Inside the Social Security Optimizer: Finding Your Best Claiming Strategy

How does a Social Security optimizer evaluate 81 claiming combinations? Learn the mechanics — two-phase Monte Carlo, dynamic scoring, survivor benefits, and tax interactions.

January 17, 20268 min read

"Delay until 70" is solid advice — until you're a couple with unequal earnings and ACA coverage to protect. Then it gets complicated.

Most Social Security advice treats it as a single decision: when should I claim? But for couples, it's 81 decisions compressed into one. And the right answer depends on things that have nothing to do with Social Security itself — your Roth conversion opportunities, your tax brackets, your Medicare premiums, whether you'll outlive your spouse.

I built the Social Security optimizer to find the combination that actually works best for your situation, not the one that sounds best in a magazine article. This isn't a recommendation engine that tells you what to do — it's a comparison engine that shows you the tradeoffs.

This post explains how it works under the hood.

The Problem: 81 Combinations, All Plausible

For a single person, claiming age is a slider from 62 to 70. Nine options. You can evaluate them by hand if you're patient.

For a married couple, both spouses choose independently. That's 9 × 9 = 81 combinations. Primary claims at 64, spouse at 68. Primary at 70, spouse at 62. Every permutation produces a different income pattern, different tax picture, different survivor outcome.

You can't evaluate 81 strategies by hand. And even if you could, you'd be comparing them on the wrong metric.

Why Breakeven Analysis Fails

The standard approach compares total lifetime Social Security payments. Claim early, get more checks. Claim late, get bigger checks. Calculate when the late-claimer catches up.

This misses everything that matters:

Tax interactions change the picture. A couple claiming $80,000 combined at age 70 faces different taxes than one claiming $50,000 at 62 — not just on Social Security, but on their entire tax return. The $80,000 might push them into a higher bracket. Or it might reduce how much they need from taxable accounts, lowering capital gains.

Roth conversion windows open and close. Between retirement and Social Security, income is often at its lowest. That's prime territory for Roth conversions. Delaying SS widens that window. Claiming early shrinks it. The optimizer coordinates these decisions automatically, in two stages: the screening pass scores all 81 combinations against one shared conversion schedule, and then the five finalists — plus the reference schedule they're compared to — each get their own full Roth re-optimization.

Survivor benefits dominate for couples. When one spouse dies, the survivor keeps the higher benefit. The other vanishes. If the higher earner claimed at 62 instead of 70, the survivor is stuck with a 30% smaller check for potentially 20+ years. For unequal-earner couples, this single factor often swamps everything else.

Success rate matters more than dollars. A strategy that maximizes expected lifetime income might still leave you broke if markets crash early. The optimizer runs Monte Carlo simulations for each strategy, measuring not just dollars but probability of running out of money.

How It Actually Works: Two-Phase Evaluation

Testing 81 strategies thoroughly is computationally expensive. Each one requires running the Roth optimizer, then simulating hundreds of market scenarios. Do that naively and you're waiting minutes for results.

The solution is a two-phase funnel:

Phase 1 (Screening): Test all 81 combinations with 75 Monte Carlo trials each, against a single shared Roth conversion schedule — one optimizer run, struck at full retirement age for both spouses, not 81 of them. That's enough to identify which strategies are in contention. The bottom 76 get eliminated.

Phase 2 (Refinement): Re-evaluate the top 5 with 300 trials each, and this time each one gets its own complete Roth re-optimization. More trials and a schedule fitted to that combination mean more accurate success rates and tax estimates. Every number you see comes from this phase.

This cuts computation by 4x while maintaining accuracy for the strategies you'll actually consider.

The Scoring System

After simulation, each strategy has three metrics:

Success rate: Percentage of market scenarios in which every year of the plan was funded — a trial fails when spending goes uncovered, not merely because the portfolio ends at zero
Lifetime taxes: Total federal, state, capital gains, NIIT, and Medicare premiums (in today's dollars)
Ending balance: Median portfolio value at life expectancy

These get combined into a single score using weights that shift based on your situation.

When any strategy has less than 90% success rate:

Success rate: 50%
Taxes: 30%
Ending balance: 20%

Safety dominates. If some strategies risk running out of money, the optimizer prioritizes avoiding that outcome.

When all strategies have greater than 90% success:

Success rate: 30%
Taxes: 40%
Ending balance: 30%

Now all paths are safe. The optimizer pivots to tax efficiency and wealth accumulation.

Why the Weights Make Sense

A natural question: why should a 2% success rate difference outweigh $50,000 in lifetime taxes?

Because the outcomes are asymmetric.

