The year is 2026, and China has just launched an invasion of Taiwan.
What happens next was the subject of a comprehensive non-classified War Game simulation convened by the Center for Strategic and International Studies. Joining me on the podcast this week is the person who lead that War Game, Retired Marine Colonel and CSIS senior advisor Mark Cancian.
I’ll cut to the chase: After 24 iterations, the most probable outcome was the defeat of a Chinese invasion of Taiwan, but at a staggeringly high cost to Taiwan and the United States.
We discuss the probable outcomes and scenarios in detail in this episode of the podcast.
To access this episode on your favorite podcast player, go here.
To get episode transcripts and receive occasional updates, sign up for our mailing list
Transcript lightly edited for clarity
Mark Leon Goldberg [00:03:43] Before we get into the substance of the findings of the war games, can you briefly explain how it worked and was designed so that later in the conversation listeners have a better sense of upon what you’re basing your assumptions?
Mark Cancian [00:03:55] We developed a war game of a U.S. China conflict over Taiwan. It’s a physical war game in that there are several maps. There are two operational maps where air and naval forces maneuver. And then there’s a ground map of Taiwan where the ground forces maneuver. And there are counters that represent the expected force structure of the various countries in the year 2026.
Mark Leon Goldberg [00:04:25] Why is that? Why 2026?
Mark Cancian [00:04:28] We picked 2026 for two reasons. The first is that this is the time that many officials have highlighted as a period of great danger. Admiral Davidson was one of the first to highlight this time period. A lot of people call this the Davidson window, this sort of 25, 26, 27, 28 period. The other reason was that at that time we developed the war game, it was the end of the Pentagon’s planning period. So they published information about forces and plans. So we had pretty good sense about what the United States was going to have in 2026.
Mark Leon Goldberg [00:05:05] So in creating the wargame, you have various experts basically play different sides of the conflict and you have rules of probability and you know the force structures of each side or have assumptions about the force structures of each side. Is that right?
Mark Cancian [00:05:22] That’s it. Exactly. And it’s important to talk about, well, maybe each one of those pieces first on the force structure, as we looked at all the different countries, the United States and China, of course, but also Japan and Taiwan, to set what we believed their structures were going to be in terms of combat resolution. We did a lot of operational analysis, looking at history, looking at weapons testing data to come up with probabilities about the results of various kinds of combat interactions. So that given if you want to fire a missile at a Chinese ship you know whether the chances that the missile will actually launch, whether the chances it will actually track with the chances to get shot down, where the chances are it will hit the ship, and then if it hits a ship with chances, it’ll sink the ship. So we did research on all of that. Those results are captured in some computer programs, Excel programs where the calculations are complicated and also in some tables. We’ve also rolled die. But it was important to us that the combat results be first objective, that it’s not reliant on a bunch of people in the back room using their judgment and that it be transparent so that someone disagreed they could see where we disagreed and we could have a conversation about that.
Mark Leon Goldberg [00:06:42] And using all this data as inputs. You ran the war game 24 times. Why that amount?
Mark Cancian [00:06:53] We wanted to get a large number of iterations so that we could explore a wide variety of scenarios and so that the results would be more analytically credible. There are many war games that are run once or twice, and those can be very educational for the participants. But as an analytic foundation for making policy recommendations, it’s very narrow because it’s highly dependent on the particular scenario, the particular participants, you know, the luck of the dice. We wanted to run it many times and have something that was analytically more defensible.
Mark Leon Goldberg [00:07:34] And because you ran it many times, you saw several outcomes repeated. And I want to spend the bulk of this conversation talking with you through how a war might play out. So first, what did your war game assume was how China would launch its invasion?
Mark Cancian [00:07:59] Our assumption was that China has made a decision to launch an invasion, and we recognize that that might not be the most probable course of action for China. But we argue that it is the most dangerous. And because of the Chinese military buildup and the rhetoric that comes out of the Chinese senior officials, it’s a plausible course of action. So the game begins. The Chinese have decided to launch an invasion and they use their missiles and aircraft to strike the U.S. and Taiwan. And then the U.S. is in the conflict and the two sides take turns moving their forces and launching attacks.
Mark Leon Goldberg [00:08:41] Where does China strike Taiwan and the U.S. in that initial volley?
