When you have only one chance to choose, the hardest part is often not recognizing a good option. It is knowing when you have seen enough options to recognize what “good” actually looks like.

That problem appears everywhere. You might be interviewing candidates for an important role, looking for a new home, comparing potential business partners, choosing a vendor, or evaluating a series of opportunities that arrive one after another. The options do not necessarily come in a convenient order, and once a decision is made, going back may not be possible.

This is where a surprisingly simple mathematical idea becomes useful: the 37% rule.

The rule comes from the classic secretary problem, a mathematical decision-making problem designed around a very specific situation. There is a fixed number of choices; they must be considered sequentially; only one can ultimately be selected; and once an option is rejected, it cannot be reconsidered.

Under those conditions, mathematics suggests a strategy that feels counterintuitive at first: do not choose immediately.

Instead, spend roughly the first 37% of your available choices learning what the market looks like. Reject those initial options, use them to establish a benchmark, and then select the first later option that is better than everything you have seen before.

For decades, the most powerful intelligence lived behind the closed doors of quant firms — billion-dollar funds whose algorithms quietly out-traded everyone else.

That era just ended.

Co-Invest by Liquid is the first way to trade directly inside Claude and ChatGPT. Ask your AI to analyze a market, stress-test an idea, or build a position sized to your comfort level, then execute, right there in the conversation. No jargon. No twelve-screen terminal. No guesswork.

It's built for people who want to invest smarter, not gamble harder. You set the risk tolerance. The AI does the heavy lifting. You approve every trade.

The institutions made the game, Co-Invest gives you a way to beat them.

Trade with Co-Invest Today

For a large number of choices, this strategy gives you approximately a 36.8% probability of selecting the single best option. That may not sound overwhelming, but it is remarkably high given how restrictive the problem is. More importantly, the principle behind the mathematics gives you a practical way to think about decisions when you have limited information and cannot simply keep searching forever.

Imagine that you need to hire one engineer and have 12 candidates scheduled for interviews.

Candidate number one looks promising. The temptation is to think, “This person could work.”

Candidate number two is even better. Now there is a new temptation: “Maybe this is the person.”

By candidate number three or four, you may already feel pressure to make a decision because you do not want to lose a strong applicant.

But there is a problem with making that decision too early: you do not yet know how strong the candidate pool actually is.

A candidate who looks exceptional in isolation may look merely average after interviewing ten more people.

This is the fundamental insight behind the 37% rule. The beginning of a decision process is valuable because it gives you information. You are not simply looking for an answer; you are learning the range of possibilities.

With 12 candidates, 37% is approximately four candidates. The mathematical strategy would therefore be to interview the first four, reject them, and remember the strongest candidate among those four. Starting with candidate five, you hire the first person who is better than that benchmark.

If candidate five is better than everyone in the initial group, the process ends there. If candidate five is not better, you continue. Candidate six gets compared against the benchmark, then candidate seven, and so on.

The important detail is that the first four candidates are not being wasted.

They are serving as your calibration period.

Tip: Treat your earliest options as information, not just opportunities. They help you establish the standard against which later choices can be judged.

The percentage is not arbitrary.

The secretary problem has been studied mathematically for decades, and under its idealized assumptions, the optimal stopping point approaches 1/e, where e is approximately 2.718. The result is about 36.8%.

That means the strategy is not simply “look at a few options first.” There is a specific trade-off involved.

If you spend too little time evaluating the initial pool, you may not develop a reliable benchmark. You could encounter an excellent option early, reject it, and then spend the rest of the process hoping something better appears.

If you spend too much time evaluating, you create the opposite problem. You may become very knowledgeable about the available choices, but the best option may already have passed.

The 37% point sits between those two risks.

You sacrifice the ability to choose from the first portion of the pool in exchange for gaining information about what the pool looks like. After that calibration period, you become increasingly capable of recognizing an unusually strong option when it appears.

That is the trade-off at the heart of the rule.

Tip: When you cannot revisit rejected choices, deliberately separate the process into two phases: learning what good looks like and then acting when you encounter it.

Real life rarely behaves as cleanly as the secretary problem.

You may not know the total number of candidates. A hiring process can be extended. A house can come back onto the market. A supplier can negotiate. A business opportunity can be revisited months later.

Most importantly, the options are rarely presented in a truly random order.

That means applying “37%” mechanically to every decision would be a mistake.

The real value of the rule is the principle underneath the mathematics: do not confuse early exposure with sufficient information.

