Game Theory in Management

I think a lot of what passes for advanced Communications Management (ProjectManagement.com’s theme for February) can challenged by employing the Game Theorists’ favorite technique, the Payoff Grid, as it applies to the feudal era role known as the court jester, or fool. As a point of reference, although it is commonly assumed that the court jester’s primary duty was to entertain whatever court he was assigned, in many (if not most) royal courts the jester’s main purpose was to criticize the decisions made by the aristocracy, up to and including the king or queen themselves (e.g., the “Fool” character in King Lear). While offering such criticisms would be typically fatal to the dukes, earls, barons, thanes, or any other land-owning political players at court, the fool, who could clearly never represent a genuine threat to the king’s position, had some level of protection against retribution for his words to the actual emperor/empress. In short, if the king was about to pursue a clearly misguided action, but the powerful members of the palace were afraid to point out the folly, the court jester/fool would be the one to challenge such plans, relying on the convention of non-retribution for the words he spoke.

Meanwhile, Back In The Project Management World…

The analogy of the feudal court to the modern-day executive boardroom is a pretty easy one to make. In those instances where an executive is pursuing an objective with a technical approach that is clearly insufficient, or even destructive to the organization or project team, and that executive has power over the trajectory of his subordinate managers’ career paths, whom among them, bereft of the protection afforded medieval fools, would be willing to give voice to the needed challenges or criticisms? I believe that much of the Communication Managers’ assertions concerning the need to “engage all stakeholders” has its basis in avoiding this exact scenario, under the assumption that someone among the “stakeholders” who is not beholden to the person setting the technical agenda would be in a position to offer the needed challenge, thereby preventing the wasteful or destructive agenda from being pursued in the first place. Without reviewing the potential efficacy of this Communications Management tactic on its face, I would like to turn reader attention to its opposite, as shown by the following Payoff Grid:

As is the case with most Payoff Grids, two of the Scenarios (in this case, B and C) reveal a completely acceptable state of affairs. If the technical agenda is bad, and someone in a position to influence its pursuit points that out, as in Scenario B, it’s an appropriate communication. Similarly, if the technical agenda is fine, and no one puts forth a challenge or criticism of it, as in Scenario C, all is well. As I stated earlier, I believe that much of the Communication Managers’ codex is oriented towards avoiding Scenario D, which raises the question: what happens when those techniques are employed, but the situation on the ground is, in fact, more analogous to Scenario A?

Merriam-Webster defines “twenty-twenty hindsight” as

the full knowledge and complete understanding that one has about an event only after it has happened.[i]

I understand that this phrase is most often invoked after a decision was made, and the evidence that it was the incorrect one becomes obvious. But my target is the other side of the Payoff Grid. I get that, in the aftermath of a poor decision, the decision-makers are subject to criticism, some of it unfair, as virtually all serious management decisions are made based on incomplete information. My focus in on those instances where the optimal technical approach (and, in most cases, the accompanying implementation strategy) have been selected, but prominent critics challenge or criticize it, seeking to prevent its use. These can be especially damaging in those cases where a truly novel approach is being employed, with no previous success stories to justify its selection. In these instances, the inappropriately articulated challenges and criticisms end up harming the development and implementation of innovative PM techniques, all because of a lack of commonly-accepted criteria to cut off “stakeholder input” once the PM has selected the technical approach.

In short, some of the stakeholders’ input can be reliably used in forging the path to the identification of the optimal technical approach to resolving the project’s central problem, and other input can be rather neutral, neither helping nor harming. But some of this input can actually be detrimental, should it be allowed to gain traction, and the people pushing it won’t dress in stripes and funny hats to tip you off. This is something the PM will have to discern on their own, but can’t do so if they’re not aware.

Now you’re aware.

[i] Retrieved from https://www.merriam-webster.com/dictionary/twenty-twenty%20hindsight on February 19, 2023, 10:32 MST.

