Decision analysis checked against the literature

Weight criteria with CRITIC, rank with TOPSIS and see every step. It runs in your browser: no sign-up, nothing uploaded.

TOPSIS ranking of five smartphones

CRITIC weights, computed live

RankPhoneCloseness C
1E0.6033
2C0.5587
3A0.5548
4D0.4997
5B0.4883

0.1872

Drag it to re-rank

CRITIC weights
Price0.1872Screen size0.1838Pixel density0.1691Thickness0.2599Mass0.2000
Show as table
CriterionWeight
Price0.1872
Screen size0.1838
Pixel density0.1691
Thickness0.2599
Mass0.2000

Data: Krishnan et al. (2021). doi:10.3390/sym13060973

CRITIC weights match the paper's Table 5. TOPSIS is computed here.

How it works

Sensitivity and cross-method checks built in

Sweep any criterion weight from 0 to 1, run 10,000 Monte Carlo draws, or compare all 16 ranking methods side by side to see where the first place switches.

Checked against published results

The test suite recomputes the published examples and compares every printed digit.

  • CRITIC

    Krishnan et al. (2021) · doi:10.3390/sym13060973

    What matched

    Weights and standard deviations match Krishnan et al. (2021) to 4 decimals.

    Tolerance ±0.00005
  • Entropy

    Zavadskas et al. (2017) · doi:10.3390/su9050702

    What matched

    Weights match Zavadskas et al. (2017) to 4 decimals, once the paper's mislabeled rows are put back.

    Tolerance ±0.0001
  • SD

    Diakoulaki et al. (1995) · doi:10.1016/0305-0548(94)00059-H

    What matched

    Weights match Diakoulaki et al. (1995) to 3 decimals, and standard deviations once the paper's population SD is converted.

    Tolerance ±0.0005
  • MEREC

    Keshavarz-Ghorabaee et al. (2021) · doi:10.3390/sym13040525

    What matched

    Weights match Keshavarz-Ghorabaee et al. (2021) to 4 decimals, and the normalized matrix, S, S' and the removal effects match every printed value.

    Tolerance ±0.00005
  • LOPCOW

    Keleş (2023) · doi:10.25287/ohuiibf.1239201

    What matched

    Weights match Keleş (2023) to 3 decimals across all three criteria sets with the sample standard deviation.

    Tolerance ±0.0005
  • CILOS

    Zavadskas et al. (2017) · doi:10.3390/su9050702

    What matched

    Weights match Zavadskas et al. (2017) to 4 decimals across both the Social (Table 10) and Economic (Table 8) criteria blocks once the mislabeled row order is restored.

    Tolerance ±0.0001
  • BWM

    Rezaei (2016) · doi:10.1016/j.omega.2015.12.001

    What matched

    Linear BWM weights match Rezaei (2016) to 4 decimals.

    Tolerance ±0.00006
  • AHP

    Saaty (2008) · doi:10.1504/IJSSCI.2008.017590

    What matched

    Eigenvector weights and consistency ratios match Saaty (2008) to 3 decimals.

    Tolerance ±0.0005
  • ROC

    Roszkowska (2013) · doi:10.15290/ose.2013.05.65.02

    What matched

    ROC, RS and RR weights match Roszkowska (2013) Tables 4 and 6 to 2 decimals, with the same SAW rankings.

    Tolerance ±0.02
  • SWARA

    Keršulienė et al. (2010) · doi:10.3846/jbem.2010.12

    What matched

    s, k, q and all six weights of Table 1 match Keršulienė et al. (2010) to 2 decimals; the 2017 textile supplier SWARA weights match too.

    Tolerance ±0.005
  • TOPSIS

    Opricovic and Tzeng (2004) · doi:10.1016/S0377-2217(03)00020-1

    What matched

    D+, D−, C and ranks match Opricovic and Tzeng (2004) to 3 decimals.

    Tolerance ±0.0005
  • SAW

    Podvezko (2011) · doi:10.5755/j01.ee.22.2.310

    What matched

    Scores and ranks match Podvezko (2011) to 3 decimals with sum normalization, and with linear max once a misprint in its Table 2 is corrected.

    Tolerance ±0.0005
  • VIKOR

    Opricovic and Tzeng (2004) · doi:10.1016/S0377-2217(03)00020-1

    What matched

    S, R, Q and the compromise A2 match Opricovic and Tzeng (2004) to 3 decimals, in both unit encodings.