When you enter an unfamiliar decision environment, you need some time to understand the distribution of possibilities. What initially appears exceptional may simply be normal. What initially seems expensive may actually be competitive. What looks like a difficult problem may become much easier once you have seen several alternatives.

This is particularly important when you are making a decision in an area where you lack experience.

If you have never hired an engineer before, for example, your first interview becomes part of your education. You are learning what strong communication looks like, how candidates explain technical decisions, which experiences matter, and which impressive-sounding qualifications actually predict performance.

The same thing happens when searching for a home, evaluating contractors, interviewing agencies, or comparing software vendors.

Your first few options teach you how to judge the next ones.

One of the biggest mistakes in sequential decision-making is falling in love with the first option that appears to satisfy your requirements.

This happens because human beings naturally want closure.

Once a candidate seems “good enough,” continuing the search feels inefficient. Once a house checks most of the boxes, viewing additional properties can feel unnecessary. Once a vendor appears competent, comparing alternatives can seem like wasted effort.

But “good enough” is not the same thing as “best available.”

The 37% rule creates a deliberate period during which you are not allowed to make that mistake.

You are forced to observe.

That forced observation can be valuable because it separates evaluation from commitment. Instead of asking after every option, “Should I choose this?”, you initially ask, “What does this option teach me about the pool?”

That subtle change can improve the quality of your thinking.

Tip: If you notice yourself wanting to commit very early, ask whether you are genuinely confident in the choice or simply relieved to have the decision almost finished.

There is an opposite mistake that can be just as costly: endless calibration.

You can keep gathering information indefinitely.

Another interview might reveal something useful. Another house might appear. Another vendor might offer a different pricing structure. Another candidate might have an interesting background.

The problem is that information has diminishing value.

At some point, additional research stops meaningfully improving your understanding and starts becoming a way to postpone commitment.

That is where the second half of the 37% principle becomes important.

Calibration has a purpose. It is supposed to improve your ability to recognize the right choice later. It is not supposed to become a permanent excuse for avoiding a decision.

Once you have enough information to establish a meaningful benchmark, the process needs to shift from learning to selection.

That transition is often harder psychologically than it sounds.

You may never feel completely certain. You may continue discovering new information after making the decision. The goal is therefore not perfect certainty. The goal is reaching a point where additional information is less valuable than taking action.

Tip: Set a deliberate calibration boundary before the decision begins. Without a boundary, research can quietly turn into procrastination.

There is a broader lesson here that extends beyond the mathematical problem.

The early part of a decision does not have to produce a winner to be useful.

It can produce a standard.

Suppose you are interviewing ten candidates for a position. The first three or four interviews can help you establish what strong, average, and weak candidates look like. You begin to understand the market.

Now imagine that candidate number seven arrives and demonstrates substantially stronger experience, clearer communication, and better judgment than anyone you have seen so far.

Without calibration, you might simply think, “This candidate seems good.”

With calibration, you can say something much more meaningful:

“This is the strongest candidate we have seen so far, and the gap is significant.”

The same principle applies to almost any sequential choice.

You are not merely collecting options. You are building the ability to compare them.

That is why the rule can be particularly helpful for people who feel overwhelmed by too many choices. Instead of trying to evaluate every option against an abstract definition of perfection, you can focus on building a useful benchmark and then looking for something that clearly exceeds it.

Hiring provides one of the clearest examples because candidates naturally arrive sequentially and organizations often face pressure to move quickly.

However, real hiring processes do not perfectly satisfy the secretary problem.

Candidates are not random. Some candidates withdraw. Interview processes can be extended. Teams can sometimes reopen conversations with previous applicants. Multiple people participate in the decision. And the cost of rejecting a candidate is not always irreversible.

So the 37% rule should not become a rigid hiring policy.

Its more useful application is to recognize that the first few interviews should often be treated as calibration.

If a team has never hired for a particular role before, the first candidates can teach the interview panel what the market looks like. Their strengths and weaknesses can expose flaws in the job description, reveal which questions produce useful information, and establish a realistic expectation for compensation and experience.

This can prevent a common mistake: changing the hiring standard after every interview.

Without a benchmark, every candidate becomes a completely new judgment.

With a benchmark, the process becomes more consistent.

Tip: Use early interviews to calibrate the hiring bar, but do not confuse the mathematical rule with a requirement to automatically reject a fixed percentage of real candidates.

TLDR Marketing is the free daily brief that 330K+ growth marketers, performance marketers, and CMOs actually read. The most interesting stories in marketing, curated and summarized in 5 minutes.

Subscribe for free

The 37% principle becomes most interesting when you face decisions involving many sequential options, limited reversibility, and incomplete information.