Mark Cancian [00:08:47] In the initial volley, they strike all of the Taiwanese targets, but particularly the Air Force and the Navy, the Taiwanese surface ships and most of their air force are destroyed in the first couple of days. The Chinese just have so many missiles available for the United States, they typically will go after the carriers. They’ll often go after Guam, if there are a lot of forces on Guam and surface ships. And then the U.S. bases in Japan. One of our assumptions is about Japan. And our assumption is that the Japanese would allow the United States to use its military bases in Japan, but Japan would not participate itself unless the Chinese attacked their homeland. So bases like Kadena [on Okinawa] the United States is using for operations, and the Chinese will often attack that.
Mark Leon Goldberg [00:09:44] So just to get back to this, the Chinese invasion of Taiwan would most probably start with China launching attacks directly against American military assets in the region.
Mark Cancian [00:09:59] That was our assumption, yes. The Chinese are in a situation like the Japanese were in 1941, that if they’re going to launch this attack in a situation where the United States has said that it will probably become involved, that they might as well strike the U.S. first and gain the advantages that go with that.
Mark Leon Goldberg [00:10:18] And I do want to get into the American response. But first, one of your key assumptions is that Taiwan would put up a fight, right?
Mark Cancian [00:10:28] Absolutely. The whole effort is predicated on the assumption that Taiwan will fight. And of course, there’s a lot of discussion about that. Taiwanese government insists that it would fight. Many Taiwanese believe that. But there are others who raise questions about whether the Chinese could undermine Taiwanese morale and resolve. In the end, we’ve got the example of Ukraine in front of us. We didn’t really make a judgment about whether Taiwan the fact we made that an assumption.
Mark Leon Goldberg [00:11:01] Because if Taiwan does not put up a fight, everything else doesn’t really matter.
Mark Cancian [00:11:05] That’s right. I mean, the whole simulation is irrelevant because the Chinese will walk in.
Mark Leon Goldberg [00:11:11] So the conflict begins with China launching airstrikes on Taiwan and on U.S. military targets, presumably also soon thereafter, China mounts a ground invasion of Taiwan.
Mark Cancian [00:11:26] Absolutely. In fact, right at the very beginning, they launched that invasion. They typically landed in the south of Taiwan. The reason is the Taiwanese have the bulk of their combat power up north protecting the capital. Not surprisingly, in a couple of iterations, the Chinese player tried to land in the north. That was very difficult. So in almost all of them, they landed in the south where there was less Taiwanese combat power. The problem for the Chinese then, is that it’s easier to get ashore, but now they have to fight their way up the entire island. And Taiwan is a very easy island to defend. The central area is very mountainous. The coastal plains have many cities and rivers. So there are many places where the Taiwanese can defend its campaign that ends up looking like Italy in World War two.
Mark Leon Goldberg [00:12:17] So China launches its airstrikes, its amphibious assault. And what are the initial American responses to having both just been attacked and to the ground assault on Taiwan?
Mark Cancian [00:12:36] The U.S. response is to attack the Chinese fleet using bombers with long range missiles, using tactical aviation with long range missiles. There’s nothing really we can do about the ground campaign because it’s so hard to break through the Chinese defensive barrier. What the United States focuses on is breaking down that barrier so that eventually it can intervene on the ground.
Mark Leon Goldberg [00:13:01] Can you just kind of flesh out that initial U.S. response a little more? So, like the U.S. basically launches strikes against naval fleets that are seeking to protect its invasion of Taiwan.
Mark Cancian [00:13:17] Exactly. The Chinese have a large surface navy. Some of the Chinese surface navy are protecting the amphibious ships. Very often, the rest of them form a picket line out east of Taiwan to intercept U.S. ships and missiles. The U.S. has to try to break through that to get at the amphibious ships. And the amphibious ships are the center of gravity for the Chinese. In other words, if they can sustain those amphibious ships, they can build up their forces on Taiwan and sustain them if they lose their amphibious ships in the matter. What else it has? The invasion will fail. So the action revolves around the Chinese amphibious ships.
Mark Leon Goldberg [00:13:56] And America’s ability to kind of destroy or deter or prevent those inferior ships from landing one way or another?
Mark Cancian [00:14:02] Exactly. And the Taiwanese, too, I will add, because the Taiwanese air and naval forces get pretty beat up early on, but they do have some ground based anti-ship missiles and those can be very effective against the Chinese ships.
Mark Leon Goldberg [00:14:18] You referenced this earlier. But I want to flesh out the vital role that Japan plays in helping to determine the outcome of this war game. So the U.S. obviously has a deep an historic security relationship with Japan and has a number of bases on Japan. But one of the key criteria of success you determine is Japan permitting America to use those bases to defend Taiwan. Why is that?
