June 2017 - SIPTA News - lists.sipta.org

[SIPTA] Fwd: [Bayes-forum: 148] Three post-doctoral positions on Bayesian machine learning

by Thomas Augustin

June 27, 2017

Re: [SIPTA] Question about the word "gamble" -- and "prevision"

by Joseph Kadane

Dear All, I wonder if "precision" is specific to the use of squared error loss, or is more general. Jay Kadane Sent from my iPhone > On Jun 12, 2017, at 2:52 AM, "angelo.gilio(a)sbai.uniroma1.it" <angelo.gilio(a)sbai.uniroma1.it> wrote: > > Dear Professor Bickis (and dear Colleagues and friends), > > my message is just to confirm that the term 'prevision' used in the > probability theory of de Finetti has a meaning completely different from > 'prediction'. Operatively, the prevision of a random quantity X is the > amount that you accept to pay (resp., to receive) in order to receive > (resp., to pay) an amount equal to the value which will be assumed by X. > To make a prediction on X is just trying to specify a value that in your > opinion will be assumed by X; this is very different from assigning a > prevision, because it may be that the prevision of X is not a possible > value of X, as it only belongs to the convex hull of the possible values > of X. > > All the best, > Angelo Gilio > > > > >> >> >> Dear Professor Bickis, >> >> In Italian the two words >> >> Previsione=Prevision and Predizione=Prediction have different acceptation. >> >> >> (Devoto-Oli Dictionary) >> >> Previsione = Elaborazione mentale relativa al futuro, sulla base di >> indizi piu' o meno sicuri >> >> (Mental thinking about the future based on more or less secure clues) >> >> >> Example: Previsioni atmosferiche =Weather forecasts >> >> Predizione = previsione autorevole e solenne riferita come oggetto di >> una previsione o profezia >> >> (Authoritative and solemn predictions referred as the subject of a >> prediction or prophecy) >> >> >> Best regards, >> Serena >> >> >> >> Professor Bickis <bickis(a)snoopy.usask.ca> ha scritto: >> >>> Thanks for all your collective erudition. I am aware that “gamble” >>> or its synonyms “wager” and “bet” have been used by many writers in >>> the context of probability and decisions theory. I was thinking of >>> the use of the word in the generic abstract sense that Walley uses. >>> Since Walley’s development picks up on de Finetti and Williams in >>> using “prevision”, I wondered if one of them had also used “gamble”, >>> but it seems that they did not. >>> >>> Googling some Italian web sites on probability, I did come across >>> the word “scommessa” in discussion of expectation/prevision, but I >>> wonder if de Finetti himself used it in the sense that Walley used >>> “gamble”. >>> >>> I am also a bit mystified by the history of the word “prevision”. >>> Writing in French his 1937 paper, de Finetti uses “prévision” rather >>> than the more common (even then) espérance. I would translate >>> “prévision” as “predicition” in English. I see that in the 1974 >>> English translation of de Finetti’s "Teoria de la Probabilità”, the >>> translators note that they are using “prevision” (where earlier >>> translators had used “foresight”), thereby making “prevision” less >>> esoteric in English. However, in the same book, de Finetti makes a >>> big deal that the “prevision is not prediction” (in the English >>> translation). I would be interested if any Italian speakers out >>> there could explain how de Finetti described this in his native >>> language. I would render the French “prévision” as “predicition” in >>> English, so naively would think that “previsione” in Italian also >>> translates as “prediction”. Google Translate, however, gives me >>> “predizione”, >>> for “predicition”. So when writing in Italian does de Finetti >>> contrast “previsione” from “predizione”? Here once can see the >>> difference between seeing and speaking, which is somewhat masked in >>> English. Is that what de Finetti was driving at? >>> >>> Interested in discussion. >>> >>> Mik Bickis >>> >>>> On Jun 6, 2017, at 08:18 AM, Samuel Cohen <cohens(a)maths.ox.ac.uk> >>>> wrote: >>>> >>>> Dear all, >>>> >>>> If you are looking more generally at the history, it's unsurprising >>>> the word 'gamble' wasn't used much before the mid-20th century. >>>> While it was a common word in english since the 18th century (but >>>> was a colloquialism), it seems to have moved up in popularity post >>>> 1900 (see for example google ngrams). >>>> >>>> De Moivre, in arguably the first textbook on probability in >>>> English, uses 'wager' throughout (which is natural, as his >>>> attention is on games of chance). >>>> >>>> Best, >>>> Sam Cohen >>>> >>>> >>>> >>>> so On 06/06/17 14:01, Teddy Seidenfeld wrote: >>>>> Dear Mik, Erik, and Friends, >>>>> >>>>> I'm no scholar of the term 'gamble', but it's been used by many >>>>> decision theorists in closely related senses. >>>>> >>>>> For instance, Savage's (1954) section 5.2 (titled 'Gambles') uses >>>>> it within his theory (of subjective expected utility) to refer to >>>>> the equivalence class of simple acts that carry the same >>>>> (personal) probability distribution over consequences. Savage is >>>>> generalizing the von Neumann-Morgenstern sense of 'gamble'. >>>>> >>>>> I grew up with the von Neumann-Morgenstern's theory of cardinal >>>>> utility for gambles -- though vN-M use an extraneous concept of >>>>> 'probability' in contrast with Savage's approach. >>>>> >>>>> Best, >>>>> Teddy >>>>> >>>>> >>>>> >>>>> >>>>> >>>>> >>>>> On Mon, Jun 5, 2017 at 7:12 PM, Professor Bickis >>>>> <bickis(a)snoopy.usask.ca> wrote: >>>>> Hello: >>>>> >>>>> Can someone answer this question? >>>>> >>>>> Was it Walley who initiated the word “gamble” to use in place of >>>>> the classical “random variable”? De Finetti uses “random >>>>> quantity”, as does Williams. “Gamble” is ubiquitous in the >>>>> current literature, but I have not seen it anywhere prior to >>>>> Walley’s 1991 book. >>>>> Walley does not seem to attribute the term to anyone else. >>>>> >>>>> Mik Bickis >>>>> _______________________________________________ >>>>> SIPTA mailing list >>>>> SIPTA(a)idsia.ch >>>>> http://mailman2.ti-edu.ch/mailman/listinfo/sipta >>>>> >>>>> >>>>> >>>>> _______________________________________________ >>>>> SIPTA mailing list >>>>> >>>>> SIPTA(a)idsia.ch >>>>> http://mailman2.ti-edu.ch/mailman/listinfo/sipta >>>> >>>> _______________________________________________ >>>> SIPTA mailing list >>>> SIPTA(a)idsia.ch >>>> http://mailman2.ti-edu.ch/mailman/listinfo/sipta >>> >>> _______________________________________________ >>> SIPTA mailing list >>> SIPTA(a)idsia.ch >>> http://mailman2.ti-edu.ch/mailman/listinfo/sipta >> >> >> >> >> >> _______________________________________________ >> SIPTA mailing list >> SIPTA(a)idsia.ch >> http://mailman2.ti-edu.ch/mailman/listinfo/sipta > > _______________________________________________ > SIPTA mailing list > SIPTA(a)idsia.ch > http://mailman2.ti-edu.ch/mailman/listinfo/sipta

