Talk:Decision theory

Untitled
There was a miserable little stub here, which got itself listed on VfD. I have now written the beginnings of a real article, and I hope no-one will still consider it deletable. But what I have put in is very much a psychologist's perspective - we need some help here from economists, mathematicians, statisticians, and management scientists. This ought to grow into a keystone article, it crosses so many boundaries. seglea 05:40, 29 Jan 2004 (UTC)

Well, this statistician has now added his two penn'orth.Blaise 08:03, 7 October 2005 (UTC)

Difference between decision theory and game theory
Sanjiv swarup (talk) 16:07, 17 June 2008 (UTC)

Can someone explain the difference between decision theory and game theory? Thanks. Samw 21:30, 27 Nov 2004 (UTC)

This link talks about the difference. http://levine.sscnet.ucla.edu/general/whatis.htm "Decision theory can be viewed as a theory of one person games, or a game of a single player against nature."


 * Also comments on this very page at pragmatic v formal.Cretog8 (talk) 16:35, 17 June 2008 (UTC)

Decision theory vs. Detection theory
In signal processing/statistics/communications, this is known as detection theory. I don't want to say that one name is correct over the other, but I also don't want two articles heavily overlaping. It seems decision theory is geared more toward "making decisions" (1 ton of guns & 3 tons of butter vs. 2 tons of guns & 1 ton of butter) while detection theory is more about choosing/detecting (is the signal present or not?).

I think there's room for both since the technical and less-technical call them different things... Cburnett 05:24, Apr 21, 2005 (UTC)
 * Further point, Category:Decision theory has only one parent category: Category:Management. Cburnett 05:25, Apr 21, 2005 (UTC)

I can't see that detection/decision theory are the same at all. They cover some of the same ground, but in quite different ways; eg decision theory seems much more general (even abstract), though MCDA etc is very applied. Possibly it could be argued that detection theory is a specific application of decision theory, but I don't know enough about it to say whether that's a helpful way of putting it. Rd232 09:31, 21 Apr 2005 (UTC)

Proposed merger
I'm not sure if these two things are the same (I suppose Rational choice theory might be ideal decision theory and thus a distinct theory), but if they're not the same, the difference between them probably needs to be discussed on each of the pages. KSchutte 19:22, 12 March 2006 (UTC)


 * Rational choice theory is the same thing as decision theory, but the are usually used in slightly different contexts. The term decision theory is more often used when describing it as a prescriptive framework for how to behave rationally, whereas the term rational choice theory is more often used to discuss it as a descriptive framework of how people really do act. In other words, rational choice theories model people as behaving according to decision theory. Closed Limelike Curves (talk) 17:18, 17 July 2024 (UTC)

Rational choice theory would be a subset of decicion theory. They should have their seperate articles.radek 01:10, 14 March 2006 (UTC)

They are nothing alike. Rational choice is a framework for examining human behavior. In many ways, decision theory is a mathematical/statistical approach to issues that may in fact extend beyond human behavior--see the discussion on decision theory vs. detection theory above. h27kim 11:08, 20 March, 2006 (PST)

This is bit imprecise, and also an overstatement. In my Decision Theory course (broader than "a mathematical/statistical approach"), I treat Rational Choice Theory as a central part of the theoretical core. Thus, I agree with Radek, above, who defines RCT as a subset of DT. Nonetheless, I do see the two fields as distinct enough to justify two entries. RCT is a body of theory largely aimed at circumscribing ad hoc (and irrational) explanations of behavior in a wide variety of fields and replacing it with hypotheses that seek to explain that behavior with explanations that give more weight to the rationality, albeit messy, thinking of the agents. In this way RCT is commonly associated with the Chicago school of economics, specifically with Gary Becker, who sees it as a duty of economists to find the rationality below the surface. Decision theory in contrast to RCT seeks instead to develop insights from multiple sources (including but not limited to RCT) while attempting to develop global (normative) prescriptions for making hard decisions. MFortunato

I agree with Radek, and others. A further distinction is that Decision Theory deals primarily with decisions under uncertainty, while Rational Choice theory does not. The intersection between the two, broadly speaking, is expected utility theory. JQ 20:42, 23 April 2006 (UTC)

Oppose -- They are at least as different as Management science and Operations research MPS 20:06, 27 June 2006 (UTC)

Oppose -- Rational choice theory is a subset of decision theory. The are many theories of decision-making (such as prospect theory) which do not posit rational choice. Rational choice and decision theory are clearly not the same.--Ossanha 21:12, 7 July 2006 (UTC)

Oppose -- Ditto, while I don't see that decision theory is that well defined, it has to encompass a range of theories about how the choice is made, RC being just one of them.

