I've been using Boldin and other retirement calculators, and I feel pretty comfortable with the projections and chance of success. Since all of my retirement money is in tax-deferred 401(k) or IRAs, I'm curious to know what benefit a financial advisor would be (vs the cost). Other than doing Roth conversions over time, all of my withdrawals will be taxed as income, so I don't know if it's worth having someone advise me on how to withdraw from my accounts over time. It seems pretty straightforward - withdraw what I need, taking into account the income tax I will have to pay on that money. But as they say, "you don't know what you don't know."

I have a decent understanding of how retirement planning works, but it is definitely not my area of expertise. Maybe a one time session with a CFP just to get an overview or info on anything I might have missed would be worth it. Thoughts?

So I just wasted a solid half-hour with the in-app AI tool, trying to eliminate IRMAA and setup our medical expenses properly. It simply can't be done. The methods for both 'estiamted' and 'itemized' expenses don't fix it, no matter how I zero it out.

Supposedly the AI opened a ticket for me, but I'm not holding my breath for any human at Boldin to look at the issue it opened. 😞

I'm just wondering if anyone has gotten it working properly if you and your spouse aren't going to be using Medicare Part B.

What are the best videos out there to tell me how to use Boldin properly? It’s fairly self explanatory in many ways. But I’m having trouble understanding what it means in some charts. Like when I have enough $ with 99% chance of success and yet a bar graph showing me shortages in this or that account and what if anything to do about it. Got any video suggestions?

Edit- thanks I didn’t realize there were videos in the app itself.

Is the chance of success just overly pessimistic or is there a flaw in its calculation?

I have a scenario that assumes a 40 year retirement with an average withdrawal rate of 2.9%/yr. Withdrawals peak around 5% in the first decade due to Roth conversions, which Boldin still treats as a withdrawal for some reason. Then after the Roth conversions are complete, the withdrawal rate drops to about 3% and eventually down to about 1.5%. So the actual spending rate is probably closer to around 2%. However, with average inflation and returns, Boldin says this scenario has an 80% chance of success.

This just does not pass the sniff test. Wade Pfau's study in 2011 showed that a 2.8% withdrawal rate over 40 years, in conservative market conditions, only failed 1% of the time. In other words a 99% success rate.

Don't get me wrong, I would prefer Boldin error on the side of caution, but this estimate seems way to pessimistic to the point where I think it may be an error. The AI agreed it should be higher. It told me that the low success rate was due to a high legacy goal. I remove that and it had no impact. Next it said it was because of my roth conversions. I removed those and the percentage went down. Then it told me I had a liquidity problem because my brokerage account would be empty before I could pull from the roth, which is nonsense because I am already well past the 5 year rule. Even still I switched to a traditional withdrawal order and the chance of success dropped to 75%.

https://papers.ssrn.com/sol3/papers.cfm?abstract_id=2544656

For the past several months I've been evaluating and testing the 4 major LLMs / AI environments to create a financial plan for myself and subsequently a few friends.

The result is several magnitudes greater than anything I've had before, but each one is very specific to each person, due to age, state of residency, filing status, sources of income, list of objectives - a laundry list of things effect a good plan.

I tried to get Claude (my tool of choice) to create a template that I could share with others as a Do-It yourself prompt document or .py script depending on their expertise. Unfortunately, this proved impractical due to all of the complexities mentioned above. Even Claude warned that this was going to be impractical as any output would be too specific and confusing for most to use.

So, instead I have created (with the help of Claude) a How-To get started guide. While this should work with any LLM, I've only tested extensively in Claude. As for Free or Pro - you could get started with Free, if you are willing to both use it after prime time hours (M-F 8am to 2pm EST) and weekends, and 'continue' a lot, but it may get you started. OR - spend $20 and jump in with both feet.

This is the link if any are interested is a shared, web published google doc: https://docs.google.com/document/d/e/2PACX-1vSDAUUn7oRZYv-YqdlJ0X2IBnS70rGeAADYlvjNZObtft0v3FVgNcThmbvFwkXRKXKmpQBpvm8KG6Jp/pub

Even though I have used AI to create my plan and it does many things Boldin (and others) cannot, it also exemplifies why something this broad would be very difficult, if not impossible, for Boldin or anyone else to implement. Boldin is great for what is and is improving. I still use it to backtest, visualize and create FPs for my adult children. BUT - it has limits on what it practically can be expected to do. It's great for those who are <40 with more typical needs. But for old codgers like myself or someone with more complex needs, this may provide you some information that is helpful.

All questions/comments welcome.

Hello, I use Boldin and am thinking about using their ability to link to my brokerage accounts. Trying to understand how many people do this and if there are risks.

