If you check the section on "Expected utility" (i.e EU) theory, you basically get ('in formal speak') a rendering of what "payoff" looks like (in a strictly Bayesian, non-normative) sense.
The section on "Unawareness" and Sequental Decisions also have some nice gems in there with regard to what you dub "uncertainty". Decision Theory has a broader outlook on what that means and this essay sort of looks at the beginnings of that.
One of the guys in my gaming group has a PhD in game theory. This article grew out of a lot of lunch-time conversations with him. The sort of guy who'll notice that you're ordering the same meal you had last week to avoid the risk that the other dish might not be as good and say "Ah I see you are engaging in counter-factual regret minimization today."
There are definitely Key overlaps in Decision Theory vs Game Theory; however the latter is concerned moreso with *just* the Descriptive. Whereas the former is looking at the Normative & the Prescriptive.
Most Game theorists today have a "just calculate" attitude to a lot of key pillars in their discipline. You have to go "meta" and step back from that & ask "Ok, but WHY this sequence of starting points & not this other one?"
A strategist who knows he's less skilled than his opponent might seek to increase the uncertainty from chance since it partially negates the value of skill.
A strategist in a bad position might seek to increase the general uncertainty and hope to get lucky.
Oooh, that's a mighty fine definition of strategy! It's admirably more elegant than what I've been using for the past few years: "Strategy is the art of coordinating lower-order means to achieve higher-order goals in the face of uncertainty."
Mostly I included the distinction between the order-of-effect you're directly controlling vs the order-of-effect you're after to have a way to meaningfully distinguish between strategy and tactics. The "coordinate" bit is to capture that it might be a bit iffy to call something a "strategy" if you can just decide to do it and make it happen yourself, but that may be leaning too hard on the original meaning of "doing what a general does."
Your four sources of uncertainty are also incredibly helpful as a method of categorization.
Very much looking forward to future pieces building on this!
Thanks! I should also have added that one of the reasons mine is a little more fiddly is that I teach at a business school, and as I'm sure you remember, few places overuse the word more, so I wanted to make sure I at least had it clear for myself.
Speaking of games, I was pondering what is an example of a game that involves pure chance, pure complexity, pure choice and pure confusion?
One that comes to mind, having devoted WAY too much time to it in my youth, mind you, is Dungeons & Dragons. The outcome of each game is determined by a combination of player skill, dice rolls, and the decisions of the game master. The proficiency of the GM can influence the complexity of the sessions and force the players to strategize in a way that accounts for chance, complexity, choice and confusion as well as balancing other factors such as roleplaying with the restriction that your character has a certain ethical alignment.
I strongly agree! I still spend WAY too much time on D&D (or, nowadays, my own variant, ACKS). I consider it one of the best tools for developing strategic skills there is.
Great article! I was a member of the West Point Wargames Committee when I was at US Military Academy and we played DnD every weekend for 12 hours at a time. It is absolutely a tool of great value.
"Provided that the probabilities or payoffs of the choices do not change after each choice is made, playing one and many iterations of a Chance-based game will not affect the strategist’s decision-making."
This is true if payoffs are construed very broadly. Typically though payoffs are consistent in some resource (eg money) but sublinear in your enjoyment of that resource. (eg, money seems to provide value to people in proprtion to it's logarithm.)
And there's ergodicity conceens too. It's been a while though and I'm not sure if those can be strictly modeled as similar to the previous case or not.
The game most relevant for our age is Illuminati. (Risk with factions and money vs. countries and armies. But the topology differs every game.)
Very well written! Relevant:
https://plato.stanford.edu/entries/decision-theory/#WhaPreOvePro
If you check the section on "Expected utility" (i.e EU) theory, you basically get ('in formal speak') a rendering of what "payoff" looks like (in a strictly Bayesian, non-normative) sense.
The section on "Unawareness" and Sequental Decisions also have some nice gems in there with regard to what you dub "uncertainty". Decision Theory has a broader outlook on what that means and this essay sort of looks at the beginnings of that.
Enjoy!