Back when I was spending a lot more time in classroom settings, teaching new Project Controls Specialists how to do their jobs, I would devote a good five minutes within the Performance Measurement module to discuss one particular aspect of preparing the documents germane to PM, specifically Baseline Change Proposals and Variance Analysis Reports. When I came to this particular point in the session, I’d turn off the projector, walk over to the white board, and write the word “his” on it.  Then I’d pick out someone in the class (usually the person who appeared to be least engaged) and ask them to read the word, which they would do.

“Is it misspelled?” I would challenge.

“No.”

Then I’d write the words “hers,” “theirs,” and “ours,” followed by the same questions, and getting the same answers, as the class slowly became more curious about this line of inquiry. Then, after erasing the previous words, I’d write the word “its.” It was at this point that most of the class would laugh, finally seeing where I was going with this little exercise.

“You’re laughing now,” I would begin, “but I can virtually guarantee that, within the coming week, I will see at least one instance of the impersonal possessive pronoun being spelled ‘it’s,’ in either a VAR or a BCP. For the record, ‘it’s’ is a contraction of the words ‘it is,’ which is something that we all should have learned definitively prior to graduating seventh grade. And, before anyone accuses me of being an insufferable grammar police nitpicker, which may or may not be a reasonable assessment of my persona, I would like to point out that I’m not the person for whom these communications are meant. It’s the customers who will likely make some rather unfortunate, negative inferences if they see errors like this. Remember, they rely on these documents to reliably inform them of the cost and schedule status of their project. If any aspect of these reports indicate that their author wasn’t paying attention in seventh grade English class, why should any of them believe that the same person understands the difference between a Cost Variance and a Schedule Variance? Or the proper way of computing an Estimate at Completion?”

The old writers’ admonition to consider the audience prior to picking up the pen has an additional layer when it comes to writing the documents germane to PM. Using the previously-stated examples of the Baseline Change Proposal and Variance Analysis Report, what are going to be the typical customer’s concerns when it comes to the reliability of these communications? I believe that

• …for the BCP, they may be on guard against a contractor attempting to gain an adjustment for some added cost that, in reality, was due to something the contractor could control, such as performance.
• …with something similar for the VAR, but more along the lines of shifting blame for negative performance away from the contractor, and towards unmanageable or unforeseen events.

To be clear, I’m not saying that this is what’s going on when these documents are prepared and presented, only that it’s at least a possibility that these concerns are in the back of the clients’ minds when they see them. In order to quell these fears, it’s incumbent on these communications’ preparers to write them in such a way as to demonstrate (a) a command of the issues and events leading to the BCP/VAR being created in the first place, (b) an acceptable level of expertise in PM strategies and techniques, and (c) sufficient proficiency in writing to convey both (a) and (b). Engaging in obfuscation or muddy writing is an indicator that the ideas being “conveyed” haven’t been thoroughly vetted, or thought through sufficiently to establish reasonable causation.

Whenever I’m on the hook to produce PM-specific documents, I try to remember to be mindful of voice. Psychiatrist Eric Berne, in his best-seller Games People Play (Grove Press, 1964), theorized three parts of the persona, namely Child, Adult, and Parent, and went on to posit that interpersonal conflict arises when the lines of communication between people become mis-aligned with these parts. For example, if a member of GTIM Nation were to post a comment on a technical aspect of one of my points, that would represent that person’s Adult communicating to my Adult. However, if I were to respond condescendingly, challenging that person’s education or depth of experience, that would represent my Parent addressing their Child, thereby crossing the lines within that transaction, and most likely leading to conflict. This aspect of PM communications may be the primary reason I find many paper presentations on the subject to be off-putting, as more than a few first-timers present a voice of absolute intellectual superiority as they spout long-known elements of PM, an approach I’ve nicknamed eat-your-peas-style hectoring. Their voice is that of a Parent scolding a Child, almost certain to generate resentment on the part of the listeners.

So, I’ll ask again: When preparing a PM-centric communication, who do you think you’re talking or writing to? And, after you’ve taken a shot at answering that question, one more remains: Is that really who you will end up talking to?

