Dear Mahmut

I have read the paper that you and your colleagues wrote on:

“Exploring the Specific Capacity of Different Multi Criteria Decision Making Approaches under Uncertainty using Data from Financial Market” – Mahmut Baydas et al

1- I highly agree with your paragraph, especially in the last sentence:

“This study is aimed to determine the capacity of MCDM methods by outputs rather than inputs, unlike the previous literature. Discussions in the recent literature points out that the capacity of a MCDM method that better fits real life problems can be higher”.

2- You say: “In addition, since different MCDMs can produce inconsistent results in a given scenario”

Which are those inconsistencies? Certainly not the different rankings, because they obey to the nature of the method, because all methods use the same data.

3- I believe that there is a mistake on this sentence of yours: “It should be noted here that an evaluation is made not on the equations of the MCDM methods, which are the inputs, but on the scores”

We are interested in rankings, not in scores. That is, what is important is the position that each alternative has, not its value. Normally, scores are different among methods, because the different procedures and assumptions, and if they use weights or not. If you start with a matrix, and use procedures without biases, even employing the same subjective weights in all of them, with no assumptions, the rankings should be the same in all of them, because you are using mathematics.

This is not a theory, there is proof of it, since you can see this in SIMUS. It departs from the same matrix, uses two different procedures, like SAW and Outranking, it produces two solutions, with different scores, but delivers identical rankings, from both procedures.

4 – I disagree with this sentence: “Therefore, if time permits, it is advisable to approach a decision problem using different decision-making methods”. There is not mathematical support for this theory. Which is the gain in determining that two MCDM methods have very similar rankings? What does it prove? Nothing, and it can even be coincidence.

5- You compare PROMETHEE II and say that it is the best method, and I agree, because it is rational, uses resources and establishes pair-wise comparisons based on real data. Unfortunately, it also uses AHP subjective weights for criteria, which does not have mathematical support, except by the use of the Eigen method. Using statistics or entropy will make it a much better method, also supported by GAIA.

6- I don’t think that normalization deteriorates information, since it divides all the values in the initial matrix by the same value.

7- I agree that sensitivity analysis can be a good indicator when determining the strength of the best solution, but, remember that different MCDM methods may yield different best alternatives, so, how do we compare them based on strength of the best solution? If two or three methods coincide in the best alternative, the procedure can work, but no if they are different.

8 - You say: “However, this study focuses directly on the ranking scores that the methods produce. In another saying, when outputs of the MCDMs are analyzed, the actual capacities of the methods can be revealed through a real-life reference tool”

Agreed, you an I discussed this subject about two years ago.

9- There is something that is stated in the paper that nags me, probably due to my ignorance in finances.

You state that: “The ultimate goal of the firm is to maximize its market value, and an actual relationship is expected between firm-based share return performances determined and approved by market participants and the financial performances of the firms”

As far as I know, all financial statements are mathematically related. Consequently, the performance, measure say by a financial indicator like ROI, has been computed using the same data utilized to compute firm-share. Considering even that the relationship between both is not linear, it is evident that one conditions the other, and thus, there could be a high correlation, even a cause-and-effect relationship.

As I see it, if equity increases because sell of shares, this is reflected in performance, which is also used to compute te financial indicators. Please correct me if I am wrong.

10 -in your conclusion you say: “Therefore, the significant correlation between the financial performance rankings of the companies obtained by MCDM and their share return rankings can give us a valid and reliable idea about the superiority of the methods”

I agree with this, provided that the MCDM method uses objective weights. In my opinion, if somebody do that using subjective weights is wasting his/her time.

Your comments will be appreciated

Nolberto Munier