Oppose -- per above. I guess that's enough to take the merger tags off. ~ trialsanderrors 18:37, 27 August 2006 (UTC)

Oppose the urge to merge --I agree with oppositions to merger expressed above. I hope no-one will still consider it deletable after strong opposition to merge above. They are different topics. Both these article should remain independent.

Prof.Sadiq 17:45, 15 January 2007 (UTC)

Oppose the urge to merge --I agree with opposition to merger.

Oppose the urge to merge -- These two subjects are related but not the same. Decision Theory is the broad term for studying how people do and should make decisions. It includes the judgment and decision making research coming from pyschology, the normative side, and the Bayesian decision theory research coming from Bayesian statistics, the prescriptive side. Decision Analysis is the convergence of these areas of research. The following description of decision analysis is taken from the Lexicon of Decision Analysis from the Decision Analysis Society "A decision maker might employ decision analysis, which is a structured way of thinking about how the action taken in the current decision would lead to a result. In doing this, one distinguishes three features of the situation: the decision to be made, the chance and unknown events which can affect the result, and the result itself. Decision analysis then constructs models, logical and perhaps even mathematical representations of the relationships within and between these three features of the decision situation. The models then allow the decision maker to estimate the possible implications of each course of action that he might take, so that he can better understand the relationship between his actions and his objectives."

User:JasonRWMerrick 3:00 pm, 16 March 2007

pragmatic v formal
I am at a loss as to the difference between the formal approach and the pragmatic approach to decision-making. Any ideas?

This is largely a scholarly distinction -- there is in most fields considerable room to study the theory of something without it leading to greater competence in the practical use of it. But in decision theory the question is especially well motivated, as it is unclear why one would wish to understand the theory of making decisions if it does not lead to making better decisions in practice. For most practitioners, the actual model of decision making they would choose to use in most situations would not be the perfect model, but instead a "good enough" model -- admitting at the outset that the theoretical crieria are not being fully or perfectly met for that approach to be theoretically perfect. (One often doesn't even know one's own preferences well enough to say whether the functional form of the utility function, for example, is just right.) Knowledge of the formal approach does however inform the pragmatic decision maker -- knowing that you are using an imperfect but approximate method, and knowing where you are making pragmatic compromises, is usually worth knowing. MFortunato

In the literature there is no consenus how exactly the terms 'decision theory', 'game theory' and alike are to be used. It seems important to me, though, that at least the following distinctions are kept in mind: -> 1 Person vs >1 person -> normative vs descriptive Sometimes we talk about what rational people do and hence are in the normative domain; at other times we explain or predict actual human behavious and hence are in the descriptive domain. -> Aximotic vs Interpretative Correctly applying axiomatic rules to axioms and other theorems is one thing, interpreting (parts of or the whole of) the calculus quite another. Whether or not rational choice theory really is about the right concept of rationality (if there is such a thing at all), for example, is a matter of interpretation rather than pure calculus.

'Game theory' tends to be used for >1 person-games that are interpreted in order to describe reality. 'Decision theory' tends to be used for 1- and >1-person-games, is mainly normative, and deals a lot with the issues of the right interpretation. GIV

Commensurable
I removed "[Editor's note: this is not an accepted use of the term 'commensurable'.]" because that sort of commentary belongs here on the discussion page. Either the term *is* the accepted term used in decision theory in which case the use is fine, or else there is some other term that is used instead, in which case you can just replace commensurable with the proper term. Or if the word just has a different meaning in decision theory than in economics, that can be pointed out without making it seem like an error. Jackdavinci 06:53, 22 July 2006 (UTC)

- Merging it with raational choice would be very confusing.