For those members of GTIM Nation who aren’t into Shakespeare in general or the history plays in particular, the quote in this blog’s title comes from the very beginning of Henry IV, Part II. At the end of Part I, King Henry and his son, the “madcap” Prince of Wales, have put down a rebellion by the powerful Percy family, led by Henry Percy, nicknamed Hotspur. At the beginning of Part II, the Earl of Northumberland, Hotspur’s father, awaits the news of what happened at the Battle of Shrewsbury. The first news he hears is all good – but it isn’t delivered by someone who witnessed the battle. Similarly, the second messenger to arrive wasn’t at the battle, but did encounter a horseman who (apparently) was. The horseman’s version is somewhat cryptic: that “young Harry Percy’s spur was cold.[i]” As the Earl and the two early messengers contemplate what that reference could mean, the third messenger, who was an eyewitness to the battle arrives, and tells of the true fate of the rebellion, which is, of course, the exact opposite of the earlier reports. In essence, Northumberland knew what he wanted to hear, and the first two messengers he encounters knew what he wanted to hear, so…

Meanwhile, Back In The Project Management World…

If Communications (ProjectManagement.com’s theme for February) as an arena of Project Management theory is to be subjected to a truly difficult stress test, I believe it would have to be in the theater of the Project Review. And what do we hear in Project Reviews? The cost and schedule performance of all the work in the portfolio, of course, especially and particularly the answer to the overriding question: Will each of the projects come in on-time, on-budget? The individual PMs know what the Program/Portfolio managers want to hear, and that’s that everything is going swimmingly – which is also the exact thing that the PMs themselves wish to report. Consider the following payoff grid:

Since Scenarios B and C represent accurate assessments of what’s going on in the Project presenting at the review, no inherent conflict in the information stream itself is indicated. We’ll also dispense with Scenario A, since no rational PM would report the Project in difficulty if it was, in fact, performing just fine, barring some ulterior motive. This leaves us with Scenario D, where the Project is in difficulty, but the PM would rather not report that, at least not at that time. The summary in the payoff grid indicates that their problems will multiply exponentially. This is due to the fact that any assets in reserve that would otherwise go to the Projects in trouble will be sent to those Projects that report that they are in trouble (Scenario C), not to those that fail to report as such. With whatever factors leading to the poorly-performing Project not addressed, or even recognized, they will continue to manifest at the very least, until they simply cannot be ignored.

Why would this happen? A variety of reasons, including the notion that whatever problems have arisen can be dealt with before they have an at-completion impact and, even if they can’t, the last thing the PM needs is a bunch of execs looking over the shoulder and kibbitzing on what strategies or techniques should be employed. If the negative consequences of being in charge of a poorly-performing Project are basically the same if those problems are brought to light early, or closer to that Project’s baseline finish date, what would be the motive for the PM to disclose those difficulties before it’s absolutely necessary?

Of course, a properly functioning Earned Value Management System (EVMS) would make this sort of legerdemain very difficult. Even that most basic of Estimate at Completion (EAC) formula, dividing the percent complete into cumulative actual costs, will yield a reliable (typically within ten points once the work is past the 20% complete point) figure rendering Scenario D highly problematic to the less-than-upfront PM. This fact may be the foundation for the calculated EAC’s unpopularity in the Project Review venue. It’s just so, well, inconvenient. Much easier to bamboozle with the so-called bottoms-up EAC (cumulative actual costs plus re-estimated remaining budget) or, better yet, call in the risk managers (no initial caps), who can muddy the epistemological waters like no others.

Who can capture the magnitude of this sort of misdirection? I’ll leave it to The Bard.

From Rumor’s tongues
They bring smooth comforts false, worse than
true wrongs.[ii]

[i] Shakespeare, William, Henry IV, Part II, Act I, Scene I.

[ii] Ibid.