Running out of money at 87 is catastrophic. You can't un-retire. You can't reclaim decades of spending. The downside is unbounded.

Paying extra taxes is suboptimal but survivable. You have less wealth, but you still have wealth. The downside is bounded.

When failure is plausible, avoiding it matters more than optimization. When failure is implausible, optimize freely. The dynamic weights encode this logic.

Couples: Survivor Benefits Are Often the Whole Story

For single filers, the optimizer is choosing between 9 relatively similar options. Success rates cluster. Taxes vary predictably.

For couples with unequal earnings, the decision is fundamentally different. The higher earner's claiming age determines what the survivor receives for potentially decades.

Consider a 65-year-old couple: primary earner's benefit at FRA is $3,000/month, spouse's is $1,200/month. Both expect to live to 90.

If both claim at 62:

Combined benefit while both alive: ~$2,940/month
Survivor benefit after first death: ~$2,100/month

If primary delays to 70, spouse claims at 62:

Combined benefit while both alive: ~$4,560/month (once primary starts)
Survivor benefit: ~$3,720/month

That's $1,620/month difference in survivor income — nearly $20,000/year for potentially 10-20 years. The "maximize survivor" strategy wins not because it produces more total dollars, but because it protects the vulnerable years.

The optimizer tests all 81 combinations and surfaces this pattern automatically. For couples with large earnings gaps, some variant of "higher earner delays, lower earner claims earlier" almost always ranks near the top.

What Gets Modeled, What Doesn't

I want to be direct about what the optimizer handles and where it has limits.

What's included:

All 81 claiming age combinations (couples) or 9 (singles)
Roth conversion re-optimization for the five finalists and the reference schedule
Monte Carlo simulation with 99 years of historical returns
Federal and state taxes, capital gains, NIIT, Medicare premiums
Survivor benefits calculated and displayed for every scenario
Custom slider testing for any combination you want to try

What's not included:

Spousal benefits aren't automatically detected. If your own work history would give you less than 50% of your spouse's PIA, you may be entitled to a higher spousal benefit. The optimizer won't find this automatically, but you can enter the spousal benefit amount directly and test scenarios with the custom sliders.

Tax law will change. Current brackets, inflation-adjusted forward. If Congress changes the rules, projections become wrong. No model predicts legislation.

Markets don't follow projections. The Monte Carlo uses 99 years of historical returns (1926-2024), which is the best stress test available. But past patterns don't guarantee future results.

Life expectancy is your estimate. You enter your expected lifespan; the optimizer uses it as fixed. Consider setting it conservatively (e.g., 95) to stress-test against longevity risk.

Running It Yourself

The optimizer is built into the Calculator page. Enter your Social Security amounts, then click "Compare Strategies." You'll see the top 5 options ranked, with success rates, taxes, and balances compared to a baseline.

You can also drag the custom sliders to test any combination. The optimizer evaluates your custom choice in real-time and shows how it compares to the ranked strategies.

For couples, make sure you've entered spouse data — the optimizer needs both benefit amounts and birth dates to search the full 81-combination space. The Social Security claiming calculator shows the ranked output for a sample couple, if you want to read one before running your own.

Common Questions

"Why does the optimizer sometimes recommend claiming before 70?"

Because delaying isn't always optimal. If you need income to bridge to Social Security and drawing from your portfolio would hurt your success rate, earlier claiming can win. If ACA subsidies are at stake, staying under income thresholds might matter more than maximizing benefits. The optimizer weighs all factors.

"How does this interact with Roth conversions?"

Directly, but not uniformly. Screening all 81 combinations shares one Roth optimization — a single run struck at full retirement age for both spouses — because 81 full optimizations would take minutes. The five finalists and the reference schedule then each get their own complete Roth re-optimization, and that is where every displayed figure comes from. So the effect is priced properly exactly where it decides the ranking: a strategy that delays SS often enables significantly more tax-efficient conversions during the gap years, and the finalist stage is what measures that.

"What if my spouse and I have very different life expectancies?"

The plan runs until the later of the two life expectancies. If you expect your spouse to outlive you by 10 years, the survivor benefit becomes even more important. The optimizer accounts for this automatically.

"Will I get the same numbers if I run it again?"

Within a session, yes — exactly. The Monte Carlo draws its market sequences from a seed fixed when the tab loads, so re-running the same plan reproduces the same success rates to the decimal. Redrawing is a deliberate action: Different Markets rotates the seed and re-runs in one step. A fresh tab starts from a fresh seed, so figures can shift by a point or two between sessions, and close contenders can swap positions. If the same strategy keeps winning across different market draws, that's a strong signal.