Mark Cancian [00:14:49] Yes, the bases are absolutely critical because that’s the only way the United States can use its fighter and attack aircraft in the battle. If the United States can’t use those bases, then these aircraft have to be stationed at Guam or even further away, and that’s too far for them to intervene. The bottom line is, if we can’t use the bases in Japan, then all of those F-35s and F-22s and F-15s and F-16s are really pretty useless. And the United States has to fight the war entirely with bombers.
Mark Leon Goldberg [00:15:23] And in most iterations of the war game, Japan indeed does permit the use of their airbases and in some iterations as well. Japan, after having been attacked, permits the use of Japanese armed forces in the conflict directly, correct?
Mark Cancian [00:15:39] That’s right. This base assumption, as we call it, you know, what we think is the most probable assumption is that the Japanese allow the United States to use its bases. And we tested that with a wide variety of very senior Japanese leaders, and they were comfortable with it. Their argument was that if Japan did not allow the U.S. to use its bases, then it would be essentially tearing up its 70 year security arrangements with the United States, and they were not willing to do that. On the other hand, they would be very reluctant to get involved off the bat themselves. So we’re quite comfortable with that assumption. We did run, I think, one or two excursions where the Japanese were strictly neutral and it was really impossible for the United States to be effective because the United States could only rely on bombers being based far away.
Mark Leon Goldberg [00:16:31] So as you mentioned, the invasion begins with an amphibious assault and airstrikes against U.S. and Taiwanese forces. The U.S. responds with its own attacks on that amphibious capability of China to prevent them from mounting their ground invasion and sustaining a ground invasion. How long does this play out?
Mark Cancian [00:16:54] The game runs about three or four weeks of game time, and by that time, it’s pretty.
Mark Leon Goldberg [00:17:01] Clear.
Mark Cancian [00:17:03] Usually how the battle is going. But it would take longer for to get to an ultimate resolution. In other words, it’s pretty clear that either the Chinese have enough forces on Taiwan and have captured maybe an airfield or a port so they can sustain those forces. And with enough time, they’ll go build up their forces and conquer the island. But that might take a couple of months. Conversely, very often the Chinese were not able to build up enough combat power. The U.S. was able to attract the Chinese amphibious ships so that the Chinese couldn’t get more forces on the island, and it would take maybe another month or so for the Taiwanese to eliminate the Chinese that were on the island. But eventually it was just units would be destroyed. So we ran 3 to 4 weeks of game time. But the full campaign, as we designed, would probably run 2 to 3 months.
Mark Leon Goldberg [00:17:58] So after running. 24 iterations of this war game. What did you find to be the most probable outcome of a Chinese decision to invade Taiwan?
Mark Cancian [00:18:14] The overall outcome was that the United States and its partners could maintain an autonomous and democratic Taiwan, but that would come at very high cost to the United States, to Japan, and to the Chinese. And as a result, our recommendation is to enhance our warfighting capabilities so that this conflict would be deterred or if not deterred, would be concluded more quickly. Having said that, we looked at a variety of scenarios so that optimistic scenarios, of course, come up better for the United States and Taiwan. The variety of pessimistic scenarios that took longer had much higher casualties, and there were a handful of very pessimistic scenarios where the Chinese actually prevailed.
Mark Leon Goldberg [00:18:59] But in the most probable scenario, which you described as something of a Pyrrhic victory, what does Taiwan, the mainland of Taiwan, look like after that probable Pyrrhic victory?
Mark Cancian [00:19:14] Yeah, Taiwan is badly beaten up. Its economy is probably in a shambles. Many of its cities have been badly damaged because of the fighting. Its transportation system has been badly damaged because of the Chinese interdiction efforts so that it has suffered greatly.
Mark Leon Goldberg [00:19:35] And on the U.S. side, what do losses look like after having been engaged in battle with the Chinese military for at least three weeks?
Mark Cancian [00:19:46] So U.S. losses are quite substantial. And I want to emphasize that the losses that we put in our report are really a floor because they don’t include losses from other areas, other regions. But the United States will lose 10 to 20 ships, including two aircraft carriers, a couple of hundred aircraft and several thousand killed. One of the things we note is that in three weeks, the United States will typically lose about half as many service members as it lost in 20 years of conflict in the Middle East.