I remember reading about one aspect of human intelligence having to do with the ability to encounter novel situations and dealing with them appropriately by remembering a similar circumstance, and either engaging the actions or decisions (with or without some level of adaptation) that were made in the recalled instance that resulted in a successful outcome, or avoiding those decisions that ended in failure. This aspect is actually the basis for the utility of Game Theory. When we learn about the Nash Equilibrium in, say, the Hawk/Dove Game, we’re not studying the various scenarios involved in order to get a handle on the amount of food ingested by a population of entirely fictitious birds, as fun as that might sound. No, the point of performing this type of analysis rests entirely on the assumption that, at some level and capacity, the interactions of the particular games’ players are analogous to real-life interactions among human players, typically in a business setting.

Some of these games can have a strong analogous relationship with their real-life counterparts. In the movie Midway (2019), a scene has Admiral Yamamoto walk in on a war game conducted by Admiral Nagumo, who is sharply criticizing his junior officers for not following some of the basic guidelines stipulated at the beginning of the game. Yamamoto asks for an explanation, and one of the offending officers explains that they had arranged for the American carrier forces to be positioned northeast of Midway, and from that vantage point those forces had surprised and sunk three of the four Japanese carriers. Nagumo points out that the only way the American forces could be so positioned would be if they had advanced notification of the Japanese intent, which was held to be an impossibility. Yamamoto orders them to re-start the war game, this time working under the assumption that the Americans did not know the Japanese intent, and could not position their forces accordingly. Of course, at the actual Battle of Midway, the Americans had intercepted almost the entire Japanese order of battle, their carrier forces were positioned northeast of Midway, they did surprise the Japanese forces, and actually sank three of the four carriers in the first attack wave.

Conversely, I also remember reading an article that described the reasons why the Tit-for-Tat strategy is so successful in multiple iterations of the Prisoner’s Dilemma, and then went on to use that analysis as an explanation for why entire nations go to war. While the Prisoner’s Dilemma may be the go-to game for analyzing cooperation and defection among non-related biological units in a common environment, I’m not at all sure insights gleaned from it can be reliably used in identifying appropriate geopolitical conflict strategies.

Even within some of the games, extracting an insight on a usable approach can prove to be very tricky. Take the Ultimatum Game as an example. In this game, the researcher/game coordinator approaches two people (who presumably know each other) with the following offer: he will give them $100 (USD) if Person #1 can propose how the money will be distributed among them and have Person #2 approve of that distribution scheme on the first try. Some Game Theory experts predicted that, in order for Person #1 to maximize their payout, they should name a distribution of $99 for themselves, and only $1 for Person #2, on the grounds that Person #2 is essentially looking at receiving either $1, or nothing at all.  When this game was tried with real people, however, that distribution was virtually never accepted. Instead of taking away from this the rather obvious challenge to the Economic Man theory (that people are rational, and will always act in their own self-interest), analysts were content to blame other factors, such as “culture.” But that’s the real irony here – that the results of using the 99-to-1 strategy, predicated on the Economic Man theory, failed repeatedly, meaning that, apart from the obvious takeaway, this game really had very little similarity to its real-life application. It wasn’t very analogous at all. 

I can almost hear GTIM Nation now: “Actually, Michael, you’re the only person I hear going on and on about Game Theory, suitable to the business environment at hand or not.” Fair enough. But the decision-makers who are carrying with them ideas, strategies, and business models that may or may not be applicable to their new surroundings aren’t announcing their thought processes as pertaining to Game Theory per se. They’re usually working off of a combination of education and experience. I can’t count the number of times I’ve encountered a new manager or co-worker who makes it a point to describe how “things” were “done” at their previous assignment, to great success, slyly cruising past any discussion of (1) if things were so swell at their previous assignment, why did they leave?, (2) even within the previous environment, was any thought given to evaluating (much less scoring) other ideas, strategies, or business models to help ensure the one used was the optimal one?, and (3) why is this person convinced that, even if the previous questions could be answered satisfactorily, the current situation be considered analogous to the point of an idea/strategy/business model being given the assumption of automatic success?

Strategies gleaned from previous successes are great, when the circumstances re-present as highly analogous to that managerial victory. When the analogy is weak, however, the new parallel being drawn may very well be sort of exactly like failure.

