Game Theory in Management

Predicting the future of Project Management is far more difficult than doing so for Asset Management. Double-entry bookkeeping, the basis for today’s Generally Accepted Accounting Principles, has been around for hundreds of years, since around the time of Machiavelli (a coincidence?), and hasn’t changed very much over that time span. Essentially, the accountants helping the pilgrims settling their affairs prior to boarding the Mayflower wouldn’t be that out of place in a modern accounting firm, save for the computerized versus pen-and-ink ledgers, and (maybe) clothes worn. This being the case, it’s a relatively safe bet that GAAP will, in another five hundred years, be eerily similar to accounting practices today. Conversely, none of those early settlers would have any idea what you were talking about if you were to ask about their Work Breakdown Structure, or Cost Performance Index. They simply would not understand you.

But, if you think about it, why would it be otherwise? After all, the whole reason double-entry bookkeeping came into existence was primarily for purposes of taxation, with its attendant management insights fulfilling something of a secondary role. Taxation has been with us since the dawn of civilization, and can be reliably predicted to stay with us right up until governments cease to exist (and, in the case of the Death Tax, a little bit afterwards). Another aspect adding to GAAP’s longevity is the fact that virtually all of its precepts are widely-known (hence the “Generally” in “Generally Accepted Accounting Principles”), if not out-and-out codified in law. With its foundations so solidly established, Asset Management is practically immortal.

Let’s compare and contrast these characteristics with those of Project Management via a simple thought exercise. What would the characteristics be of one company with outstanding talent within its Accounting team, barely competent (or even inadequate) capability in its Project Management Office, versus another company with the opposite talent profile? I think it would look something like this:

All of which brings me to another issue somewhat germane to PM and its future. If you had developed an algorithm that significantly increased the odds that you would win a large lottery’s jackpot, would you be interested in sharing that formula with others, or would you keep it a secret so as to maximize your own payouts? The answer is obvious, which leads to a follow-on question: If you had an insight into PM techniques, one that gave an advantage in achieving the benefits attributed to Company B in the above table, how eager would you be to share that outside of your own organization? Such information is considered, by definition, company sensitive, in that it is part of the structure that grants a competitive advantage. On the other hand, when has any organization throughout history gained a competitive advantage due to the discovery of a superior accounting practice? Before you answer with the Hostile Takeover phenomena, I would assert that such tactics do not stem from a novel evaluation of a given company’s assets to test if they are “undervalued,” as much as it’s a strategy to eliminate a competitor from a specific market.

Which brings us back to being able to accurately predict the future of Asset Management versus Project Management. New and reliably profitable techniques don’t get presented in the Asset Management realm because, with the exceptions involving avoiding taxes and a few others, they don’t exist, and they don’t get presented in the PM world because, if they really work, there’s little incentive for the authors of those papers to give away the metaphorical winning lottery numbers. In summary, the future of accounting is easy: I think it’s safe to say that it’s going to be fairly similar to the way it is now. The future of PM? Nobody really knows, because the ones who hold that future in their hands aren’t talking (outside of ProjectManagement.com’s blog pages, of course).

In the 1988 action film Die Hard, Bruce Willis plays the role of New York City detective John McClaine, who becomes caught up in an attempted terrorist takeover of the offices of the Nakatomi Corporation, where his estranged wife works. As one could imagine from films of this genre and era, there’s a fair amount of violence before (and after) the Los Angeles Police Department is alerted that something is amiss, but, even after an initial investigation, they remain unaware. This is only remedied when McClain drops the body of a slain terrorist many stories onto the hood of the police car driven by the investigating officer. Having done so, McClain looks out the shot-out glass pane that the terrorist just exited, and spits out the line “Welcome to the party, pal!” The still shot of Bruce Willis uttering that line has become the basis for a popular internet meme, typically meant to communicate a sense of disappointment (or even disgust) that someone assigned to a specific job should remain oblivious to critical and obvious developments well past the time that they should have at least started doing something about them.

