There is a point in some deals where you stop trying to make money. You start trying to prove that you were right. The problem is, nobody sends you an email when you cross that line.

It usually starts with something perfectly reasonable. I bought the property because the numbers worked. I hired the contractor because he had a good track record. I chose the partner because we had complementary skills. I backed the borrower because I believed the problem was temporary.

Then something goes wrong. The $80,000 rehab becomes $95,000. The contractor misses a deadline. The partner starts becoming difficult. The tenant stops paying. The borrower needs another extension. And you say: "Okay. Problems happen. Let's fix it." Which is exactly what a sensible person should say.

Then the $95,000 rehab becomes $110,000. Another deadline passes. The partner has another explanation. The tenant promises Friday. Again. The borrower needs just a little more time. Now there are two numbers involved. The money I might lose if I walk away. And the money I have already spent.

That second number is dangerous. Because $50,000 already spent feels like a reason to spend another $10,000. Three months invested in a relationship feels like a reason to give it another month. Twenty phone calls feel like a reason to make the twenty-first. Otherwise, I have to admit something I really don't enjoy admitting: "Maybe I was wrong."

Economists have a nice clean phrase for this: sunk cost. Real life isn't nearly as clean. Because here's the other side of it: am I supposed to run away the moment a deal gets difficult?

If I fired every contractor after his first missed deadline, I'd eventually run out of contractors. If I abandoned every investment when the first unexpected expense appeared, I probably shouldn't be investing in real estate. If I terminated every business relationship the first time we hit turbulent water, I wouldn't have many long-term relationships. Some of my best relationships have survived arguments, mistakes, bad months and situations where one of us screwed up. And some good deals look absolutely terrible halfway through.

That's what makes this difficult. “Cut your losses” sounds wonderfully intelligent when you're not the one standing in the middle of the mess. Sometimes another $10,000 really does save the original $100,000. Sometimes another month really does solve the problem. Sometimes the contractor who screwed up is exactly the guy I want fixing it because he knows what happened and takes responsibility.

Persistence is not stupidity. But neither is persistence automatically a virtue. So how do I know which one I'm doing?

I've started thinking there are two very different reasons to put more money or time into a troubled situation. The first is: something has changed, and I now have a credible reason to believe the outcome will be better. The second is: i've already come this far.

Those sentences may sound similar when you're emotionally involved. They're not. If the contractor missed the deadline because a supplier failed and now the material is physically sitting on the job site, that's new information. If the borrower needed an extension because a sale was delayed and now there's a signed closing date, that's new information. If the rehab needs another $10,000 but that $10,000 gets a $300,000 property finished and marketable, I can do the math. But if I'm hearing the same explanation for the fourth time? Nothing changed. I'm not financing a solution anymore. I'm financing hope.

And there's another test I like even less: if I walked into this situation today, knowing everything I know now, would I put the next dollar into it? Not yesterday's dollar. That one's gone. The next one. That's brutal because it doesn't care about my pride, my original analysis or how many times I told everybody this was going to work. It only asks whether the decision in front of me makes sense today.

Maybe the answer is yes. Good. Fight for it. Good businesses, investments and relationships are not built by people who jump overboard every time the boat hits a wave. But if the only reason I'm still on the boat is because I was the guy who insisted it was seaworthy... I may have a different problem.

So perhaps the question isn't: When should I give up? It's: What would have to happen for me to admit that I was wrong? If I can't answer that before the next problem arrives, I can probably lose an astonishing amount of money proving that I wasn't.