This is what the coach does, in full, with nothing cut.
Below is an entire Grindstone session, taken from a debrief a learner chose to share. Their identity and firm are removed; every word of the exchange, the scores, and the coach's assessment is exactly as it happened. Read it and decide for yourself whether it is worth your time.
Ring I · Analytical judgment
An office building where the model was clean and the rent roll was not
Nine tenants, a 38% anchor expiring inside the hold, and a model that applied one renewal probability to all of them.
Private coach
“You have the model, the rent roll, and 48 hours. What do you do, and specifically, what are you looking for?”
You
Before I touch the IRR, I want to stop treating the rent roll as the deal. A uniform 75% renewal across nine tenants is a modeling convenience, not a thesis. I'd pull the actual leases for the top five exposures and build a tenant-by-tenant renewal view: anchor first, since the insurance firm is 38% of the building and its Month 38 expiration sits inside the hold.
Private coach
“Good. Model the anchor's departure at Month 38, not its renewal: what does the deal look like then, and who do you actually need to call this week to price the probability that call goes against you?”
You
At the anchor's departure I'd hold the 24,200 SF dark for 12 months, re-tenant at market with downtime and TI of $40 to 60 per SF, and assume the cap moves out 75 to 100 bps on the exit, because a 51% leased building doesn't trade as stabilized. I'd call the anchor's regional facilities lead and the listing broker for any competing block in the submarket; if either of them is quiet, that's the signal.
Private coach
“Right. 'Government-backed' is a framing, not a credit. What specifically would you need to see in the federal contractor's lease, and in the parent contract behind it, before you let that label carry any weight in the renewal probability?”
You
I'd need the lease's appropriations clause and any early termination right tied to non-renewal of the underlying federal contract. Then the contract itself: who's the awarding agency, what's the period of performance, are there option years, and is funding obligated or just authorized. If the contract expires inside our hold and the lease lets them out on non-renewal, the federal label is worth zero.
Private coach
“Now the harder one. If you do all of this and your senior associate still says the model is clean and IC is Friday: what specifically do you bring into the room, and what are you asking the IC to decide?”
You
I'd bring a one-page tenant-by-tenant renewal view with the anchor's departure modeled, two scenarios on the federal contractor, the nonprofit's termination right quantified, and a revised IRR band, not a point estimate. I'm not asking IC to kill the deal; I'm asking them to price the anchor's departure as a base case, not a downside, and to decide what the deal looks like at 51% occupancy.
Private coach
“That is the move. The model accepts; the analyst interrogates. The 75% probability was never the question; the question was always whether nine different businesses each independently want to be in that building in 36 months. Why is that the only question that matters here?”
You
Because the deal is priced as if the rent roll is the cash flow. It isn't. The rent roll is a snapshot of leases that happen to overlap today. The cash flow is nine separate business decisions made by nine counterparties who don't care about our IRR. The job at this stage isn't to defend the model; it's to replace it with an underwriting that takes each of those decisions seriously, anchor first.
The debrief
You refused to defend the model and put the deal back where it belongs: on a tenant-by-tenant business view, with the anchor's departure priced as the base case rather than a downside scenario.
Modeled the anchor's departure with downtime, TI, and an exit-cap adjustment rather than a haircut.
Asked for the appropriations clause and the underlying federal contract by name, not a summary.
Named the uniform 75% renewal as the structural fault and reframed nine tenants as nine business decisions.
Treated 'government-backed' as a framing and priced the anchor's facilities lead as the real signal.
Brought the disagreement to IC as a repriced underwriting, not as a kill recommendation.
What they did well
- Named the 75% uniform assumption as the structural fault before touching the IRR.
- Treated the anchor's Month 38 expiration as the base case, with explicit re-tenanting downtime, TI, and an exit-cap adjustment.
- Reframed the IC ask from 'kill the deal' to 'price the anchor's departure correctly', preserving the relationship and the analytical point.
What to press next
- What does the building's market position look like once a 38% anchor has left in this submarket: who is the next user, and at what rent?
- Which of the remaining eight tenants are signaling the same as the anchor, and how would you know before they tell you?
- What price would you actually pay for this deal underwriting the anchor's departure as a base case?
- If IC declines to reprice and instructs the team to close anyway, what is your responsibility from that point forward?
The clean model was never the asset; it was the trap. The deal you can defend is the one where the model is the floor of the work, not the ceiling.
Your session will read differently.
The coach does not run a script. It follows what you say, and presses where your reasoning is thin. The only way to know what it finds in yours is to work one.