One of the guys in my gaming group has a PhD in game theory. This article grew out of a lot of lunch-time conversations with him. The sort of guy who'll notice that you're ordering the same meal you had last week to avoid the risk that the other dish might not be as good and say "Ah I see you are engaging in counter-factual regret minimization today."
Interesting!
There are definitely Key overlaps in Decision Theory vs Game Theory; however the latter is concerned moreso with *just* the Descriptive. Whereas the former is looking at the Normative & the Prescriptive.
Most Game theorists today have a "just calculate" attitude to a lot of key pillars in their discipline. You have to go "meta" and step back from that & ask "Ok, but WHY this sequence of starting points & not this other one?"
To do that properly, Decision Theory is needed.
I am going to give him grief on this next time I see him and accuse him of being bad at decisions. Hehe
Got to figure out a way to use "counter-factual regret minimization" in a PowerPoint.
Some other examples:
A strategist who knows he's less skilled than his opponent might seek to increase the uncertainty from chance since it partially negates the value of skill.
A strategist in a bad position might seek to increase the general uncertainty and hope to get lucky.
Great points. I have absolutely done the latter in various wargames etc.
US foregin policy?
Oooh, that's a mighty fine definition of strategy! It's admirably more elegant than what I've been using for the past few years: "Strategy is the art of coordinating lower-order means to achieve higher-order goals in the face of uncertainty."
Mostly I included the distinction between the order-of-effect you're directly controlling vs the order-of-effect you're after to have a way to meaningfully distinguish between strategy and tactics. The "coordinate" bit is to capture that it might be a bit iffy to call something a "strategy" if you can just decide to do it and make it happen yourself, but that may be leaning too hard on the original meaning of "doing what a general does."
Your four sources of uncertainty are also incredibly helpful as a method of categorization.
Very much looking forward to future pieces building on this!
Your definition seems rock-solid to me!
Thanks! I should also have added that one of the reasons mine is a little more fiddly is that I teach at a business school, and as I'm sure you remember, few places overuse the word more, so I wanted to make sure I at least had it clear for myself.
Waiting for part 2 with bated breath!
>Poker is game where uncertainty arises from Choice, Complexity, Chance, and Confusion.
They say that poker was Von Neumann's inspiration for game theory
I didn't know that, but that makes good sense.
Speaking of games, I was pondering what is an example of a game that involves pure chance, pure complexity, pure choice and pure confusion?
One that comes to mind, having devoted WAY too much time to it in my youth, mind you, is Dungeons & Dragons. The outcome of each game is determined by a combination of player skill, dice rolls, and the decisions of the game master. The proficiency of the GM can influence the complexity of the sessions and force the players to strategize in a way that accounts for chance, complexity, choice and confusion as well as balancing other factors such as roleplaying with the restriction that your character has a certain ethical alignment.
I strongly agree! I still spend WAY too much time on D&D (or, nowadays, my own variant, ACKS). I consider it one of the best tools for developing strategic skills there is.
There's a case to be made that TTRPGs like DnD can be useful tools for practising decision making under uncertainty.
https://warontherocks.com/2023/08/embrace-the-nerd-dungeons-dragons-and-military-intelligence/
Great article! I was a member of the West Point Wargames Committee when I was at US Military Academy and we played DnD every weekend for 12 hours at a time. It is absolutely a tool of great value.
*decision making or decision support (intelligence falls in the latter category)
A lot of people like myself got into statistics (or had an easier time learning it) in college because of D&D introducing me to statistics and odds.
Same!
Excellent write up. Nothing to add.
I like this new topic. Most of what I know about decision making under uncertainty comes from Taleb's books.
Thoughts on the game of Go?
I've never played it!
"Provided that the probabilities or payoffs of the choices do not change after each choice is made, playing one and many iterations of a Chance-based game will not affect the strategist’s decision-making."
This is true if payoffs are construed very broadly. Typically though payoffs are consistent in some resource (eg money) but sublinear in your enjoyment of that resource. (eg, money seems to provide value to people in proprtion to it's logarithm.)
And there's ergodicity conceens too. It's been a while though and I'm not sure if those can be strictly modeled as similar to the previous case or not.
I appreciated the taxonomy of uncertainty though. Enjoyable post.
Thanks!