June 12, 2017

Re: [SIPTA] Question about the word "gamble" -- and "prevision"

by angelo.gilio＠sbai.uniroma1.it

Dear Professor Bickis (and dear Colleagues and friends), my message is just to confirm that the term 'prevision' used in the probability theory of de Finetti has a meaning completely different from 'prediction'. Operatively, the prevision of a random quantity X is the amount that you accept to pay (resp., to receive) in order to receive (resp., to pay) an amount equal to the value which will be assumed by X. To make a prediction on X is just trying to specify a value that in your opinion will be assumed by X; this is very different from assigning a prevision, because it may be that the prevision of X is not a possible value of X, as it only belongs to the convex hull of the possible values of X. All the best, Angelo Gilio > > > Dear Professor Bickis, > > In Italian the two words > > Previsione=Prevision and Predizione=Prediction have different acceptation. > > > (Devoto-Oli Dictionary) > > Previsione = Elaborazione mentale relativa al futuro, sulla base di > indizi piu' o meno sicuri > > (Mental thinking about the future based on more or less secure clues) > > > Example: Previsioni atmosferiche =Weather forecasts > > Predizione = previsione autorevole e solenne riferita come oggetto di > una previsione o profezia > > (Authoritative and solemn predictions referred as the subject of a > prediction or prophecy) > > > Best regards, > Serena > > > > Professor Bickis <bickis(a)snoopy.usask.ca> ha scritto: > >> Thanks for all your collective erudition. I am aware that “gamble” >> or its synonyms “wager” and “bet” have been used by many writers in >> the context of probability and decisions theory. I was thinking of >> the use of the word in the generic abstract sense that Walley uses. >> Since Walley’s development picks up on de Finetti and Williams in >> using “prevision”, I wondered if one of them had also used “gamble”, >> but it seems that they did not. >> >> Googling some Italian web sites on probability, I did come across >> the word “scommessa” in discussion of expectation/prevision, but I >> wonder if de Finetti himself used it in the sense that Walley used >> “gamble”. >> >> I am also a bit mystified by the history of the word “prevision”. >> Writing in French his 1937 paper, de Finetti uses “prévision” rather >> than the more common (even then) espérance. I would translate >> “prévision” as “predicition” in English. I see that in the 1974 >> English translation of de Finetti’s "Teoria de la Probabilità”, the >> translators note that they are using “prevision” (where earlier >> translators had used “foresight”), thereby making “prevision” less >> esoteric in English. However, in the same book, de Finetti makes a >> big deal that the “prevision is not prediction” (in the English >> translation). I would be interested if any Italian speakers out >> there could explain how de Finetti described this in his native >> language. I would render the French “prévision” as “predicition” in >> English, so naively would think that “previsione” in Italian also >> translates