A Retirement Advisor Can Optimize Your Money. No One Can Optimize Your Life.

The Job the Software Does

Ask ThunderHarbor which withdrawal order produces less lifetime tax and it gives you a number. Ask it which ACA plan qualifies for the biggest subsidy and it gives you another number. Both answers are real and worth having. The mistake is assuming the smaller number is automatically the better decision, because the software has no way to weigh in on what you gave up to get it. It only sees the dollars that moved. It never sees the doctor you switched away from, or the deductible you are now paying out of pocket at exactly the age you are most likely to need it.

A Plan Is a Baseline, Not a Script

A projection cannot predict what the next twenty or thirty years of your life will look like, no matter how many scenarios it runs. Markets will move on their own schedule. Health will change without asking permission. A grandchild will need help nobody budgeted for, or a market drop will open a door nobody planned around. None of it will play out exactly the way the numbers on the page describe.

That gap between the plan and the life you live does not make the plan pointless. A baseline exists so that the moment something changes, you already know whether your numbers can absorb it. Without one, every surprise turns into a guess made under pressure. With one, a health scare, a bad year in the market, or a grandchild’s unplanned wedding become questions you can answer instead of problems you only react to.

An Example: The Plan With the Bigger Subsidy Wasn’t the Better Plan

Walter is 64, retired two years ago, and not yet on Medicare. His income qualifies him for an ACA plan around $640 a month in New York City, comfortably under the 400% federal poverty line cutoff. He looked at that plan for about a week, then bought a different one for $1,080 a month instead.

The subsidized plan he skipped had a $9,200 deductible and a narrow network that did not include his cardiologist of eleven years. The plan he chose has a $2,500 deductible and keeps every doctor he already sees. The difference costs him about $440 a month, $5,280 a year, money the cheaper plan’s own subsidy math says he did not have to spend.

Walter is not confused about the subsidy. He read the numbers and picked the more expensive plan anyway, on purpose, because keeping his cardiologist and avoiding a $9,200 bill in a bad year is worth more than $5,280 to him. That is not a mistake a better calculator would catch. It is a decision the calculator was never equipped to make in the first place.

This Shows Up Far Beyond Insurance

The same trade shows up in smaller ways all over a plan. Someone delays a long-planned trip by a year because taking it now would realize enough capital gains to trigger an extra $2,000 in Medicare surcharges eighteen months later. Someone puts off a roof repair they can clearly afford because the withdrawal to pay for it would push them $3,000 over the ACA subsidy cliff for one year. In both cases the math is correct. The thing being protected, a trip, a safe roof, is not made of dollars, and the plan never asked whether it was worth protecting.

Where Optimization Should Stop

None of this is an argument against optimizing. Converting to Roth in the right bracket, harvesting a gain at 0%, and timing a withdrawal around an IRMAA threshold are all real, quantifiable dollars. Leaving them on the table for no reason is still a mistake. The argument is about what happens after that work is done and the plan already holds up.

A finished optimization pass and an unfinished one look identical from the inside. Each additional dollar you find still feels like the same kind of progress it always did, so nothing tells you to stop. The habit that got the plan funded, treat every choice as a place to save money, does not come with an off switch. It keeps running on whatever is left, even once there is nothing left worth cutting for.

What makes this easy to miss is the asymmetry between the two kinds of cost. Underspending on optimization shows up as a number, a few thousand dollars a year, easy to calculate and easy to regret. Overspending on it shows up as an absence, a doctor you stopped seeing, a trip you keep putting off, a roof repair you keep delaying past the point it needed fixing. None of that appears in a projection. A plan can look strictly better on paper while it is quietly costing more than it saves.

We are not going to tell you where that line sits in your plan. That number does not exist, and software that claimed to know it would be lying to you the same way the subsidy math lied to Walter about which plan was cheaper. What software can do is make sure the question gets asked at all. Once a plan already covers what you need it to cover, is the thing you are protecting worth more than what protecting it costs? Years spent converting, harvesting gains, and staying under a subsidy cliff bought room, not a smaller number on a page. Once the room is there, the only mistake left is forgetting to use it.

There is a plainer reason the line matters too. Nobody reaches retirement by accident. It takes three or four decades of showing up for work, saving money that could have gone somewhere else, and running a life on someone else’s schedule. What retirement earns you is the first stretch of life spent answering to no one but yourself, built slowly and on purpose, not handed out the day you turn a certain age. A plan that keeps shrinking every number in sight long after the baseline already holds ends up spending part of that earned freedom still working, without a paycheck.

Walter’s choice was never about health insurance. He already knew, from the plan itself, that his baseline could absorb $5,280 a year without putting anything else at risk, so he spent it on the one thing all that saving was ever supposed to buy him, a comfortable retirement with the doctor he already trusts. The calculator’s job ended the moment it answered that question. What Walter did with the answer, choosing comfort over a bigger number in a spreadsheet, was his to make, and it was the whole point of building the baseline in the first place.