Merge with decision analysis
I propose these pages are merged as they seem to be covering the same material. Thoughts? Andeggs 15:29, 23 December 2006 (UTC)
 * I would oppose this, as decision analysis seems to be a prescriptive approach advocating rational choice based models in making decisions, and the concept is geared towards business people. Decision theory is (to me) a field of applied math (economics, AI, etc) that attempts to model different ways of making decisions, which includes rational choice models, but also includes more subjective models. Smmurphy(Talk) 04:04, 25 March 2007 (UTC)
 * I also oppose for similar reasons. -Gomm 19:46, 26 March 2007 (UTC)

Where is mention of Behavioral Decision Theory? It has been blossoming over the last couple of decades, manifesting itself even in US federal policy decisions. Wish I were the guy to write it, but I'm a lowly decision analyst. DaveBees 03:32, 3 May 2007 (UTC)

I think there has been a lot discussion about decision analysis being a prescriptive approach. I differ and so would some people in the field. Even if it is assumed to be prescriptive it does seem to be a part of Decision Theory. I strongly support this merger.


 * My interpretation of decision analysis is based in the article, as I don't know much about it. Looking at the bibliography and seeing Raiffa, maybe they are the same thing after all, I'm not sure.  Anyway, the last two lines of decision analysis about it being prescriptive don't seem to fit with how some might see decision theory. Less those two lines, everything in that article could be merged here (most of it is here already).  If we did this, I think that a separate section on decision analysis or "applied decision theory" for its uses in business and government might help, and might show why the other name redirects here. Smmurphy(Talk) 02:52, 4 June 2007 (UTC)

Calibrated Probability Assessments article
I'm going to write an article on calibrated probability assessments. This is based on the observed phenomenon that training has been shown to improve a person's ability to place odds on uncertain estimates and events. In other words, of a large number of times when a calibrated person says they are 80% confident in a projection, they will right about 80% of the time.

I was considering making a separate article but, after looking at the sub-articles herein, maybe it should be merged with this. Any ideas?

The research comes from Sarah Lichtenstein, J. Edward Russo, and more. Alternate terms for same are "calibration of probabilities" or "calibrated assessment of probabilities".Hubbardaie 03:56, 16 June 2007 (UTC)

Introductory sentence.
This article starts off like an advertising campaign rather than a neutral presentation.


 * Decision theory is an area of study of discrete mathematics, related to and of interest to practitioners in all branches of science, engineering and in all human social activities."

Decision theory is "of interest to" pretty much everybody who ever lived? I humbly disagree. It is of interest to persons who want to study models of decision-making (especially normative models). But it is a bit overboard to claim that all engineers and scientists are interested in (or ought to be interested in) this particular model, much less housewives and others involved in "human social activities".

This article isn't really my pet, so I hope that someone else will edit the sentence and tone down the hyperbole. Phiwum 19:17, 28 September 2007 (UTC)

Computer science project
I've added it to WP:CS for it's relevance to Artificial Intelligence and Decision engineering. Diego (talk) 13:47, 10 July 2009 (UTC)

[Unsigned, chaos]
Please excuse me if I have incorrectly placed my thoughts. Curiously I was looking to find an article related to decision making and its process. Wondering myself that if at any given time as choices and decisions are being made that(hypotheticaly)given the same information and within the same environment would the same decision have been made, not knowing any better at the time? Wouldn't this then somehow make the preamble of "How we make a choice" and understanding how unilateral choices truely are? Assuming of course that I am able to rationalize within a set of predetermined parameters the out come of my choice (chosing to do anything at all) has already been made.

Typing this text right here right now? Should I now be able to back in time without knowing the outcome I would have made the same choice to go ahead and type this text.

Perhaps if that is such the case then maybe we are pre-destined for everything. As if the course then of our lives have been fully planned for our eventuality.

Afterall the only thing that we really know is that we don't know anything. :) —Preceding unsigned comment added by Cjuner (talk • contribs) 20:04, 13 October 2009 (UTC)

First use of "decision theory" citation needed
Article states:

The phrase "decision theory" itself was first used in 1950 by E. L. Lehmann.[citation needed]

Would that be:

Lehmann, E.L.; Scheffé, H. (1950). "Completeness, similar regions, and unbiased estimation. I.". Sankhyā: The Indian Journal of Statistics 10 (4): 305–340. MR39201. JSTOR 25048038 -- (from E. L. Lehmann? I don't have convenient access to article on JSTOR.  If you do, and can confirm, please cite as such.  Alternatively, delete mention of his first use.

OED, Second edition, 1989; online version November 2010, states (not first use, certainly, as Chernoff & Moses published Elementary Decision Theory in 1959, based on nine years of course notes):

decision theory n.