as “prediction”. Google Translate, however, gives me >> “predizione”, >> for “predicition”. So when writing in Italian does de Finetti >> contrast “previsione” from “predizione”? Here once can see the >> difference between seeing and speaking, which is somewhat masked in >> English. Is that what de Finetti was driving at? >> >> Interested in discussion. >> >> Mik Bickis >> >>> On Jun 6, 2017, at 08:18 AM, Samuel Cohen <cohens(a)maths.ox.ac.uk> >>> wrote: >>> >>> Dear all, >>> >>> If you are looking more generally at the history, it's unsurprising >>> the word 'gamble' wasn't used much before the mid-20th century. >>> While it was a common word in english since the 18th century (but >>> was a colloquialism), it seems to have moved up in popularity post >>> 1900 (see for example google ngrams). >>> >>> De Moivre, in arguably the first textbook on probability in >>> English, uses 'wager' throughout (which is natural, as his >>> attention is on games of chance). >>> >>> Best, >>> Sam Cohen >>> >>> >>> >>> so On 06/06/17 14:01, Teddy Seidenfeld wrote: >>>> Dear Mik, Erik, and Friends, >>>> >>>> I'm no scholar of the term 'gamble', but it's been used by many >>>> decision theorists in closely related senses. >>>> >>>> For instance, Savage's (1954) section 5.2 (titled 'Gambles') uses >>>> it within his theory (of subjective expected utility) to refer to >>>> the equivalence class of simple acts that carry the same >>>> (personal) probability distribution over consequences. Savage is >>>> generalizing the von Neumann-Morgenstern sense of 'gamble'. >>>> >>>> I grew up with the von Neumann-Morgenstern's theory of cardinal >>>> utility for gambles -- though vN-M use an extraneous concept of >>>> 'probability' in contrast with Savage's approach. >>>> >>>> Best, >>>> Teddy >>>> >>>> >>>> >>>> >>>> >>>> >>>> On Mon, Jun 5, 2017 at 7:12 PM, Professor Bickis >>>> <bickis(a)snoopy.usask.ca> wrote: >>>> Hello: >>>> >>>> Can someone answer this question? >>>> >>>> Was it Walley who initiated the word “gamble” to use in place of >>>> the classical “random variable”? De Finetti uses “random >>>> quantity”, as does Williams. “Gamble” is ubiquitous in the >>>> current literature, but I have not seen it anywhere prior to >>>> Walley’s 1991 book. >>>> Walley does not seem to attribute the term to anyone else. >>>> >>>> Mik Bickis >>>> _______________________________________________ >>>> SIPTA mailing list >>>> SIPTA(a)idsia.ch >>>> http://mailman2.ti-edu.ch/mailman/listinfo/sipta >>>> >>>> >>>> >>>> _______________________________________________ >>>> SIPTA mailing list >>>> >>>> SIPTA(a)idsia.ch >>>> http://mailman2.ti-edu.ch/mailman/listinfo/sipta >>> >>> _______________________________________________ >>> SIPTA mailing list >>> SIPTA(a)idsia.ch >>> http://mailman2.ti-edu.ch/mailman/listinfo/sipta >> >> _______________________________________________ >> SIPTA mailing list >> SIPTA(a)idsia.ch >> http://mailman2.ti-edu.ch/mailman/listinfo/sipta >> > > > > > > _______________________________________________ > SIPTA mailing list > SIPTA(a)idsia.ch > http://mailman2.ti-edu.ch/mailman/listinfo/sipta >

