/**
 * Statistical Functions
 *
 * Native RuntimeValue implementations.
 */
import type { RuntimeValue } from "../runtime/values.js";
type NativeFn = (args: RuntimeValue[]) => RuntimeValue;
export declare const fnMEDIAN: NativeFn;
export declare const fnLARGE: NativeFn;
export declare const fnSMALL: NativeFn;
export declare const fnRANK: NativeFn;
export declare const fnSTDEV: NativeFn;
export declare const fnSTDEVP: NativeFn;
export declare const fnVAR: NativeFn;
export declare const fnVARP: NativeFn;
export declare const fnNORMSDIST: NativeFn;
export declare const fnNORMDIST: NativeFn;
export declare const fnNORMSINV: NativeFn;
export declare const fnNORMINV: NativeFn;
export declare const fnPERCENTILE: NativeFn;
export declare const fnPERCENTILEEXC: NativeFn;
export declare const fnQUARTILE: NativeFn;
export declare const fnQUARTILEEXC: NativeFn;
export declare const fnMODE: NativeFn;
/**
 * MODE.MULT — returns a vertical array of every mode (dynamic array).
 * When the dataset is multimodal Excel spills all of them; for a single
 * mode it behaves like MODE.SNGL.
 */
export declare const fnMODE_MULT: NativeFn;
export declare const fnCORREL: NativeFn;
export declare const fnSLOPE: NativeFn;
export declare const fnINTERCEPT: NativeFn;
export declare const fnRSQ: NativeFn;
/**
 * STEYX(known_y, known_x) — standard error of the predicted y-value for
 * each x in a regression. Matches Excel's definition:
 *   SE = sqrt((1/(n-2)) × (S_yy − S_xy² / S_xx))
 * where S_xx, S_yy, S_xy are centred sums of squares / cross-product.
 */
export declare const fnSTEYX: NativeFn;
export declare const fnFORECAST: NativeFn;
export { fnFACT, fnFACTDOUBLE, fnCOMBIN, fnCOMBINA, fnPERMUT } from "./math.js";
export declare const fnGEOMEAN: NativeFn;
export declare const fnHARMEAN: NativeFn;
export declare const fnTRIMMEAN: NativeFn;
export declare const fnDEVSQ: NativeFn;
export declare const fnAVEDEV: NativeFn;
export declare const fnCONFIDENCENORM: NativeFn;
/**
 * CONFIDENCE.T — confidence interval half-width for the mean using the
 * Student's t distribution (small sample / unknown population variance).
 */
export declare const fnCONFIDENCE_T: NativeFn;
/** COVARIANCE.P — population covariance. */
export declare const fnCOVARIANCE_P: NativeFn;
/** COVARIANCE.S — sample covariance (divide by n-1). */
export declare const fnCOVARIANCE_S: NativeFn;
/**
 * RANK.AVG — average-tie rank. Identical to RANK.EQ except that tied
 * positions return the average of the ranks they would otherwise span.
 */
export declare const fnRANK_AVG: NativeFn;
export declare const fnFISHER: NativeFn;
export declare const fnFISHERINV: NativeFn;
export declare const fnAVERAGEA: NativeFn;
export declare const fnMAXA: NativeFn;
export declare const fnMINA: NativeFn;
export declare const fnPOISSON_DIST: NativeFn;
export declare const fnBINOM_DIST: NativeFn;
/**
 * BINOM.DIST.RANGE(trials, probability, number_s, [number_s2]) —
 * probability of a binomial trial outcome between `number_s` and
 * `number_s2` (inclusive). When `number_s2` is omitted, returns the
 * probability of exactly `number_s` successes.
 */
export declare const fnBINOM_DIST_RANGE: NativeFn;
export declare const fnBINOM_INV: NativeFn;
export declare const fnHYPGEOM_DIST: NativeFn;
export declare const fnNEGBINOM_DIST: NativeFn;
export declare const fnCHISQ_DIST: NativeFn;
export declare const fnCHISQ_INV: NativeFn;
export declare const fnCHISQ_DIST_RT: NativeFn;
/**
 * CHISQ.INV.RT(probability, df) — right-tailed inverse of chi-square.
 * Equivalent to CHISQ.INV(1 - probability, df). Probabilities of 0 or 1
 * return +∞ or 0 respectively; values outside (0, 1] are #NUM!.