1961 				  			  Jrnl. Acoustical Soc. Amer. XXXIII. 358/1  				An algorithm based on statistical *decision theory.

1964 				  			  T. W. McRae Impact of Computers on Accounting v. 120   				Decision theory is probing the psychology of decision making, and attempts to provide an algorism for taking decisions. -- Paulscrawl (talk) 04:20, 3 December 2010 (UTC)


 * I do have access to the full text of the 1950 Sankhyā article but it doesn't mention 'decision theory'. (It does mention 'decision functions' but it's certainly not the first use of that term as it references a 1947 paper by Abraham Wald with that term in the title.) --Qwfp (talk) 13:51, 3 December 2010 (UTC)

OK, thanks. I've searched Google Scholar fairly thoroughly and can't find Wald's usage of "decision theory" at all -- looks like that while Wald was the pioneer, Lehman gets the coinage prize after all; I've found a few sources:

1. SOURCE: Jeff Miller, Earliest Known Uses of Some of the Words of Mathematics, March 22, 2010.

Statistical DECISION THEORY was essentially a 20th century development although some of the ideas can be found in earlier work: see MEAN ERROR.

The theory exists in both classical and Bayesian versions.

The classical theory was founded by Abraham Wald in 1939 ("Contributions to the Theory of Statistical Estimation and Testing Hypotheses," Annals of Mathematical Statistics, 10, 299-326) and developed in his book Statistical Decision Functions (1950).

The phrase "decision theory" appears in E. L. Lehmann's "Some Principles of the Theory of Testing Hypotheses," Annals of Mathematical Statistics, 21, (1950), 1-26.

2. SOURCE: H. A. David, First (?) Occurrence of Common Terms in Statistics and Probability, Dec 17, 2008.

Decision function, statistical Wald, A. (1945a, title) --- theory Lehmann, E. L. (1950, p. 5)

Wald, A. (1945a). Statistical decision functions which minimize the maximum risk. Ann. Math., 2nd series, 46, 265-280. Lehmann, E. L. (1950). Some principles of the theory of testing hypotheses. AMS, 21, 1- 26.

3. And a peer-reviewed article, albeit in another field of research, S. Hickey Nutrient risk assessment in a decision theoretic context. Journal of Nutritional&Environmental Medicine September 2008; 17(3): 184–194

The term decision theory was coined in 1950 by Lehman [13] for a branch of cybernetics concerned with optimal decision-making, assuming an ideal and rational decision maker. Lehmann EL. Some principles of the theory of testing hypotheses. Ann Math Stat 1950;21:1–26.

So, that's it for now. I'll add Lehman citation to article if no one can find an earlier source. Seems odd -- there was a LOT of decision theoretic work in operations research during WWII and I am surprised it took five years after the war for the phrase to appear in print. I tried Kenneth Arrow and RAND, but no such luck. - Paulscrawl (talk) 18:19, 3 December 2010 (UTC)

✅ -- Paulscrawl (talk) 11:25, 4 December 2010 (UTC)

#Paradox_of_choice is missing
https://en.wikipedia.org/wiki/Decision_theory#Paradox_of_choice is missing. Where can I find this paragraph? --80.243.61.37 (talk) 07:33, 14 September 2015 (UTC)

External links modified
Hello fellow Wikipedians,

I have just added archive links to 1 one external link on Decision theory. Please take a moment to review my edit. If necessary, add after the link to keep me from modifying it. Alternatively, you can add to keep me off the page altogether. I made the following changes:
 * Added archive http://web.archive.org/web/20060705052730/http://www.infra.kth.se/~soh/decisiontheory.pdf to http://www.infra.kth.se/~soh/decisiontheory.pdf

When you have finished reviewing my changes, please set the checked parameter below to true or failed to let others know (documentation at ).

Cheers.—cyberbot II  Talk to my owner :Online 02:15, 19 March 2016 (UTC)

results depend s on another (but not on each other) / did you mean: results are interdependent?
If so correct the phrasing. If not, write within a parenthesis the sense in simple English.


 * I would have written: results are co-configured via an interdependence connectome
 * because the connectome is important: a. not all results are linked, b. not all results are immediately linked without intermediate stages, c. not all relationships have a single input (mono-link), d. some of the results might not connect (dump result: dump results aren't necessarily wrong if the program is [self]evolving. the perfect program tests some alterations with some dumped results, and usually few dumped results if wisely (and arbitrarily) selected are beneficial), e. some (usually few and not all) results are not interdependent (for example direct input as output, in few cases that's the correct answer)