Mark Leon Goldberg [00:20:20] I mean, that’s like a staggering number. Both of U.S. equipment lost and U.S. American service members killed. Something almost like 3000 American service members would die within like three weeks of combat, which is a rate of death in combat not seen in a generation in America.
Mark Cancian [00:20:38] Yeah, In effect, it’s a rate that hasn’t been seen really since the Second World War or particularly for the equipment. Yes, this will be a shock to the U.S. population and to the U.S. military, particularly to the Navy, the Air Force that who have essentially operated in sanctuary since the Second World War. And that’s why we worried that this might be a Pyrrhic victory and could conceivably be followed by a return to isolationism, as we saw after the First World War, and therefore, why we need to deter this war rather than fight it.
Mark Leon Goldberg [00:21:16] And just also, in terms of the U.S. fleet that’s lost. I mean, that’s a big chunk of the U.S. Pacific fleet, is it not?
Mark Cancian [00:21:24] It’s a good chunk. Yes. We have to keep in mind that, you know, the U.S. fleet runs about 290 ships. You know, we lose ten or 20. We do lose to aircraft carriers, however. And, you know, that’s a large loss. And as I say, you know, this is a flaw because there’s also fighting going on in the South China Sea. And we stop at three or four weeks. That would go on for a couple more weeks. Also, the problem for the U.S. fleet is that these ships are very hard to replace. It would take probably a decade to replace the surface combatants. The destroyers and the cruisers and the aircraft carriers could probably never really be replaced, given current capacity for building aircraft carriers.
Mark Leon Goldberg [00:22:09] And what does your scenario suggest about what happens to China after suffering a defeat?
Mark Cancian [00:22:16] China would lose a lot of equipment also, including in the unsuccessful invasions, many prisoners of war on Taiwan. These losses would likely be large enough that they might destabilize the Chinese communist powers regime. And one of our hopes is that the Chinese will recognize this and it will act to make them more cautious. One piece of good news is that we know that the Chinese government and the Chinese media have seen this report and we hope, you know, have taken it to heart.
Mark Leon Goldberg [00:22:51] And one thing that I find interesting is that one of your key assumptions, correct me if I’m wrong, is that the U.S. would not strike the Chinese mainland. Correct in the event of a Chinese attack on Taiwan. Why is that?
Mark Cancian [00:23:06] We actually played it both ways. About half the games we played where the U.S. was allowed to strike the mainland in about half the games. The United States was not. The reason we played it both ways is that a lot depends on who you talk to, whether this would be authorized. If you talk to people in the Pentagon, they say, absolutely, we’re going after the mainland. If you talk to people, you know, come out of the State Department or the White House, they say, well, you know, maybe not. That might be too escalatory. The Chinese are, of course, a nuclear power, and striking their homeland might result in escalation to nuclear weapons. So we played it both ways.
Mark Leon Goldberg [00:23:42] In any of your models, was that nuclear scenario envisioned and tested? I mean, I remember reading a article in the academic journal International Security, maybe like a year ago that suggested the most probable path of a nuclear exchange between the U.S. and China would occur in the context of a conventional war in which the United States struck targets in China that the Chinese perceived to be attacks against their nuclear capabilities and thus would launch. Did a nuclear exchange result in any of your outcomes?
Mark Cancian [00:24:20] The choice was no, because we took nuclear off the table for two reasons. I mean, one is that some people argue that a conflict could be kept at the conventional level because of concerns of escalation. But the other one was that including nuclear play first makes it a very different game. A very often takes much longer. And of course, the dynamics are different on the battlefield if you’re using nuclear weapons. We are hoping to do a follow on project that includes nuclear weapons and nuclear decision making.
Mark Leon Goldberg [00:24:51] So if your most probable outcome was this Pyrrhic victory in which the U.S. geostrategic position was severely damaged and that the naval fleet was damaged, all these service members died. What recommendations did your war game suggest to prevent that Pyrrhic victory and prevent a Chinese invasion of Taiwan?