 */
export declare const fnCHISQ_INV_RT: NativeFn;
export declare const fnF_DIST: NativeFn;
export declare const fnF_INV: NativeFn;
export declare const fnT_DIST: NativeFn;
export declare const fnT_INV: NativeFn;
export declare const fnT_DIST_2T: NativeFn;
export declare const fnT_DIST_RT: NativeFn;
export declare const fnT_INV_2T: NativeFn;
export declare const fnBETA_DIST: NativeFn;
export declare const fnBETA_INV: NativeFn;
export declare const fnGAMMA: NativeFn;
export declare const fnGAMMALN: NativeFn;
export declare const fnGAMMA_DIST: NativeFn;
export declare const fnGAMMA_INV: NativeFn;
export declare const fnEXPON_DIST: NativeFn;
export declare const fnWEIBULL_DIST: NativeFn;
export declare const fnLOGNORM_DIST: NativeFn;
export declare const fnLOGNORM_INV: NativeFn;
export declare const fnPHI: NativeFn;
export declare const fnGAUSS: NativeFn;
export declare const fnERF: NativeFn;
export declare const fnERFC: NativeFn;
export declare const fnSTANDARDIZE: NativeFn;
export declare const fnFREQUENCY: NativeFn;
export declare const fnGROWTH: NativeFn;
export declare const fnTREND: NativeFn;
export declare const fnLINEST: NativeFn;
export declare const fnLOGEST: NativeFn;
/**
 * F.DIST.RT(x, d1, d2) — right-tail probability of the F-distribution.
 *
 * Equivalent to `1 - F.DIST(x, d1, d2, TRUE)`. Using the symmetry of the
 * regularized incomplete beta function, this can be expressed as
 * `I(d2/(d2 + d1*x), d2/2, d1/2)`, which avoids subtracting from 1 and
 * is numerically stable in the upper tail.
 */
export declare const fnF_DIST_RT: NativeFn;
/**
 * F.INV.RT(p, d1, d2) — inverse right-tail of the F-distribution.
 * Returns x such that P(F > x) = p. Implemented via binary search on the
 * right-tail CDF (monotonically decreasing from 1 at x=0 to 0 at x=∞).
 */
export declare const fnF_INV_RT: NativeFn;
/**
 * SKEW — sample skewness.
 * Formula: n / ((n-1)(n-2)) * Σ((xi-mean)/s)^3, where s is the sample stdev.
 */
export declare const fnSKEW: NativeFn;
/**
 * SKEW.P — population skewness.
 * Formula: (1/n) * Σ((xi-mean)/σ)^3, where σ is the population stdev.
 */
export declare const fnSKEW_P: NativeFn;
/**
 * KURT — sample excess kurtosis.
 * Formula: n(n+1) / ((n-1)(n-2)(n-3)) * Σ((xi-mean)/s)^4 - 3(n-1)^2 / ((n-2)(n-3)).
 */
export declare const fnKURT: NativeFn;
export declare const fnPERCENTRANK: NativeFn;
export declare const fnPERCENTRANK_INC: NativeFn;
export declare const fnPERCENTRANK_EXC: NativeFn;
/**
 * PROB(x_range, prob_range, lower_limit, [upper_limit]) — probability that
 * values in x_range are between lower_limit and upper_limit inclusive.
 *
 * Excel rules:
 *   - x_range and prob_range must have the same dimensions
 *   - prob_range entries must sum to 1 (±ε); otherwise #NUM!
 *   - Any prob entry ≤ 0 or > 1 → #NUM!
 *   - upper_limit omitted → probability that x = lower_limit
 *   - Result is the sum of prob_range entries for which
 *     lower_limit ≤ x_range value ≤ upper_limit
 */
export declare const fnPROB: NativeFn;
/**
 * Z.TEST(array, x, [sigma]) — one-tailed probability value for z-test.
 *   mean = AVERAGE(array), n = COUNT(array)
 *   sigma = provided OR sample standard deviation of array
 *   z = (mean - x) / (sigma / √n)
 *   Z.TEST = 1 - NORM.S.DIST(z, TRUE)
 */
export declare const fnZ_TEST: NativeFn;
/**
 * F.TEST(array1, array2) — two-tailed F-test probability comparing
 * the variances of two samples.
 *
 *   F = var(larger) / var(smaller)  (always >= 1)
 *   P = 2 × P(F_dist(df1, df2) > F)
 */
export declare const fnF_TEST: NativeFn;
/**
 * T.TEST(array1, array2, tails, type) — tails in {1, 2}, type in {1, 2, 3}.
 *   type 1: paired
 *   type 2: two-sample, equal variance
 *   type 3: two-sample, unequal variance (Welch's)
 */
export declare const fnT_TEST: NativeFn;
/**
 * CHISQ.TEST(actual_range, expected_range) — chi-square independence test.
 *
 *   χ² = Σ (actual_i - expected_i)² / expected_i
 *   df = (rows-1)(cols-1) for a contingency table, or n-1 for 1-D
 *   p = 1 - CHISQ.DIST(χ², df, TRUE)
 */
export declare const fnCHISQ_TEST: NativeFn;