Meanwhile, Back In The Project Management World…

I’ve written often about PM techniques’ unequalled capability to accurately predict such vital management information as how long a particular activity, task, or even project will last, or how much it will cost at completion. The use of action item lists, updated via responsible managers’ input along the lines of whether the activity in question is on-time or late, is a common way of trying to get a handle on when things will be completed, but it’s completely unreliable. Similarly, using the general ledger to provide a so-called “burn rate,” or amount the project is spending by reporting period (usually monthly), and then multiplying that figure by the number of reporting cycles (hypothetically) left in the project to derive an Estimate at Completion (EAC) only produces wildly misleading results, even when these figures are averaged out, or massaged via other Gaussian Curve methods.

No, the only valid way of deriving an accurate figure for activity/task/project duration or total costs belongs to PM, specifically Critical Path Methodology (CPM) for duration and Earned Value (EV) for both duration and cost. The simplest version of the formula is to divide the activity’s/task’s/project’s percent complete into its cumulative actual costs for a reliably accurate Estimate (cost) at Completion. This formula also works for duration: divide the percent complete figure into the amount of time that has elapsed since the work began, and you’ll have a usable end-date.

But here’s the thing: this simple and effective technique has been around for decades, if not longer. “If that’s the case, Michael” I can hear GTIM Nation say, “then why aren’t those techniques much more widely used, if not commonly-recognized as the standard?” I believe the answer has to do with factors that are outside nominal management science boundaries, and firmly in the realm of Organizational Behavior and Performance. Perhaps the most obvious example is provided by our friends, the Accountants, who typically hold that any piece of management information that is quantified in monetary units must originate from the general ledger, and Generally Accepted Accounting Principles, or GAAP. Think I’m exaggerating? Consider the use of the term “the bottom line.” Commonly used as shorthand for “I believe that X is the ultimate impact of the conditions/issues we have been discussing,” my guess is that it’s originally derived by how people read and understand a profit-and-loss statement, where the summary answer to how profitable a given organization is shown on the literal bottom line of that report. Since the standard Earned Value reports Cost/Schedule Status Report (C/SSR) and Cost Performance Report (CPR, Format I) have the aforementioned calculated Estimate at Completion documented in the far right-hand column, I think we PM-types should stop using the term “bottom line,” and substitute “right-hand column” instead. Also, no real project (as distinct from an organization, or function) spends money at a consistent level month-to-month, and pretending to the contrary doesn’t make it right.

Then we have our other friends, the risk managers (no initial caps). They have managed to convince themselves (and, sadly, many others) that the use of Gaussian curves will yield usable management information on what things will go wrong on a given project, and what the cost and/or schedule impact will be. Of course, the Earned Value and Critical Path systems will accurately inform on cost and schedule performance, whether they are influenced by things that management anticipated or not. And, again, this type of analysis has been around for decades. Somehow, the risk managers (no initial caps) have influenced the selection of analysis techniques away from the optimal and towards their, ummm, sub-optimal versions. It’s maddening.

All of which leaves us PM-types in the unenviable position of having to witness “managers” choosing the wrong analysis techniques, hawked by Accountants and risk managers, while remaining oblivious to the optimal PM ones. And, when these managers finally use these PM techniques, we can only hope to imitate Bruce Willis’ steady and mildly ironic tone when we say “Welcome to the party, pal!”

In last week’s blog I asserted that the principals undergirding modern Project Management started off in the realm of the personal and migrated to the business arena, and the blog prior to that one argued that PM techniques have that rare quality in the management sciences field of being completely scalable. Taken together, it stands to reason that many of the principals contained in the PMBOK Guide® would have readily-recognizable counterparts in our common, personal vernacular. I think that it also implies that management science hypotheses that have been shoe-horned into the PM codex, but don’t really belong there, would tend to stand out because of their lack of any real-world, or personal utility. Based on the old-style structure of the PMBOK Guide®, let’s take a look at some of those real-world applications, shall we?