Mark Cancian [00:25:20] We had a wide variety of recommendations. I mean, to hit some of the big ones, we recommend that the United States and Japan build hardened shelters for aircraft. And the reason is that 90% of aircraft losses occur on the ground to Chinese missiles. I was astonished to find out that the U.S. base on Guam, Andersen Air Force Base, has no hardened shelters. So that makes aircraft very vulnerable to Chinese missile attack. Building hardened shelters and dispersing aircraft would make them much more survivable. Another major insight was the need to buy more long range precision munitions, particularly anti-ship munitions. The United States runs out of these after the literally the first couple of days of fighting and then and then has other munitions it can use. But these require the launching platform, the aircraft, to get close to the Chinese. Therefore, you have much higher losses. And that’s something that the department has picked up on. We briefed many people in the Pentagon. The Pentagon’s FY 2024 budget includes a lot more of those munitions. We like to think that our report had something to do with that, combined with the experience in Ukraine, where, again, we’re running out of certain kinds of munitions. Beyond that, we had a wide variety of insights for the United States. And I’ll give you just two more. One is that the United States needs to be very careful about forward deploying forces in a crisis. Our policy and our history is that when there’s a crisis brewing, we send forces forward. We send the carriers forward to enhance deterrence and to increase war fighting capability. The comes to that. The problem is with China, you have now put these assets inside their defensive bubble where they’re very vulnerable. That’s why we lose two carriers right off the bat. People forget that in the spring of 1941, the United States Navy moved its battle fleet from San Diego to Pearl Harbor to deter the Japanese from adventurism. And we all know how that turned out. Tom Schelling had a great quote where he said, the problem with a strong deterrent is it also makes a great target. And another important insight is that there’s no Ukraine model for Taiwan. With Ukraine. Of course, once the war began, the United States ramped up its transportation of weapons and equipment, as did many of our allies. The Russians have tried to interdict that flow, but they’ve been unable to do that. So there has been this huge flow of weapons and equipment to Ukraine all the way through the war. That’s not possible with Taiwan. The Chinese defensive bubble over Taiwan is so powerful that nothing can get in there for at least the first month or two. Several of our U.S. players tried to do that. They sent some convoys in amphibious ships, U.S. amphibious ships. They were always sunk. One participant tried to fly in an Army brigade. It was shot down. So Taiwan is going to have to fight the war with everything it needs beforehand and not count on getting resupplied.
Mark Leon Goldberg [00:28:39] So the project did not take a position on whether the benefits of defending Taiwan outweigh the costs. But the point was to put information out there. Let policymakers decide. But after having run the war game, do you personally take a position on whether the benefits outweigh the costs in that most likely Pyrrhic victory scenario?
Mark Cancian [00:29:01] I do. It’s my personal opinion that the United States should beef up these defenses. We have increasingly put out a policy that we would defend Taiwan. You see that coming out of Congress. You see it coming out of the White House. The president has said that several times, despite the walk backs by his advisers. So if we are going to have an announced policy of defending Taiwan and I think that’s reasonable, then we need to back that up with a military posture that’s actually capable of doing that. So personally, yes, I would execute these enhancements and defend Taiwan. But in the report, the reason we didn’t as a group publish that is that people have different opinions and that judgment relies on a wide variety of foreign policy judgments. And I should say at this point that this project was run by three people. There were three principal investigators. I was the thesis lead and sort of the manager, Eric Higginbotham, who’s up at MIT, who’s with ran for many years, did the China Asian part and also a lot of the wargaming and Matthew can’t Ian, who’s at the Naval War College running war games he did a lot of the war game design. And if our last names seem to be similar, it’s because he is my son.
Mark Leon Goldberg [00:30:21] Lastly, I know we’re over time. Is there anything else you want to add or any point you wanted to make?
Mark Cancian [00:30:27] The one point I would emphasize is, you know what a shock this is going to be for not just the American people but the military services, particularly the Air Force and the Navy. Their leadership has been saying all the right things. That is that a great power conflict would be very different from what they’ve done for the last 25 plus years, and that services need to be ready for this different kind of environment. The problem is that the majors, the lieutenant colonels, you know, the middle ranks, their experience of war is in sanctuary. And culturally, that’s very hard to change. When I talk to Air Force audiences, I tell them, imagine a situation where the reinforcing aircraft arrive on Okinawa, Kadena Air Force Base. They land on a bumpy runway because it’s been struck by Chinese missiles so many times. They taxi past literally hundreds of wrecked aircraft that were destroyed on the ground. They move into a barracks that was vacated by the previous squadron because they were all killed in the previous missile strikes. The hospitals full of wounded. The golf course has been turned into a cemetery and they’re told, welcome to Okinawa. Tomorrow you fly over Taiwan. And that’s an experience the Air Force hasn’t had since 1945.
Mark Leon Goldberg [00:31:46] Mark, thank you so much for your time. Very sobering scenarios that you outline.
Mark Cancian [00:31:52] Thanks for having me on the program.