View Poll

Here's what it said. So the tool designed to optimize tax outcomes doesn't understand tax brackets? Seriously...!!!!

The Logic is "Greedy," Not Strategic

The algorithm has one single goal: minimize the total sum of tax dollars paid between now and age 100.

• The Growth Problem: It sees your ~$4M in tax-deferred accounts growing at ~8% (your moderate return assumption). By the time you hit RMD age, that balance is projected to be massive.
• The "Rip the Band-Aid" Bias: It calculates that paying 37% on $4M today is "cheaper" in total nominal dollars than paying 24% on a future $15M+ balance.
• Bracket Blindness: It does not "know" that jumping from the 24% bracket to the 37% bracket is a bad move. It only knows that the future "tax bomb" is mathematically larger than the current one.

Have any of you exported your Boldin plan to Claude (or similar AI model)? If so, how’d you do it, and did you find the comparative analysis helpful?

Did you just use the Boldin export or did you go the MCP/agent route (or both)?

I'm a Federal employee that will NOT be taking Medicare B/D in retirement due to IRMAA costs. I will instead maintain my Federal BCBS health care in retirement. I set Medicare Expenses (65 through longevity) to Itemized (no expenses) with a cost of $0 through my wife and my lifetime estimate. Despite this, Boldin insists on adding in an IRMAA cost.

1. How do I eliminate this cost?

2. Boldin - will you please just give us a "Decline Medicare" option toggle that turns all of the Medicare costs off?

Hello reddit,

I'm trying out Bolden now. We've tracked our expenses for years so I have a pretty good handle on our "must spend" vs "like to spend" in our categories.

When does "Must spend" have any affect on anything in the detailed budgeter?

As an example, in the sheet I'm playing with, I have a chance of success rate /Monte carlo of 99%.

If I now add a Travel line, with Must Have: $5000 and "Like to spend": $5000, then the chance to success and monte carlo drops to 81% success. Fair enough.

What I don't understand is, if I set "Must Have: $0" and "Like to spend:" to $5000, then the chance to success and monte carlo stays at 81%. Why does "Must Have" have no affect here?

I would expect that to be MUCH closer to 99%, since IMHO must:0,like:5000 means "I will spend up to $5000 if I can afford to" which shouldn't be hammering my chance of success.

I'm trying to "balance things out" somewhat in my monte carlo. It's an 81% success rate, but my final "Monte carlo median" value is 7x my starting amount. It feels like I should be able to spend more/guardrail harder, and on average be able to spend more without hurting my overall chance of success. But I can't figure out the settings for that.

I overlooked this. Instructions are to enter gross pay less pre-tax deductions. But post-tax deductions, health and life insurance, Roth 401K are not accounted for when entering income data. Given my post tax deductions are 6x more than pre-tax this is significant. Do these get entered into expenses? How do the pre/post deductions that are entered manually impact social security calaculations?

I am trying to find my Roth conversion sweet spot. Instead of up to 24%, on suggestion of the often wrong AI tool, I changed that to tax optimized. With retirement age of 56 the "optimized" strategy converts $4 million in like 5 years, by age 60. I asked the AI to explain and it told me the system is trying to avoid a massive RMD. That is both a bonkers result and explanation.

I checked my birthday and my wife's in the tool and they are correct. The Roth conversion tool is great!! But, we were both born after 1/1/1960 which pushes our RMDs to 75. But, the tool has it starting at 72. How do I correct it? Sorry, I should have said that this is from the Roth Conversion Explorer. Where should your converted funds go?>>>>>What strategy would you like to model? (I picked lowest lifetime taxes)>>>>Is it okay to pay for the tax liability of Roth conversions with tax-deferred funds? YES>>>>>Would you like to set a specific start and stop age for conversions? NO>>>>SUMMARY.

SOLVED: humblequest22

Your spouse turns 75 in 2035 and that is the first year there is a non-zero RMD. Your age isn't a factor yet.

(I already gave an answer, but this is the direct answer, so I posted again.)

https://preview.redd.it/5uq2smep2wzg1.png?width=996&format=png&auto=webp&s=426bc5b49271fbfc4db78d8f282f767c8eb714b8

