Forty parcels out of 1,100, and the bid-down math keeps collapsing

Your improvement value screen is doing less work than you think. Assessed improvement value is a tax figure, not a condition report, and in most counties it's rolling forward off a reassessment cycle that could be years stale. A house that was worth 60% of the parcel value at last assessment can be missing a roof now, and a property with delinquent taxes is exactly the property where deferred maintenance stacks up. If any of your 26 are ones you'd actually be content to end up owning, drive them. I've seen structures go from occupied to gutted for copper inside one redemption period.

5.25% gross, minus your $90 to $110 of per-cert cost, minus the work of tracking redemptions and paying subs, and you're at what, three and a bit? I can get close to that without leaving my chair. If the answer is that you need bigger tickets or thinner counties to make it work, that's the strategy telling you something.

Slightly behind everyone here. When you say the rate gets bid down to 3%, does the amount you pay stay the same? I had assumed bidding was on price. So the winner still hands over the full $2,100 of back taxes and just accepts less interest when the owner redeems?

@ember I asked my county the same thing and yes, you pay the taxes owed, the bidding is only over the return. What I don't understand is the deposit. Mine wants 10% wired a week before and says unused funds go back in 15 business days. Bramble, does yours tie up cash on parcels you don't win?

Your fixed cost per certificate is the whole underwriting problem and you've already found it. At $95 of load on a $2,100 cert you need about 4.5% just to break even on a 14 month hold, and every month past 14 helps you rather than hurts you, which is a strange position to be in.

Two things I'd check before deciding anything. First, pull three years of results, not one. A single year of median 5.25% tells you nothing about whether a large bidder showed up once. If the same handful of registered bidder numbers win 60% of the improved lots every year, your pile A is not available to you at a price you like and no amount of parcel screening changes that. Second, get the county's redemption curve if they publish one, or ask what percentage of certificates from the 2022 and 2023 sales are still outstanding. Your entire yield assumption is the month 14 number and you've given no reason for it.

On the bigger tickets idea, it works arithmetically and it concentrates you. Four $12,000 certificates carry the same load as forty $1,200 ones and one of them going sideways is now a quarter of your book. That's a real trade, just make it deliberately.

Pile B I'd leave alone entirely. A 16% rate on a parcel you'd refuse to own is a rate on an asset you can't foreclose your way out of, and the sliver lots are sitting at 16% for exactly that reason.

@verity that's the argument I've used on myself for a year and it's why I still own nothing. Bramble has an actual results file and forty screened parcels, which is more than most people bring. Bid on six of the improved ones with a hard rate floor, see what actually clears and how long redemption takes, then decide whether the neighboring deed sales are worth the drive. The one thing the results file can't tell you is what it feels like to service a certificate through year two.