Consider searching for a home. You may have a budget, a preferred location, and a list of requirements, but you do not know what the available market looks like until you start viewing properties. The first few homes teach you what your budget actually buys.

Or consider choosing a long-term service provider. The first few conversations can reveal normal pricing, common contract terms, typical service levels, and warning signs.

Even something as ordinary as recruiting a freelancer can benefit from this thinking. If you accept the first proposal because it appears reasonable, you may never discover that the market contains substantially stronger candidates at a similar price.

In each case, the initial options create information that improves later judgment.

The key question becomes:

How much information do you need before additional searching becomes less valuable than choosing?

The 37% rule provides one mathematical answer under very specific assumptions. Real decisions require judgment on top of it.

There is something particularly useful about the secretary problem because it exposes a tension that appears in almost every important decision.

You want more information, but information takes time.

You want the best option, but waiting can cause the best option to disappear.

You want certainty, but many decisions cannot provide it.

The solution is not to eliminate uncertainty. It is to design a process that makes uncertainty manageable.

That is what the 37% rule ultimately represents.

It tells you to accept a period of uncertainty at the beginning so that you can make better decisions later. It also tells you that observation has to end. At some point, the purpose of gathering information has been fulfilled, and the next valuable action is choosing.

For someone juggling work, responsibilities, opportunities, and too many decisions at once, this distinction matters.

You do not need to investigate every possibility forever.

You need enough exposure to understand the landscape, a clear benchmark for what qualifies as excellent, and a rule for recognizing when an option crosses that threshold.

The most useful interpretation of the 37% rule is not “always reject the first 37%.”

It is:

Calibrate before you commit.

If you have 100 genuinely sequential, irreversible choices, the mathematical rule gives you a clear starting point: examine roughly the first 37, establish your benchmark, and then choose the first subsequent option that beats it.

If you have 12 choices, that becomes roughly four.

If you have 20, it becomes roughly seven.

But if your decision allows you to return to previous options, negotiate, gather more information, or reopen discussions, the classical mathematics no longer applies directly.

And that is perfectly fine.

The purpose of understanding the mathematics is not to turn every decision into a formula. It is to understand the underlying trade-off well enough to recognize when the principle applies.

Sometimes the right decision is to move quickly.

Sometimes the right decision is to keep looking.

The important part is knowing why.

Tip: Use 37% as a calibration benchmark for sequential, hard-to-reverse decisions—not as a universal rule for every choice you make.

The next time you face a decision involving many options, resist the urge to ask immediately, “Which one should I choose?”

Start with a different question:

“How much of the available landscape do I need to see before I can recognize an unusually strong option?”

That question changes the entire process.

You stop expecting the first option to provide certainty. You stop treating every new option as a completely isolated decision. You begin building a benchmark.

Then, once the calibration phase is complete, you can switch modes.

You are no longer asking whether an option is merely acceptable. You are asking whether it is better than the standard established by everything you have already seen.

That is the real power of the 37% rule.

It does not guarantee that you will choose the best option. The mathematics only gives you the maximum probability of selecting the best option under the assumptions of the secretary problem, and even then, the probability is only about 36.8%.

What it does give you is something more practical: a disciplined way to balance patience with action.

You gather enough information to know what good looks like, but you do not gather information forever.

For a busy person making consequential decisions, that may be the most useful lesson of all.

The best choice does not always require knowing everything.

Sometimes, it requires knowing when you have learned enough to recognize it.

What’s your next spark? A new platform engineering skill? A bold pitch? A team ready to rise? Share your ideas or challenges at Tiny Big Spark. Let’s build your pyramid—together.

That’s it!

Keep innovating and stay inspired!

If you think your colleagues and friends would find this content valuable, we’d love it if you shared our newsletter with them!

As the world becomes increasingly digital, this question will be on the minds of millions of people seeking new income streams in 2026.

The answer is—Absolutely!

Thank you for taking the time to read today’s email! Your support allows me to send out this newsletter for free every day. 

 What do you think for today’s episode? Please provide your feedback in the poll below.

Share the newsletter with your friends and colleagues if you find it valuable.

Disclaimer: The "Tiny Big Spark" newsletter is for informational and educational purposes only, not a substitute for professional advice, including financial, legal, medical, or technical. We strive for accuracy but make no guarantees about the completeness or reliability of the information provided. Any reliance on this information is at your own risk. The views expressed are those of the authors and do not reflect any organization's official position. This newsletter may link to external sites we don't control; we do not endorse their content. We are not liable for any losses or damages from using this information.