SPOTIS (Stable Preference Ordering Towards Ideal Solution)

SPOTIS is a ranking method for when you hold a decision table filled with numbers and want the alternatives placed in a single order. Its output is a distance value for every alternative, showing how far it sits from a fixed ideal point, together with the rank that value produces; a smaller distance is better. Dezert, Tchamova, Han and Tacnet proposed it in 2020, designed explicitly, as its name suggests, to be "rank reversal free." It shares TOPSIS's idea of "closeness to the ideal," but departs from TOPSIS at the root in where that ideal point comes from. It does not generate weights; weights are supplied from outside.

Classical "closeness to the ideal" methods such as TOPSIS and VIKOR share a common weak point. Their ideal and anti-ideal points are built from the best and worst values found within the current alternative set itself. This resembles grading a class on a curve set by its own strongest and weakest student: add a new student to the class and the curve shifts, changing everyone's grade. SPOTIS rejects this idea. Its ideal and anti-ideal are built, before the analysis begins, from the decision-maker's answer to the question "what is the realistically best and worst value conceivable on this criterion"; they work like a fixed ruler set from outside. Every alternative is measured against this fixed ruler, not against the others.

The consequence of this philosophy is written directly into the method's name: a stable ranking. Because the ruler does not change when an alternative is added or removed, the remaining alternatives' distance to it does not change either; the rank reversal seen in TOPSIS and VIKOR is thereby structurally prevented. This comes at a philosophical price of its own: drawing the ruler, that is, setting the bounds, is now the decision-maker's responsibility rather than the data's, and that responsibility must be discharged defensibly.

The method proceeds through three steps.

First, fixing the bounds and the ideal point. For every criterion, the method takes the pair of "best possible" and "worst possible" bounds the decision-maker has set before the analysis. These bounds come not from the best/worst values observed in the data set but from the decision-maker's knowledge of the criterion's realistic range. Each criterion's ideal point is built from these bounds: for a "higher is better" criterion, the ideal is the upper bound; for a "lower is better" criterion, the ideal is the lower bound. This is precisely the choice that makes SPOTIS resistant to rank reversal: the ideal point comes from pre-fixed bounds, not from the alternatives themselves.

Second, the scaled distance. The method takes the difference between each alternative's value on each criterion and that criterion's ideal point, and scales it to between 0 and 1 by dividing by the width of the criterion's fixed bound range. Criteria in different units (TL, days, points) thereby become comparable on the same scale; the scaling is built from the fixed bounds, not from the data set.

Third, the weighted total distance and the ranking. The method multiplies each alternative's scaled distances by the criterion weights and sums them; this total summarises the alternative's whole distance to the fixed ideal point in a single number. The method then ranks the alternatives from smallest to largest on this number. Unlike TOPSIS, here a smaller distance is better.

The formulas behind each step, the intermediate tables and the citation formats are given on the DecisionMind method page; this card carries no formulas.

The distance value shows how far an alternative sits from the pre-fixed ideal point: the closer to 0, the better; the closer to 1, the worse. This value may look like TOPSIS's closeness score, "a ratio between 0 and 1," but its meaning differs. In TOPSIS the ideal is built from that analysis's own alternatives, so it cannot be compared across analyses. In SPOTIS the ideal is fixed. As long as the same fixed bounds are used, this distance value carries a more consistent reference across different analyses or comparisons made at different times. Even so, this value is not a percentage or a probability; it only shows a position against the fixed ruler.

The distance value can sometimes exceed 1, or the scaled distance on one criterion can come out greater than 1. This is not a computational error; it signals that the bounds were set too narrowly. If an alternative's actual value falls even beyond the bound the decision-maker set as "worst possible," the distance exceeds its theoretical upper limit, and in that case the bounds should be reviewed.

For this reason:

"According to SPOTIS, A2 scores 0.425, that is, 57.5 per cent successful"

should be written instead as:

"With these fixed bounds, A2 is the alternative closest to the ideal point (distance 0.425); as long as these bounds remain unchanged, this distance can also be compared against another analysis"

SPOTIS works with crisp data: one number per cell. You need alternatives in rows, criteria in columns, one number per cell; direction information for every criterion; weights summing to 1; and, for every criterion, a lower and upper bound pair set by the decision-maker before the analysis. These bounds are mandatory for SPOTIS to work and must not be derived automatically from the data itself. It does not produce weights, it asks for them; they may be drawn from sources such as AHP, BWM, Entropy or CRITIC. DecisionMind holds five SPOTIS family members alongside the base method (including balanced, fuzzy, negative and positive extensions). A minimum of two alternatives and two criteria is required; there is no upper limit, and three to twelve criteria work comfortably.