June 10, 2017

Re: [SIPTA] Question about the word "gamble" -- and "prevision"

by Serena Doria

Dear Professor Bickis, In Italian the two words Previsione=Prevision and Predizione=Prediction have different acceptation. (Devoto-Oli Dictionary) Previsione = Elaborazione mentale relativa al futuro, sulla base di indizi piu' o meno sicuri (Mental thinking about the future based on more or less secure clues) Example: Previsioni atmosferiche =Weather forecasts Predizione = previsione autorevole e solenne riferita come oggetto di una previsione o profezia (Authoritative and solemn predictions referred as the subject of a prediction or prophecy) Best regards, Serena Professor Bickis <bickis(a)snoopy.usask.ca> ha scritto: > Thanks for all your collective erudition. I am aware that “gamble” > or its synonyms “wager” and “bet” have been used by many writers in > the context of probability and decisions theory. I was thinking of > the use of the word in the generic abstract sense that Walley uses. > Since Walley’s development picks up on de Finetti and Williams in > using “prevision”, I wondered if one of them had also used “gamble”, > but it seems that they did not. > > Googling some Italian web sites on probability, I did come across > the word “scommessa” in discussion of expectation/prevision, but I > wonder if de Finetti himself used it in the sense that Walley used > “gamble”. > > I am also a bit mystified by the history of the word “prevision”. > Writing in French his 1937 paper, de Finetti uses “prévision” rather > than the more common (even then) espérance. I would translate > “prévision” as “predicition” in English. I see that in the 1974 > English translation of de Finetti’s "Teoria de la Probabilità”, the > translators note that they are using “prevision” (where earlier > translators had used “foresight”), thereby making “prevision” less > esoteric in English. However, in the same book, de Finetti makes a > big deal that the “prevision is not prediction” (in the English > translation). I would be interested if any Italian speakers out > there could explain how de Finetti described this in his native > language. I would render the French “prévision” as “predicition” in > English, so naively would think that “previsione” in Italian also > translates as “prediction”. Google Translate, however, gives me > “predizione”, > for “predicition”. So when writing in Italian does de Finetti > contrast “previsione” from “predizione”? Here once can see the > difference between seeing and speaking, which is somewhat masked in > English. Is that what de Finetti was driving at? > > Interested in discussion. > > Mik Bickis > >> On Jun 6, 2017, at 08:18 AM, Samuel Cohen <cohens(a)maths.ox.ac.uk> wrote: >> >> Dear all, >> >> If you are looking more generally at the history, it's unsurprising >> the word 'gamble' wasn't used much before the mid-20th century. >> While it was a common word in english since the 18th century (but >> was a colloquialism), it seems to have moved up in popularity post >> 1900 (see for example google ngrams). >> >> De Moivre, in arguably the first textbook on probability in >> English, uses 'wager' throughout (which is natural, as his >> attention is on games of chance). >> >> Best, >> Sam Cohen >> >> >> >> so On 06/06/17 14:01, Teddy Seidenfeld wrote: >>> Dear Mik, Erik, and Friends, >>> >>> I'm no scholar of the term 'gamble', but it's been used by many >>> decision theorists in closely related senses. >>> >>> For instance, Savage's (1954) section 5.2 (titled 'Gambles') uses >>> it within his theory (of subjective expected utility) to refer to >>> the equivalence class of simple acts that carry the same >>> (personal) probability distribution over consequences. Savage is >>> generalizing the von Neumann-Morgenstern sense of 'gamble'. >>> >>> I grew up with the von Neumann-Morgenstern's theory of cardinal >>> utility for gambles -- though vN-M use an extraneous concept of >>> 'probability' in contrast with Savage's approach. >>> >>> Best, >>> Teddy >>> >>> >>> >>> >>> >>> >>> On Mon, Jun 5, 2017 at 7:12 PM, Professor Bickis >>> <bickis(a)snoopy.usask.ca> wrote: >>> Hello: >>> >>> Can someone answer this question? >>> >>> Was it Walley who initiated the word “gamble” to use in place of >>> the classical “random variable”? De Finetti uses “random >>> quantity”, as does Williams. “Gamble” is ubiquitous in the >>> current literature, but I have not seen it anywhere prior to >>> Walley’s 1991 book. >>> Walley does not seem to attribute the term to anyone else. >>> >>> Mik Bickis >>> _______________________________________________ >>> SIPTA mailing list >>> SIPTA(a)idsia.ch >>> http://mailman2.ti-edu.ch/mailman/listinfo/sipta >>> >>> >>> >>> _______________________________________________ >>> SIPTA mailing list >>> >>> SIPTA(a)idsia.ch >>> http://mailman2.ti-edu.ch/mailman/listinfo/sipta >> >> _______________________________________________ >> SIPTA mailing list >> SIPTA(a)idsia.ch >> http://mailman2.ti-edu.ch/mailman/listinfo/sipta > > _______________________________________________ > SIPTA mailing list > SIPTA(a)idsia.ch > http://mailman2.ti-edu.ch/mailman/listinfo/sipta >