    Tolerance ±0.0005
  • EDAS

    Keshavarz Ghorabaee et al. (2015) · doi:10.15388/Informatica.2015.57

    What matched

    All 47 appraisal scores and every cell of Table 2 match Keshavarz Ghorabaee et al. (2015) to 2 decimals, rounded as the paper did.

    Tolerance ±0.006
  • WASPAS

    Aytaç Adalı and Tuş Işık (2017) · doi:10.18825/iremjournal.335408

    What matched

    All 55 scores of Table 7, λ from 0 to 1, match Aytaç Adalı and Tuş Işık (2017) to 4 decimals, with the three rank orders the paper reports.

    Tolerance ±0.0001
  • COPRAS

    Özcan and Çelik (2021) · doi:10.4995/ijpme.2021.14734

    What matched

    All seven Q values of Table 11 and every cell of Table 10 match Özcan and Çelik (2021) to 4 decimals; the printed rank row swaps A2 and A3, the printed Q does not.

    Tolerance ±0.0001
  • MARCOS

    Stević and Brković (2020) · doi:10.3390/logistics4010004

    What matched

    All 23 utilities f(K) of Table 5 match Stević and Brković (2020) to 3 decimals, with the same 23 ranks; the paper applies its C3 and C4 weights swapped against its own FUCOM figure, and so does this check.

    Tolerance ±0.001
  • CoCoSo

    Yazdani et al. (2019) · doi:10.1108/MD-05-2017-0458

    What matched

    S, P, ka, kb, kc, k and all seven ranks of Table VI match Yazdani et al. (2019) to 3 decimals; k only within 0.0015, because the paper rounded it at another stage.

    Tolerance ±0.002
  • MABAC

    Pamučar and Ćirović (2015) · doi:10.1016/j.eswa.2014.11.057

    What matched

    The border G, the scores S and all seven ranks of Tables 10 and 12 match Pamučar and Ćirović (2015) to 4 decimals.

    Tolerance ±0.0001
  • MOORA

    Brauers and Zavadskas (2010) · doi:10.3846/tede.2010.01

    What matched

    All nine column norms, every normalized cell and the three ratio-system scores match Brauers and Zavadskas (2010) to 3 decimals, with the same three ranks.

    Tolerance ±0.0005
  • ARAS

    Bakır and Atalık (2018) · doi:10.20491/isarder.2018.410

    What matched

    S₀, S, K and all 11 ranks of Table 11 match Bakır and Atalık (2018) to 4 decimals, with Tables 9 and 10 once THY's reprinted row is corrected.

    Tolerance ±0.00002
  • CODAS

    Badi et al. (2018) · doi:10.31181/dmame180101b

    What matched

    E, T, H and all six ranks of Tables 3 and 4 match Badi et al. (2018) to 3 decimals; the two examples of the 2016 origin paper match too.

    Tolerance ±0.0005
  • PROMETHEE II

    Goswami (2020) · doi:10.2478/fman-2020-0008

    What matched

    π, φ⁺, φ⁻, φ and all six ranks of Tables 13, 15, 16 and 18 match Goswami (2020) to 4 decimals; the 1986 hydroelectric example matches too.

    Tolerance ±0.00002
  • ELECTRE I

    Wang and Triantaphyllou (2008) · doi:10.1016/j.omega.2005.12.003

    What matched

    Concordance C, discordance D, outranking graph S and kernel {A2, A3, A5} match Wang and Triantaphyllou (2008) Section 3.1 to 3 decimals; Soner and Önüt (2006) net concordance matches too.

    Tolerance ±0.0005
  • ELECTRE III

    Wang and Triantaphyllou (2008) · doi:10.1016/j.omega.2005.12.003

    What matched

    All 110 off-diagonal credibility entries of Table 3 and both distillation chains match Wang and Triantaphyllou (2008) within 0.005.

    Tolerance ±0.005
How verification works

Where the tolerance comes from: papers print results rounded to a fixed number of decimals (for example, a 4-decimal value like 0.6033 carries a ±0.00005 rounding margin), so each tolerance above equals the rounding precision of that publication.

Additional checks: alongside published paper tables, the automated test suite verifies permutation, scale and shift invariants, exact algebraic equivalences across methods (MARCOS with linear SAW, V-shape PROMETHEE II with MABAC, distributive AHP with sum SAW), and constant-column or extreme-value boundary cases.