Sherlock Holmes was known for being taken by surprise by references his associate, Dr. Watson, would make to what could probably be best described as pop culture in Victorian England. Should Watson express amazement that Holmes would be ignorant of such things, Sherlock would explain his desire to not occupy his mind with trivial things in order to maximize his effectiveness as a consulting detective. If one could isolate the kind of mental filter that would function as a differentiator between data items that contribute to Holmes’ infamous deduction capabilities and ones that do not, I believe that such a filter could have but one name: relevance.

Meanwhile, Back In The Project Management World…

GTIM Nation may be aware of one of my favorite axioms, “All things fail by irrelevant comparisons.” Combine this with one of Hatfield’s Incontrovertible Rules of Management (#22):

All useful management information has the following three characteristics:

• • It is accurate,
  • It is timely,
  • …and it is relevant.

This is what irks me about much of what passes for insightful management techniques, particularly within the PM domain. They can be rather irrelevant. What should be the test for relevance, at least within PM space? I think simply asking the question “does this (technique, information stream, business practice) help bring the project in on-time, on-budget?” is a great starting point. While asserting that this question is simple, I also readily acknowledge that discerning relevance is multi-layered. Prior to using it as a management technique litmus test, let’s add one key element of structure, the old saw “Quality, Availability, Affordability: pick any two” (original author unknown). With this structure and question in-hand, let’s take a look at some business practices that may or may not have been relevant.

First up: Six Sigma Quality Management. Six sigma shows up on quite a few lists of “worst management fads” when those words are entered into a search engine, and for good reason. Far from being universally applicable, there are actually many scenarios where committing significantly greater resources into the quality a given service or product could be harmful to that organization’s health. Besides potentially contributing to scope creep (and thereby endangering an on-time, on-budget completion), an organization that succeeds because of the affordability and availability of its output may well be making a significant error in moving towards a high-quality product. In all likelihood, such an organization would be already known for providing affordable, available services or products, meaning that, even if they should move towards higher quality, the output would not necessarily be seen as such. Imagine your favorite fast-food chain suddenly offering rib eye steaks. See what I mean?

Next, I want to use our little two-tiered test with our friends, the Communications Managers. Even if one could argue that their engage-all-stakeholders strategy could help attain project objectives indirectly, by intercepting potential barriers erected by people who would not have otherwise opposed the scope had they only been informed of it earlier in its development cycle (a tentative take, at best), it also works against organizations that have the foresight to select whichever of the two of three targeted characteristics for their overall strategy. For example, if you were to launch an organization that could produce, say, a top-quality acoustic guitar for the same price as those companies producing low or mid-priced instruments, how anxious would you be to let people outside your organization (the very definition of “all stakeholders”) know exactly how you pulled it off? Such information is referred to as “trade secrets” for a reason, and releasing it to a broader set of people than is absolutely necessary in order to feel good about your enlightened approach to management is, in my opinion, a really bad idea.

For our last entry in the deducing-relevance sweepstakes, I’ll address one of my favorite targets, the risk managers (no initial caps). For this evaluation, we don’t even need the second-tier filter/test, since it flunks the first: I defy any risk manager to quantify how a risk register assists the PM in bringing in the scope, on-time, on-budget. Show me the PM who has spent real coin on the risk management function, and I’m confident that there’s going to be a customer who wants (or even requires) such an analysis. I have yet to meet the PM who, left to their own devices, would bother with the whole Monte Carlo schedule analysis schtick, or even consult it prior to making key decisions in the case of one being foisted upon them.

Meanwhile, Back At 221B Baker Street…

It’s my contention that, by using the two-tiered test of (1) asking the question “does this help bring the project in on-time, on-budget?” and (2) identifying how the proposed technique/strategy fits in the Affordable, Available, Quality: pick any two framework, many irrelevant business practices can be filtered out of the PMs’ repertoire, thereby enhancing the odds of identifying the killer optimizing the project’s technical approach. So, you see, it’s really rather elementary.