Laboratori bilong Rulet · Laboratori · Statistik · 8 min

Chi-Square Tes

Tes ol i save yusim bilong askim: “ol 37 namba i olsem stretpela wil bai givim?”

Sotpela bekim

χ² = Σ (wanem yu lukim − wanem yu wetim)² / wanem yu wetim, long olgeta 37 hul. Skelim wantaim chi-square distribiusen long 36 degri bilong fridom bilong kisim p-velyu.

namba bilong tes
Σ (O−E)²/E
df
36
mak long 5%
51.0
mak long 1%
58.6

Kaunim em

Avarej namba long wan wan hultulHulWanpela long ol 37 hul i gat namba long rota. Narapela nem: slot, cup.Ritim olgeta → em n/37. Long wan wan hul, kisim senis namel long wanem yu lukim na wanem yu wetim, skwerim, na dilaim long wanem yu wetim; bungim ol 37 namba. Stretpela wil i givim namba klostu 36 long avarej (ol degri bilong fridomlaboratoriDegri bilong fridomNamba bilong ol velyu i no hangamap long narapela insait long wanpela statistik. Long chi-square tes long 37 hul em i 36, long wanem ol kaunim i mas bung i kamap olgeta namba.Ritim olgeta →), wantaim standet diviesenmatematikStandet diviesenSkwea rut bilong varians; hamas longwe resal i save stap long ekspektet velyu.Ritim olgeta → klostu 8.5. Antap long 51.0 i makim samting long mak 5%; antap long 58.6 long 1%. Wheel Bias Analyzer i kaunim namba bilong tes na stretpela p-velyulaboratorip-velyuSans long lukim senis i bikpela olsem o i winim dispela yu lukim, sapos wil i stret.Ritim olgeta → long regularized incomplete gamma function.

n tanimavarej long wan wan hulσ long wan wan hulhul long +2σhul long +3σ
370103.11619
1,000275.13742
3,7001009.9120130
10,00027016.2303319
37,0001,00031.21,0621,094
σ ≈ √(n·p·(1−p)) wantaim p = 1/37. Long 1,000 tanim, hul i kamap 37 taim em i samting nating; hul i kamap 42 taim em i gutpela long kisim namba tu samting yu kaunim.

Wanpela piksa

Consider a hypothetical sample of 1,000 independent spinsmatematikOl tanim i no hangamap long narapelaOlgeta tanim i gat wankain sans, maski wanem samting i kamap bipo. Olgeta kaunim long dispela sait i sanap long dispela.Ritim olgeta →. Each pocket has an expected count of 1,000/37 ≈ 27.03. If pocket 8 appears 41 times, its contribution to the statistic is about 7.22. Suppose the total statistic is approximately 45.2. On 36 degrees of freedom, the approximate p-value is 0.14: under the fair-wheel model, a statistic at least this large has a probabilitymatematikSansSans bilong wanpela resal, long 0 i go 1 o 0% i go 100%. Wanpela namba: 1/37 = 2.70%.Ritim olgeta → of about 14%. This is not significant at the 5% level, and it does not prove the wheel is fair. A pocket selected after examining these counts needs independent confirmation; another 1,000 spins is not guaranteed to resolve a small departure.

Ol samting tes i ting i tru

The usual approximation assumes independent observations and adequate expected counts. The common five-per-cell rule gives n ≥ 185 for 37 equal pockets; it is a guideline for the approximation, not a hard boundary for all possible tests. Pooling wheels changes the population being tested. Dependence may invalidate the reference distribution without necessarily increasing the statistic.

Pasin bilong Han bilong Dila long RuletOl Tanim bilong Rulet i No Hangamap long Narapela

Pearson na Monte Carlo

Karl Pearson i putim tes ples klia long 1900. Wanpela bilong ol namba wan piksa bilong en em 26,306 tanim bilong rulet long Monte Carlo ol i bin raitim. Em i painim olsem ol i no wankain tumas bilong kamap long laki, na em i tokaut olsem “Monte Carlo roulette… is not a game of chance” (rulet bilong Monte Carlo i no pilai bilong sans). Bihain ol narapela raita i tok olsem ol nius man i raitim ol tanim bilong wanpela niuspepa i bin giamanim planti bilong ol. Em i wanpela skul tu long save gut long we ol namba i kam long en.

Tes bilong Pearson long 1900 i painim olsem ol tanim bilong Monte Carlo i no wankain tumas; ol man bilong niuspepa i raitim i bin giamanim planti bilong ol.

Rulet long Monte Carlo

Wanpela tes tru: ol wil i wok long em yet long Las Vegas

A published analysis of specific automatic wheels is evidence about those recorded samples, not all automatic roulette. A non-significant uniformity test cannot prove the absence of steering, and combining several wheels does not identify the behavior of each one. Check the original study’s data and design before drawing that conclusion.

Moa long chi-square

Long wanpela hul tasol yu ting i no stret, binomial tes i strong moa. Long wanpela hap bilong wil yu ting i no stret, yusim z-tes bilong seksenwilSeksenWanpela hap bilong wil we ol hul i stap wan bihain long narapela, i narakain long wanpela grup long laiaut.Ritim olgeta → long pes bilong skelim ol seksen. Chi-square em i tes bilong olgeta samting: em i tokim yu olsem wanpela samting i no stret, tasol i no tok wanem samting. Yusim em pastaim, na bihain lukluk.

Skelim Ol Seksen

Fomula

χ² = Σᵢ (Oᵢ − Eᵢ)² / Eᵢ
Eᵢ = n / 37 · df = 36
ol mak: 51.0 (5%) · 58.6 (1%)

Ol askim

Bilong wanem 36 degri bilong fridom na i no 37?

Ol 37 namba i mas kamap n wantaim, olsem na 36 tasol i fri long senis.

Wil i gat strongpela bias bai soim wanem χ²?

Wil i gat wanpela hul long 1/25 long 10,000 tanim bai givim namba klostu 60 long dispela hul tasol; χ² klostu 100, na p klostu siro tru.

Mi ken mekim long 200 tanim?

Long lo bilong tes, yes, wantaim avarej 5.4 long wan wan hul. Tasol em bai i no painim wanpela samting, sapos ol i no pasim wanpela hul olgeta long glu.

Liklik lap

For 36 degrees of freedom, the approximate 5% upper-tail cutoff is 51.0.

For 36 degrees of freedom, the approximate 5% upper-tail cutoff is 51.0.

Ol As bilong Save

  1. Pearson, K. “On the criterion that a given system of deviations from the probable…” https://www.itl.nist.gov/div898/software/dataplot/refman1/auxillar/chsqgood.htm
  2. Feller, W. An Introduction to Probability Theory and Its Applications, Vol. 1. https://www.amstat.org/asa/files/pdfs/P-ValueStatement.pdf
model: Yurop, wanpela sirometodoloji ↗ol as bilong save ↗