• Scope: Keep your eye on the target.
• Cost: This is a story from Walt Garrison’s book, Once A Cowboy (Random House, 1988), discussing an incident where Don Meredith spent $400 (real money back in the 1960s) buying beer for friends and teammates at a bar. “I just made $400” Garrison quotes Meredith. “How do you figure that?” “I just had $800 worth of fun.”[i]
• Schedule (neutral): Time is money.
• Schedule (negative): Well, that meeting was time I’ll never get back.
• risk (no initial caps): “Sir! The odds of surviving a direct attack on an Imperial Star Destroyer are—”[ii]
• risk (no initial caps) (appropriate response): “Shut up!”[iii]
• Procurement: Caveat emptor.
• Communications: “Keep a secret, it’s your slave. Tell it, and it’s your master.” – Henry Wilson Allen.[iv]
• Quality: “Quality is never an accident. It is always the result of intelligent effort.” – John Ruskin[v].
• Human Resources: “Never go to a doctor whose office plants have died.” – Erma Bombeck[vi]

Now, the alert GTIM Nation citizen will quickly notice that the personal-level PM applications for risk (no initial caps) and Communications tend to run contrary to what much of the higher-level analysis posits, namely that risk management (no initial caps) provides a much-needed information stream, central to real-time decision-making, and that Communications Management is needed to ensure that all stakeholders are identified and fully-informed of your project’s goings-on. In the risk management (no initial caps) instance, the quote from The Empire Strikes Back is actually the second time in that movie where the android C3PO is attempting to advise Han Solo on his course of action. The first time went like this:

C3PO: “Sir, the possibility of successfully navigating an asteroid field is approximately 3,720 to one.”

Han Solo: “Never tell me the odds.”[vii]

By this point in the Star Wars films, the C3PO character has been established as being very intelligent, of occasional (if accidental) help in some situations, utterly useless in crises, lacking a practical sense of perspective or proportion, and possessing a mechanical personality that’s almost human – the perfect combination of traits for a character who gives marginally relevant statistical information at various times to the real decision-makers.

Next, we have the utter failure of the Communications Managers to produce a variant of their recommendations that has traction in the personal sphere. In fact, it’s quite the opposite: the notion that all “stakeholders” should be identified and informed of your personal plans and activities sounds absurd on its face, and there’s a good reason for that. Not everyone who is “impacted” by your plans and activities wants those efforts to come to fruition, and some of these can be expected to actively work against them. To alert this entire group of people is foolish on the personal and macro level. Consider also how often the business world is compared to team sports, or even warfare, with similar axioms and strategies having efficacy in both. Enemy nations employ spies to discover the other side’s plans and activities, and spend considerable resources to prevent their own from being distributed beyond those who can demonstrate a need-to-know. While the utter destruction of your competitors’ Project Teams isn’t a normal aspect of management in general and PM in particular, out-performing them in terms of Scope, Cost, and Schedule is a universal goal, and broad dissemination of your plans and activities is a sure-fire way of lowering the odds of attaining it.

For those who insist on heeding the advice from risk and Communications Managers, I’ll leave you with one more Hans Solo quote, said after the Millennium Falcon has just experienced a large jolt, and C3PO has offered that the asteroid on which they’ve landed “is not entirely stable.”

“’Not entirely stable.’ I’m glad you’re here to tell us these things. Chewie, take the professor into the back, and plug him into the hyperdrive!”[viii]

[i] Garrison, Walt, Once A Cowboy, Random House, 1988, pp.77.

[ii] C3PO, The Empire Strikes Back, 1980.

[iii] Princess Leia, The Empire Strikes Back, 1980.

[iv] Retrieved from http://thinkexist.com/quotation/keep-a-secret-it-s-your-slave-tell-it-and-it-s/395805.html on September 20, 2022, 20:28 MDT.

[v] Retrieved from https://www.azquotes.com/quotes/topics/quality.html on September 19, 2022, 18:18 MDT.

[vi] Retrieved from https://www.azquotes.com/author/1616-Erma_Bombeck on September 19, 2022, 18:26 MDT.

[vii] The Empire Strikes Back, 1980.

[viii] Ibid.