If the alternative set is expected to change over time, for instance if a new supplier or a new candidate will be added later, and the existing alternatives' order must not be affected by that addition, SPOTIS is a suitable choice. It also suits repeated comparisons on the same criteria across different periods or different groups. If you can set defensible, realistic best/worst bounds for the criteria in advance, drawing on an accreditation threshold, an engineering tolerance, or a range known from past years, the method sits on solid ground.

SPOTIS should not be used when the criteria's realistic bounds are not known in advance or are disputed. If the bounds are set arbitrarily or too narrowly, the distance values can become meaningless, for instance some alternatives may exceed the theoretical upper limit, and the "fixed ruler" advantage turns into a disadvantage. If a criterion allows no compromise at all, that is, if veto logic is required, SPOTIS does not provide it; it is a compensatory method.

The alternative set will change over time, keeping the ranking stable matters → SPOTIS

The criteria's realistic bounds are known and defensible in advance → SPOTIS

Building the ideal point automatically from the data set suffices, the burden of setting bounds is unwanted → TOPSIS

The bounds are unknown or disputed → clarify the bounds first, or use a data-driven method (TOPSIS, VIKOR)

No compromise is accepted on one criterion (veto logic) → the ELECTRE family

SPOTIS's most important advantage is that it directly delivers what its name promises. Because the ideal point is built from fixed bounds, the other alternatives' distance values do not change when a new member is added to, or removed from, the alternative set; the rank reversal problem known from TOPSIS and VIKOR is structurally prevented by this design (Dezert et al., 2020). Second, its computational burden is small and it is as explainable as TOPSIS: the idea of "distance to a fixed ruler" is easy to convey to a decision-maker. Third, as long as the same fixed bounds are kept, it offers a more consistent basis for comparison across analyses run at different times or with different subgroups. Fourth, this property of the method has already been carried over into fuzzy settings since its publication (Shekhovtsov et al., 2022), which shows the design is extensible.

Its limitations stem from the same design. First, setting the fixed bounds is now the decision-maker's responsibility rather than the data's; if these bounds are chosen arbitrarily, the method's "stability" advantage turns into a "subjectivity" disadvantage. Second, if the bounds are set narrowly, that is, if an alternative's actual value falls even beyond the "worst possible" bound, the scaled distance can exceed its theoretical upper limit and the result becomes uninterpretable. Third, SPOTIS, like TOPSIS, rests on the assumption of full compensation: a weakness on one criterion can be papered over by strength on another; there is no veto logic. Fourth, the method treats criteria as independent, and the quality of the weights lies outside the method itself; this carries over the general MCDM limitations found across the TOPSIS family. More broadly, rank reversal in multi-criteria ranking methods is documented across a wide literature (Li and Abbas, 2026). SPOTIS offers a direct solution to only one of these problems, the ideal point's dependence on the alternative set; it does not remove sensitivity arising from other sources, such as uncertainty in the choice of weights.

The most common mistake is using the best/worst values observed in the data set as the bounds instead of fixed, pre-set bounds; this effectively turns SPOTIS into TOPSIS and entirely removes the method's resistance to rank reversal. A second mistake is setting the bounds narrower than the realistic range and failing to notice that an alternative falls outside them; in that case the scaled distance exceeds 1 and the report must explain this. A third mistake is reading the distance value in the same direction as TOPSIS's closeness score. In SPOTIS a smaller value is better and a larger value is worse; reading it the other way round reverses the ranking. A fourth mistake is setting the bounds without offering any justification and failing to defend that choice in the report; the bounds are SPOTIS's most critical input and must be justified. A fifth mistake is resetting the bounds afresh, and differently, each time the same analysis is repeated; this also destroys the method's advantage of stability over time.

The governing principle is this:

SPOTIS's stability rests on the ideal point coming from pre-fixed bounds rather than from the alternatives; if the bounds are chosen arbitrarily or too narrowly, that stability stops being an advantage and the result itself becomes contestable.