From the Required Minimum Distributions (RMDs) chart

I think one of the biggest variables that is overlooked in financial planning is what numbers are you using for your longevity numbers. Financial planners tend to want to use 95 for both or sometimes 90 for the male 95 for the female. But everybody's different. Anyway it's a big variable and so I wanted to make sure I wasn't being too conservative by having our longevity numbers too high. So I use Claude. I asked it to create me a questionnaire of all the things they would need to know to calculate our longevity ages. Things like our current lifestyle and information about our parents and sisters and brothers and aunts and uncles in terms of when they died and from what. List of medications were taking list of supplements were taking, etc. after filling that out and uploading it to Claude in addition to our quest lab reports for each of us it came back with a very nice report that showed a conservative age of most likely age and an optimistic age. Included all of the items that are working in our favor and also all the items that are working against us. So what I did is for my baseline I used the most likely numbers. Then to stress test that I created a scenario with the optimistic numbers. And then compare the scenarios, some very interesting results. Living longer lowered the chance of success which one might expect, but living longer also left us more at the end of the plan. When inquiring about this, we have more money at the end of the plan because we have more years of collecting Social security, and since I'm not going to collect Social security, either of us, until age 70, we're well beyond the break-even point in the long longevity scenario. But the reason the chance of success goes down even though value and plan goes up, as Claude explained it, is that it's five more years that the Monte Carlo simulations are using which means five more years of potential downturns in the market.

Just thought I would share this in case anyone else was interested in giving it a try. I feel I have a better, more realistic plan, of course no one knows when our time will come, so the best we can do is take our best guess based on data.

As of now, this is some hot garbage. It gives me instructions to go to pages that don't exist, click on icons that don't exist, etc. Hoping it gets much better but until then, don't trust it with any other information if it can't get those simple things correct.

As retired military and federal employee, i have both traditional and ROTH TSP assets. They are not strictly like an IRA or a 401K. But Boldin does not have TSP as an account type. Surprising considering how many retired military and federal workers there are. How are folks modeling TSP?

Any security issues with adding a Boldin pdf plan into the AI platforms outside Boldin? I noticed the other day using Google Gemini that it tied a query of about a month ago to the present query, but it was either my address or retirement date. Something personal. So wondering if what we put out to AI with Boldin retirement planning questions should be done differently. Any techniques better for using AI? Inside Boldin for AI I assume is okay, but maybe not so either?

My wife and I have a plan for charitable giving that involves targeting most of our pre-tax IRA holdings to charity while mostly living off Roth and Brokerage holdings plus Social Security. We have a number in mind (in today's dollars) of how much we want to set aside for charity. My plan is to use Roth conversions to reduce our pre-tax IRA balances so that when we match our future RMDs with QCDs and set the beneficiary on the pre-tax IRA accounts to our designated charities, the QCDs plus the remainder amounts achieve our giving goals. That way, everything that goes to the charity is pre-tax/untaxed.

That seems to be the most tax-efficient way to do charitable giving considering our balances and giving goals and it also means that whatever we leave to our heirs (Roth and brokerage balances) does not incur taxes. Another advantage is that if a sudden life change, market crash or other event occurs that breaks our plan, we can reduce our planned charitable giving as a buffer (sorry charities).

So the plan is to use RMDs in early retirement to reduce our pre-tax IRA holdings to a level where the projected RMDs and remaining balances basically match our giving goal. I've struggled with how to model this in Boldin, but finally found a way that seems to work.

If I simply let the RMD explorer maximize estate value, it ends up creating larger conversions in the early years, draining the pre-tax IRA balances to the point where there isn't enough left to do the charitable giving we want to do. And we incur taxes now converting amounts that would otherwise have gone tax-free to charities.

I've now modeled this in Boldin by entering a series of expenses from our IRA accounts at percentages that match what our RMDs would be for that year. Marking them as 'tax-deductible' apparently allows Boldin to treat them like QCDs and entering them as percentages means they adjust automatically when I model different Roth conversions. This was tedious (20+ entries for each of us, each with a distinct year and percentage), but it got the job done.

Now I can tinker with my planned Roth conversions and look at the projected RMDs and what the remainder balances are at our death to try to arrive at the right amount of charitable giving.

A few minor issues...

• Boldin's Roth Conversion Explorer really wants to get reduce those RMDs, even though they're offset by QCDs. No big deal--it just means I can't use the options for maximizing estate value, but it does mean that I get a persistent warning in my dashboard and financial wellness report telling me I'm missing out on Roth conversion opportunities. In the end, I've created a manual series of Roth conversions that seem pretty optimal for our goals.
• I do wish Boldin had more explicit QCD support. I suspect if our RMDs exceeded the QCD limits, Boldin would be oblivious to it.
• Entering all those QCD expense entries was extremely tedious. I'm not sure how Boldin could address this to make it easier since it seems like a pretty unique requirement, but it sure would be nice. The Boldin AI did do a good job of auditing it and telling me where I'd made mistakes on a few of the entries.
• Adding up all the RMDs and final balances is a pain and I'm not totally sure if I'm thinking about "today's dollars" vs "future dollars" correctly when looking at the amounts. For now, I've offloaded this to the Boldin AI and I'm hoping that it's correct when it tells me the total amount in today's dollars.

I'd be interested in what others think of this approach, or if others have found another way to accomplish this.