With ProjectManagement.com’s September theme of Personal PM, I (naturally) found my thoughts wandering towards Impersonal PM. What is Impersonal PM? A couple of examples from the movies include:

• “This is business, not personal!” The Godfather (1972), though some version of this quote occurs throughout all three movies in the series.
• “It was all business, it was nothing personal.” Taken (2008)

Such an assertion often comes about when some particularly loathsome antagonist is about to receive a rather violent comeuppance, and they are attempting to defend their otherwise indefensible actions, though in The Godfather it’s usually invoked in an attempt to avoid a reflexive (albeit extremely) fierce response to some provocation.

I think that, if I were to attempt a paraphrase of “It’s not personal, it’s business,” it would be “Taking things personally will compromise one’s ability to select an optimal strategy in the impersonal world of business.” Consider that PM is part of the Management Science realm. The scientific method has no room for the personal. This is not to say that an acknowledgement of Organizational Behavior and Performance factors is not central to developing things like a workable implementation strategy – they absolutely are essential to success. What I am saying is that developing the optimal technical approach to any given problem is going to be difficult enough even before the human factors are brought in to the fray, and forming initial responses from the personal aspect will almost always result in poor choices.

Of course, impersonally-derived decisions are often the basis for some evil choices in the movies. Such was the case for the HAL 9000 computer from 2001: A Space Odyssey (1968). HAL controls the interplanetary ship Discovery One, including the life support systems of three crew members who have been placed in suspended animation. Two crew members who have not been placed in suspended animation, Dave Bowman and Frank Poole, discover HAL acting erratically, something supposedly impossible for its series of computers. They decide to disconnect HAL, but HAL murders the crew members in suspended animation, kills Frank, and attempts to kill Dave. Why does HAL do this? It’s not entirely clear, but some clues are offered up by HAL itself, both in 2001: A Space Odyssey, and in the follow-on movie, 2010: The Year We Make Contact. In the former, HAL claims that it was constructed for “the accurate processing of information, without distortion or concealment,” while in the latter it is revealed that HAL knew of the Discovery’s real mission, but was instructed to not disclose it to either Frank or Dave. But even tossing in the remaining key decisional element, that the successful completion of the mission was considered a higher priority than the lives of the crew, doesn’t quite get me to understanding why HAL concluded that murdering the crew outright was a swell idea. Oh, well, willing suspension of disbelief in order to accommodate a plot device, I suppose.

Meanwhile, Back In The Project Management World…

One of the elements of (valid) PM Information Systems that I discussed in last week’s blog was their scalability. Another aspect that makes them rare and highly valued in the business world is their impersonal nature. “But Michael!” I can hear GTIM Nation challenge, “it’s not rare! The general ledger is just as impersonal as an Earned Value Management System!” Well, yes and no. A particular department’s budget for the fiscal year might be high or low with respect to its actual value to the organization, with managers in possession of exceptional political skills typically being in receipt of the former. This effect is, by definition, a personal influence on those cold, hard general ledger figures. But a cost variance from an EVMS is merciless. You have either performed as planned, better than planned, or worse, and it’s captured, quantified, and reported for all the “stakeholders” to see. Sometimes people in positions of authority will attempt to soften or ward off information that points to poor performance for which they are responsible. It’s the reason that savvy PMO Directors will often instruct their Project Controls Specialists to not accept percent complete estimates over 90%, or even 85%, unless the task/activity is truly finished, and will insist on calculated Estimates at Completion as opposed to the easily manipulated “bottoms-up” variety. Valid cost and schedule performance management systems are no respecter of persons, making them far more reliable than their subjective data-containing counterparts in the Management Information System portfolio.

Unfortunately, some executives will take Project Management Information System output personally. The same people who wouldn’t dream of getting rid of the company’s payroll system if salaries unexpectedly spike would be perfectly fine with eliminating an Earned Value or Critical Path requirement if it points to their lack of performance. It’s when these people push the idea of scaling back the PMO’s purview that the more rational decision makers should quote HAL, and reply “I’m sorry, Dave (or whomever), I’m afraid I can’t do that.”