June 7, 2017

Re: [SIPTA] Question about the word "gamble" -- and "prevision"

by Gert De Cooman

June 7, 2017

Re: [SIPTA] Question about the word "gamble" -- and "prevision"

by Professor Bickis

Thanks for all your collective erudition. I am aware that “gamble” or its synonyms “wager” and “bet” have been used by many writers in the context of probability and decisions theory. I was thinking of the use of the word in the generic abstract sense that Walley uses. Since Walley’s development picks up on de Finetti and Williams in using “prevision”, I wondered if one of them had also used “gamble”, but it seems that they did not. Googling some Italian web sites on probability, I did come across the word “scommessa” in discussion of expectation/prevision, but I wonder if de Finetti himself used it in the sense that Walley used “gamble”. I am also a bit mystified by the history of the word “prevision”. Writing in French his 1937 paper, de Finetti uses “prévision” rather than the more common (even then) espérance. I would translate “prévision” as “predicition” in English. I see that in the 1974 English translation of de Finetti’s "Teoria de la Probabilità”, the translators note that they are using “prevision” (where earlier translators had used “foresight”), thereby making “prevision” less esoteric in English. However, in the same book, de Finetti makes a big deal that the “prevision is not prediction” (in the English translation). I would be interested if any Italian speakers out there could explain how de Finetti described this in his native language. I would render the French “prévision” as “predicition” in English, so naively would think that “previsione” in Italian also translates as “prediction”. Google Translate, however, gives me “predizione”, for “predicition”. So when writing in Italian does de Finetti contrast “previsione” from “predizione”? Here once can see the difference between seeing and speaking, which is somewhat masked in English. Is that what de Finetti was driving at? Interested in discussion. Mik Bickis > On Jun 6, 2017, at 08:18 AM, Samuel Cohen <cohens(a)maths.ox.ac.uk> wrote: > > Dear all, > > If you are looking more generally at the history, it's unsurprising the word 'gamble' wasn't used much before the mid-20th century. While it was a common word in english since the 18th century (but was a colloquialism), it seems to have moved up in popularity post 1900 (see for example google ngrams). > > De Moivre, in arguably the first textbook on probability in English, uses 'wager' throughout (which is natural, as his attention is on games of chance). > > Best, > Sam Cohen > > > > so On 06/06/17 14:01, Teddy Seidenfeld wrote: >> Dear Mik, Erik, and Friends, >> >> I'm no scholar of the term 'gamble', but it's been used by many decision theorists in closely related senses. >> >> For instance, Savage's (1954) section 5.2 (titled 'Gambles') uses it within his theory (of subjective expected utility) to refer to the equivalence class of simple acts that carry the same (personal) probability distribution over consequences. Savage is generalizing the von Neumann-Morgenstern sense of 'gamble'. >> >> I grew up with the von Neumann-Morgenstern's theory of cardinal utility for gambles -- though vN-M use an extraneous concept of 'probability' in contrast with Savage's approach. >> >> Best, >> Teddy >> >> >> >> >> >> >> On Mon, Jun 5, 2017 at 7:12 PM, Professor Bickis <bickis(a)snoopy.usask.ca> wrote: >> Hello: >> >> Can someone answer this question? >> >> Was it Walley who initiated the word “gamble” to use in place of the classical “random variable”? De Finetti uses “random quantity”, as does Williams. “Gamble” is ubiquitous in the current literature, but I have not seen it anywhere prior to Walley’s 1991 book. >> Walley does not seem to attribute the term to anyone else. >> >> Mik Bickis >> _______________________________________________ >> SIPTA mailing list >> SIPTA(a)idsia.ch >> http://mailman2.ti-edu.ch/mailman/listinfo/sipta >> >> >> >> _______________________________________________ >> SIPTA mailing list >> >> SIPTA(a)idsia.ch >> http://mailman2.ti-edu.ch/mailman/listinfo/sipta > > _______________________________________________ > SIPTA mailing list > SIPTA(a)idsia.ch > http://mailman2.ti-edu.ch/mailman/listinfo/sipta

June 6, 2017

Re: [SIPTA] Question about the word "gamble"

by Samuel Cohen

June 6, 2017

Re: [SIPTA] Question about the word "gamble"

by Teddy Seidenfeld

June 6, 2017

Re: [SIPTA] Question about the word "gamble"

by Erik Quaeghebeur

Dear Mik, Dear list, > Was it Walley who initiated the word “gamble” to use in place > of the classical “random variable”? […], but I have not seen it > anywhere prior to Walley’s 1991 book. It dates from prior to his book. Already together with Fine he uses the term in their '79 paper “Varieties of modal (classificatory) and comparative probability” <https://link.springer.com/article/10.1007%2FBF00869449>. Furthermore, Fine uses the term in his '73 book “Theories of probability: an examination of foundations” <https://books.google.com/books?id=U9lEAAAAIAAJ&q=gamble>. I don't have access to the book's full text here, so going back further in this line is for someone else. So, no, Walley did not initiate its use. Of course others have used the word ‘gambling’ in the context before (e.g., Smith in '61 <https://www.jstor.org/stable/2983842>), but I guess you're focussing on the term ‘gamble’ as formalized by Walley, then for now we have Fine '73. Best, Erik -- https://ac.erikquaeghebeur.name

June 6, 2017

[SIPTA] Question about the word "gamble"

by Professor Bickis

Hello: Can someone answer this question? Was it Walley who initiated the word “gamble” to use in place of the classical “random variable”? De Finetti uses “random quantity”, as does Williams. “Gamble” is ubiquitous in the current literature, but I have not seen it anywhere prior to Walley’s 1991 book. Walley does not seem to attribute the term to anyone else. Mik Bickis

June 5, 2017