Millennium Challenge 2002 (MC02) was a major war game exercise conducted by the United States armed forces in mid-2002, likely the largest such exercise in history. The exercise … cost $250 million, involved both live exercises and computer simulations. MC02 was meant to be a test of future military “transformation”—a transition toward new technologies that enable network-centric warfare and provide more powerful weaponry and tactics.

Red, commanded by retired Marine Corps Lt. General Paul K. Van Riper, used old methods to evade Blue’s sophisticated electronic surveillance network. Van Riper used motorcycle messengers to transmit orders to front-line troops and World War II light signals to launch airplanes without radio communications. … In a preemptive strike, Red launched a massive salvo of cruise missiles that overwhelmed the Blue forces’ electronic sensors and destroyed sixteen warships. This included one aircraft carrier, ten cruisers and five of six amphibious ships. An equivalent success in a real conflict would have resulted in the deaths of over 20,000 service personnel. … Another significant portion of Blue’s navy was “sunk” by an armada of small Red boats, which carried out both conventional and suicide attacks that capitalized on Blue’s inability to detect them as well as expected.

At this point, the exercise was suspended, Blue’s ships were “re-floated”, and the rules of engagement were changed. … The war game was forced to follow a script drafted to ensure a Blue Force victory. … Red Force was ordered to turn on all his anti-aircraft radar in order for them to be destroyed, and Red Force was not allowed to shoot down any of the aircraft bringing Blue Force troops ashore. … They also ordered Red Force not to use certain weapons systems against Blue Force and even ordered that the location of Red Force units to be revealed. This led to accusations that the war game had turned from an honest, open free play test of America’s war-fighting capabilities into a rigidly controlled and scripted exercise intended to end in an overwhelming American victory. …

Due to his criticism regarding the scripted nature of the new exercise, Van Riper resigned his position in the midst of the war game. … Navy Capt. John Carman, Joint Forces Command spokesman, said the war game had properly validated all the major concepts which were tested by Blue Force, ignoring the artificially imposed restrictions placed on Van Riper’s Red Force which led them to succeed. (more)

War colleges, where people learn to be soldiers, often have war simulations where different people play different parts of a war between “us” and “them.” Students and others are told that these are realistic, or at least as realistic as is feasible given the simplifications that simulations and games require.

But I’ve now heard personally from enough independent expert insider sources that I’m willing to post it: the above example was not a rare exception; war games are mostly fake.

They are designed so that our side wins, and so that the official strategy that we teach students actually prevails. Every once in a while some joker plays “them” cleverly and wins, and then their career is over. The games also make sure terrifying outcomes never happen, even in a game where “we” win. For example, it is forbidden to sink an aircraft carrier.

I’m told that that fake war games and simulations are common in the rest of the military too. Simulations that show realistic levels of outcome uncertainty and variance tend to be rejected in favor of low variance ones that suggest outcomes just can’t get very bad.

You might have thought that because in the military most everyone’s freedom and lives are on the line, the military at least would try hard to create realistic estimates of the outcomes of their policies. But you’d be wrong. Organizational disfunction plagues them as well. Apparently military leaders think it is more important to instill confidence in the troops and citizens than to actually find out how wars would go.

Now ask yourself: since your freedom and lives aren’t on the line in your organization, just how much more dysfunctional might be your outcome estimating processes?