{R} \cdot (\sqrt{R} - 1)^{-1}$ is eventually equal to $1 + (\sqrt{R} - 1)^{-1}$ as $R \to \infty$.
Unfolding the definition of eventual equality along the filter $\mathrm{atTop}$, it suffices to show equality for all $R \ge 2$. For any $R \ge 2$, the term $\sqrt{R} - 1$ is non-zero, since $\sqrt{R} = 1$ would imply $R = 1$, contradicting $R \ge 2$. Rewriting $\sqrt{R}$ as $(\sqrt{R} - 1) + 1$ and distributing the reciprocal gives $(\sqrt{R} - 1)(\sqrt{R} - 1)^{-1} + 1 \cdot (\sqrt{R} - 1)^{-1}$. Because $\sqrt{R} - 1 \ne 0$, the first term cancels to 1, yielding $1 + (\sqrt{R} - 1)^{-1}$.
uses
Filter.EventuallyEq,Filter.eventually_atTop,sub_eq_zero.mp,Real.sqrt_eq_one,sub_add_cancel,add_mul,mul_inv_cancel₀,one_mulhave heq : (fun R : ℕ => 1 + (Real.sqrt (R : ℝ) - 1)⁻¹) =ᶠ[Filter.atTop] (fun R : ℕ => Real.sqrt (R : ℝ) * (Real.sqrt (R : ℝ) - 1)⁻¹) := by rw [Filter.EventuallyEq, Filter.eventually_atTop] use 2 intro R hR have hne : Real.sqrt (R : ℝ) - 1 ≠ 0 := by intro h have h_sqrt : Real.sqrt (R : ℝ) = 1 := sub_eq_zero.mp h rw [Real.sqrt_eq_one] at h_sqrt have : (R : ℝ) ≥ 2 := by exact_mod_cast hR linarith symm calc Real.sqrt (R : ℝ) * (Real.sqrt (R : ℝ) - 1)⁻¹ = (Real.sqrt (R : ℝ) - 1 + 1) * (Real.sqrt (R : ℝ) - 1)⁻¹ := by rw [sub_add_cancel] _ = (Real.sqrt (R : ℝ) - 1) * (Real.sqrt (R : ℝ) - 1)⁻¹ + 1 * (Real.sqrt (R : ℝ) - 1)⁻¹ := by rw [add_mul] _ = 1 + (Real.sqrt (R : ℝ) - 1)⁻¹ := by rw [mul_inv_cancel₀ hne, one_mul]
The limit of $1 + (\sqrt{R} - 1)^{-1}$ transfers to $\sqrt{R} \cdot (\sqrt{R} - 1)^{-1}$ by eventual equality.
Because the two sequences are eventually equal along the filter $\mathrm{atTop}$, the limit of 1 established for $1 + (\sqrt{R} - 1)^{-1}$ applies to $\sqrt{R} \cdot (\sqrt{R} - 1)^{-1}$ as well.
exact h4.congr' heq
Real.tendsto_sqrt_atTop- The real square root function tends to infinity as its input tends to infinity.
tendsto_natCast_atTop_atTop- The canonical cast from the natural numbers to an Archimedean ordered semiring tends to infinity as the natural number tends to infinity.
Filter.tendsto_atTop_add_const_right- Adding a constant to a function that tends to infinity preserves divergence to infinity.
tendsto_inv_atTop_zero- The inversion (reciprocal) map tends to 0 as its input tends to infinity.
tendsto_const_nhds- A constant function tends to that constant value in the neighborhood filter of that value.
add_zero- Adding zero to any element leaves it unchanged ($a + 0 = a$).
Filter.EventuallyEq- The relation stating that two functions agree on a set belonging to a given filter.
Filter.eventually_atTop- A predicate holds eventually along the filter $\mathrm{atTop}$ if and only if it holds for all elements greater than or equal to some threshold.
sub_eq_zero.mp- If the difference $a - b$ equals 0, then $a = b$.
Real.sqrt_eq_one- For a real number $x$, $\sqrt{x} = 1$ if and only if $x = 1$.
sub_add_cancel- Subtracting and then adding the same element restores the original value ($a - b + b = a$).
add_mul- Right-distributivity of multiplication over addition ($(a + b) \cdot c = a \cdot c + b \cdot c$).
mul_inv_cancel₀- Multiplying any non-zero element by its reciprocal yields 1 ($a \cdot a^{-1} = 1$).
one_mul- Multiplying by 1 on the left leaves an element unchanged ($1 \cdot a = a$).
- The final step uses dot notation `h4.congr' heq` to invoke `Filter.Tendsto.congr'`, transferring the limit along the eventual equality.
Written by a model (gemini-3.8-flash) from the Lean proof and the goal states Lean reported at each step. Lean verified the proof; this explanation is a reading aid and is not itself machine-checked.
Lean source view module on GitHub
Choose a tactic — a marked line — to see the goals Lean had before and after it, as in Lean’s infoview. Step with ← → or j k.
/-- The sequence sqrt(R) * (sqrt(R) - 1)⁻¹ converges to 1 as R tends to infinity. -/theorem op819_sqrt_mul_div_sqrt_sub_one_tendsto : Filter.Tendsto (fun R : ℕ => Real.sqrt (R : ℝ) * (Real.sqrt (R : ℝ) - 1)⁻¹) Filter.atTop (nhds 1) := by have h1 : Filter.Tendsto (fun R : ℕ => Real.sqrt (R : ℝ)) Filter.atTop Filter.atTop := Real.tendsto_sqrt_atTop.comp tendsto_natCast_atTop_atTop have h2 : Filter.Tendsto (fun R : ℕ => Real.sqrt (R : ℝ) - 1) Filter.atTop Filter.atTop := Filter.tendsto_atTop_add_const_right Filter.atTop (-1 : ℝ) h1 have h3 : Filter.Tendsto (fun R : ℕ => (Real.sqrt (R : ℝ) - 1)⁻¹) Filter.atTop (nhds 0) := tendsto_inv_atTop_zero.comp h2 have h4 : Filter.Tendsto (fun R : ℕ => 1 + (Real.sqrt (R : ℝ) - 1)⁻¹) Filter.atTop (nhds 1) := by have h := (tendsto_const_nhds (x := (1 : ℝ))).add h3 rw [add_zero] at h exact h have heq : (fun R : ℕ => 1 + (Real.sqrt (R : ℝ) - 1)⁻¹) =ᶠ[Filter.atTop] (fun R : ℕ => Real.sqrt (R : ℝ) * (Real.sqrt (R : ℝ) - 1)⁻¹) := by rw [Filter.EventuallyEq, Filter.eventually_atTop] use 2 intro R hR have hne : Real.sqrt (R : ℝ) - 1 ≠ 0 := by intro h have h_sqrt : Real.sqrt (R : ℝ) = 1 := sub_eq_zero.mp h rw [Real.sqrt_eq_one] at h_sqrt have : (R : ℝ) ≥ 2 := by exact_mod_cast hR linarith symm calc Real.sqrt (R : ℝ) * (Real.sqrt (R : ℝ) - 1)⁻¹ = (Real.sqrt (R : ℝ) - 1 + 1) * (Real.sqrt (R : ℝ) - 1)⁻¹ := by rw [sub_add_cancel] _ = (Real.sqrt (R : ℝ) - 1) * (Real.sqrt (R : ℝ) - 1)⁻¹ + 1 * (Real.sqrt (R : ℝ) - 1)⁻¹ := by rw [add_mul] _ = 1 + (Real.sqrt (R : ℝ) - 1)⁻¹ := by rw [mul_inv_cancel₀ hne, one_mul] exact h4.congr' heq