It’s a pretty amazing coincidence that the September theme for ProjectManagement.com is Personal PM, seeing as my son is getting married later this month. Given that the typical wedding service lasts between twenty minutes and one hour, such occasions may be, on a minute-by-minute basis, the most thoroughly planned and scheduled events in the universe. Smallbusiness.chron.com asserts that Critical Path Method (CPM) scheduling began in ancient times, and cites the pyramids (2589 BC) as evidence,[i] but marriages may pre-date the pyramids by thousands of years, making them, in my opinion, the most likely driver for the invention/discovery of CPM. This provides quite the perspective flip, no? In the 21^st Century, the preeminent organization for advancing Project Management, PMI®, assigns as a monthly theme for its writers the notion of “personal” PM, when it may very well be that PM originated and was personal all along, and only moved to industry and management science later. But this perspective points to something that I think we’ve known all along, but only rarely articulated: Project Management has that scarce and unique attribute of being completely scalable.

To be sure, not all of the management concepts that seek to belong to the PM club are scalable. Consider risk management (no initial caps). Those techniques might aid in making decisions in a gambling establishment, but most table games in, say, Las Vegas have very specific, inviolable rules, and are conducted in a highly controlled environment. Given the number of cards in a deck, or number of faces on dice, or number of slots on a roulette wheel, Gaussian Curve-based analysis can provide for some better-informed decisions. Business, in general, and Project Management in particular have very few very specific rules, and never takes place in a highly controlled environment. Dealing with chaos is part and parcel of the PM function, and chaotic events are, definitionally, impossible to quantify with respect to their impact(s) or odds of occurrence. So, risk management (no initial caps) utterly fails the scalability test.

What of this blog’s titular analysis technique, Game Theory? Well, this is more of a mixed bag, and mostly depends on the game being analyzed, along with its strength of analogy to the management problem being addressed. John Nash won a Nobel Prize in economics for his contributions to Game Theory and management sciences, but Game Theory itself also suffers from a limited capacity for scalability, with perhaps the best example being the Ultimatum Game. This game entails approaching two people (Person A and Person B) and making them the following offer: the Experimenter will give them $100 (USD) if Person A can produce a scheme for how the money will be split between them, and this plan is accepted by Person B as proposed the first time. If Person B rejects Person A’s plan to split the money, neither participant receives anything. Conventional Game Theory wisdom held that Person A would maximize their payoff by proposing that Person A receives $99, while Person B receives only $1, on the notion that Person B would essentially be presented with a choice to receive either $1 or nothing at all, and could be reliably projected to choose the former. In actuality, this proposed split was almost never accepted, usually because Person B didn’t believe that such a proposal was “fair.” In certain cultures, even a 50/50 split was routinely rejected, for what the researchers called “cultural” reasons. But another way of saying “the experiment didn’t work out as our theories predicted due to cultural reasons” would be “our analysis failed to take into account a sufficient number of parameters – much less accurately quantifying them – to generate a reliable predictor of the subjects’ decisions.” In short, the difficulty in recognizing and quantifying the necessary number of parameters to reliably invoke Game Theory in business applications serves as an upper limit on its scalability.

Compare this limited scalability to Earned Value and Critical Path Methodologies (EVM and CPM, respectively).  To get a pretty accurate idea of how much a certain effort, large or small, will cost at its conclusion, simply assess how much has been spent by the time the 50% complete point has been reached, and double it. By extension, dividing the percent complete figure into cumulative actual costs generates a fairly reliable Estimate at Completion. Interestingly, this formula also works for schedules (though it’s not, technically, CPM). To get a usable estimate of how long a specific piece of Scope will take to complete, again, large or small, divide the percent complete figure into its cumulative duration. As for CPM-specific applications, all successful projects had to be directed by a person or persons who knew which activities needed to be completed before others could start, well before anybody thought to coin the phrase “schedule logic.”

To sum up, the essential analysis techniques of PM, Earned Value and Critical Path, are completely scalable, from modern mega-projects, to ancient mega-projects, all the way down to arranging a wedding. Indeed, I think we should petition to include a chapter on bouquet-throwing in the next edition of the PMBOK Guide®!

[i] Retrieved from https://smallbusiness.chron.com/history-critical-path-method-55917.html on September 3, 2022, 19:45 MDT.