History and Horse Races - Run Only Once

July 5, 2011
Posted by Jay Livingston

Here’s sentence from a New York Times op-ed by David Greenberg last week.
Modern Republican isolationism began with the 1919 battle over joining the League of Nations, when Senate Republicans, led by so-called Irreconcilables like William Borah of Idaho, killed the deal - even though without American guidance, European affairs were doomed to explode again.
I’ve read sentences like this hundreds of times – I’ve probably written some myself. But last week, I had been reading Everything is Obvious* (*Once You Know the Answer) by Duncan Watts, and I got stuck on doomed. “European affairs were doomed to explode again.”

Were they? Was the explosion inevitable? And was the absence of US guidance a necessary cause of the explosion?

It certainly seems that way. We know that Europe exploded in war twenty years after that 1919 decision by the US to stay out of the League. But we know that only in retrospect. At the time, there may have been several other equally plausible outcomes.

It’s like saying that the 2011 Belmont Stakes was “doomed” to be won by Ruler on Ice. True, he did win. But before the start of the race, there were many other plausible outcomes, most of them more likely. If the same race were held tomorrow – the same twelve horses running on a similarly sloppy track – would Ruler on Ice inevitably win? I doubt it. (If he were again 24-1, I might be tempted to put a small bet on his nose, but I wouldn’t be at all confident of collecting.)

The trouble with history is that, like the 2011 Belmont, it’s run only once. And after it’s run, historians and op-ed writers do the same thing that horseplayers do after a race: they go back to the Racing Form, the past performances, and pick out the information nuggets, often pebble-sized, that account for the results. Sometimes they even add their own speculation as fact. A horseplayer might say that Ruler on Ice would also have won the Kentucky Derby if only he’d been entered, a statement for which we have zero observations. An op-ed writer might say that the US presence in the League of Nations would have prevented World War II, a speculation based on a similar number of observations.

Watts suggests a different way of looking at history. Here, for example, is his take on the surge in Iraq. In the fall of 2007, the US upped its Iraq force by 30,000 troops. By the next summer, violence had substantially decreased. Conclusion: the surge worked. It reduced the violence.

Or did it?
Many other things happened between the fall of 2007 and the summer of 2008 as well. Sunni resistance fighters, seeing an even greater menace from hard-core terrorist organizations like Al Qaeda than from American soldiers, began to cooperate with their erstwhile occupiers. The Shiite militia – most importantly Moktada Sadr’s Mahdi Army – also began to experience a backlash from their grassroots, possibly leading them to moderate their behavior. And the Iraqi Army and police forces, finally displaying sufficient competence to take on the the militias, began to assert themselves, as did the Iraqi government. Any one of these factors might have been at least as responsible for the drop in violence as the surge. Or perhaps it was some combination. Or perhaps it was something else entirely.
We can’t know because history is run only once.

Ideally, we should be able to go back, examine our variables, look at the possible scenarios, and assign each outcome a probability. Maybe the outcome that did happen would have a lower probability than others, and we would wind up saying that what happened was a fluke or at least improbable. But that’s not what we do.

Rather than producing doubt, the absence of “counterfactual” versions of history tends to have the opposite effect – namely that we tend to perceive what actually happened as having been inevitable.
As I was reading this part of Watts’s book, I kept thinking of Hans J. Morgenthau, the great political scientist. This was many years ago, perhaps during Vietnam, and after Morgenthau’s analysis, typically thoughtful and insightful, someone in the audience asked him what we could expect as the outcome. In his elegant, German-accented English. Morgenthau said,
Well, the answer to that is that I am a professor, not a prophet. I cannot tell you what will happen. I can only tell you why what did happen was absolutely bound to happen.

Will Ruler on Ice beat those same horses again? Will another surge reduce violence? Maybe, but don’t bet on it.


Bob S. said...


Why does it have to be an 'either/or' situation regarding the surge?

Perhaps the surge worked because it coincided with the other events happening.

Isn't it possible that if the surge hadn't happened the Sunni resistance would have cooperated with Al-Qaeada or or stepped up their efforts to take control instead of cooperating?

Isn't it possible that if the surge hadn't happened, the Iraqi army and police wouldn't have had the confidence and support to step up?

I'm not saying the surge by itself reduced violence but it was an integral factor in the events.

Seldom can one item be the sole cause of success or failure, but the absence of that item can make that difference.

Jay Livingston said...

Yes, it’s possible. That’s exactly Watts’s point. Lots of scenarios are possible. The trouble is that people – even people who style themselves social scientists – talk about an outcome not as one of many possibilities but as a unique inevitability. The trouble is that we cannot know in any scientific way.

If you have a sore shoulder, and I tell you to do some aromatherapy – inhale the fumes of hibiscus, lavender, and crabgrass – and tomorrow, you shoulder feels fine, it’s possible that the aromatherapy worked. It’s also possible that something else, or nothing except time, was what made the soreness go away. And it's possible that the aromatherapy, even though it didn't cure your shoulder all by itself, was a necessary factor. In order to be more certain, we can do studies with lots of people who have sore shoulders, some getting placebo, others getting the real deal aromatherapy.

But that’s sore shoulders. History is run only once. So, Watts says, we ought to be much more careful about conclusions like “the surge worked,” or even that it was “an integral factor” – i.e., a necessary but not sufficient condition. If we really wanted to test that idea, we would have to do what we cannot do – go back to pre-surge Iraq and rerun history this time without the surge.

Bob S. said...


This is where I disagree.

I think we can say the surge worked. I do think we need to expand on WHY it worked.

But the fact of the matter, during the time of the surge things improved.

Did the training of the Iraqi Army & Police work? Yes.

Did the pressure on militias work? Yes.

The question you seem to be struggling with is "did it work all by itself"?

and the answer is no.

But did the Surge have achieveable metrics?

Yes. Number of bodies in country increased.

Number of patrols/missions/operations/buildings build/people cared for/ etc -- those things are measureable and all evidence indicates the "Surge" worked.

I have the feeling you would have no trouble making the statement "The surge failed" had things turned out differently.

Let me ask you a simple question appropriate for the occasion -- "Did the Revolutionary War work?"

Jay Livingston said...

Watts's point is that you can go back and find evidence that supports an idea about what caused what. But that doesn't mean that this causal mechanism was actually at work.

If violence had not decreased, would I conclude that the surge failed? Yes, but not because of political preferences but because of logic. If my aromatherapy patient's shoulder improves, I cannot conclude that it was the aromatherapy. But if the shoulder does not improve, it's fairer to conclude that the aromatherapy did not work (I guess we'll have to try it with organic crabgrass next time). But even so, I could not then conclude that aromatherapy does not work.

As for the Revolution, it depends on how you define "work." That is, it depends on the goals. If the goal was to remove British rule, then I would agree that it worked. (But if I were going to be absolutely true to Watts's approach, I'd have to look for other possible factors that might have led to the withdrawal of British rule.)

Bob S. said...


You sound like you want to have it both ways.

Can't say that the Surge worked to reduce violence because of other factors but you ignore those other facts if the "Surge" hadn't worked?

How can it be?

But if the shoulder does not improve, it's fairer to conclude that the aromatherapy did not work

By your own logic this doesn't make sense.

The aromatherapy could have worked but other factors (fell and hurt shoulder, didn't stop working, used it too much etc) could overwhelm the influence of the aromatherapy.

Jay Livingston said...

Yes, exactly. I was going to add something to that effect -- the operation was a success, but the patient died. But if we define "success" or "working" as achieving the desired outcome, and if we don't get that outcome, it's hard to say that it worked. That's why I said that even though the shoulder didn't improve, we can't reject the idea that aromatherapy is still a useful treatment (at least, we can't reject that idea yet, just on the basis of this one case).