Ask ten experienced homebrewers which IBU formula to trust and you'll get close to ten different answers, several of them contradictory, most of them delivered with real conviction. The honest answer is that "which formula is right" is the wrong question to be asking in the first place — the actual lab measurement both formulas are trying to approximate sits behind a spectrophotometer, not behind either equation, and neither Tinseth nor Rager has any special claim to being closer to it in general. What the size of the gap between them tells you, on any given recipe, is genuinely useful information though, and that's worth understanding rather than picking a favourite and moving on.
The gap on a moderate beer
30 g of an 11% alpha hop, boiled the full 60 minutes, in a 19 L batch at a standard 1.050 boil gravity: 40.1 IBU by Tinseth, 53.5 by Rager — a gap of about 13 points, or roughly a third of the Tinseth figure. Check this yourself. That's a genuinely large-looking gap for what's a completely ordinary, moderate-strength beer, and it's the first thing that surprises brewers who've only ever used a tool defaulting to one formula.
The gap on a big beer
The same weight, alpha acid and boil time, but in a 1.090 boil-gravity imperial beer with a doubled hop weight to compensate (56 g instead of 30): 66.4 IBU by Tinseth, 106.0 by Rager — a gap approaching 40 points. Check this one too. Notice the gap didn't just grow proportionally with the bigger numbers — it grew disproportionately. On the moderate beer, Rager reads about 33% higher than Tinseth. On the big beer, Rager reads about 60% higher. That widening isn't noise; it's the direct, visible signature of the two formulas' gravity-correction methods being built completely differently.
Why gravity, specifically, is where they split hardest
Picture the two gravity-correction behaviours as two different shapes rather than two numbers. Tinseth's is a smooth downward slope starting the moment gravity rises past 1.000 — every additional gravity point shaves utilization down a little further, gradually, with no threshold anywhere. Rager's is closer to a flat line that suddenly bends downward once boil gravity crosses 1.050, staying essentially uncorrected below that mark and then correcting more and more aggressively above it. A recipe boiling at 1.045 sees almost no daylight between the two shapes — both are near their starting point. A recipe boiling at 1.090 is comparing a formula still gently sloping (Tinseth) against one that's been bending steeply downward for 40 gravity points already (Rager), and that structural difference, not any single coefficient, is what produces the widening gap.
Boil time is the other split, and it behaves differently
Shorten the moderate-beer example's boil time from 60 minutes to just 5, and Tinseth reads 8.0 IBU against Rager's 9.9 — a gap of under 2 points, both formulas agreeing that a very short boil contributes very little bitterness. Stretch it to 90 minutes instead, and Tinseth reads 42.9 against Rager's 55.4 — the gap actually shrinks slightly in proportional terms compared to the 60-minute version, because both formulas' boil-time curves are approaching their own asymptotes by 90 minutes and have less room left to diverge further. Boil time drives a real but comparatively modest and fairly stable disagreement; gravity is where the real divergence lives, and it keeps growing rather than levelling off.
Where each formula actually came from
Glenn Tinseth's formula first circulated online in the mid-1990s, built as an exponential-decay model of utilization rising toward a ceiling as boil time increases, with a continuous "bigness factor" correcting for gravity. Jackie Rager's formula, published in brewing literature around the same era, took a different mathematical shape entirely — a hyperbolic tangent curve for boil-time utilization, centred around roughly 30 minutes, with the step-function gravity correction described above layered on top. Neither author was working from the same underlying dataset or the same modelling assumptions as the other, and neither formula was ever positioned as a replacement for lab measurement — both are explicitly estimates, built to be "close enough" for recipe formulation without requiring a spectrophotometer in every home brewery.
Why software defaults vary, and why that's not a red flag
Different brewing software picked different defaults over the decades since, generally without much fanfare or a strong technical justification for choosing one over the other — it was frequently a case of whichever formula the original developer happened to implement first, or happened to find better documented at the time. This is exactly why running the identical recipe through BeerSmith, Brewer's Friend, a spreadsheet built from a magazine article, and this site can produce four different numbers even when every one of those tools is implementing its chosen formula correctly. See why do my IBUs differ between calculators for the fuller list of places tools diverge beyond formula choice alone — pellet-vs-cone assumptions, which gravity value gets used, and how hop-stand or whirlpool additions get modelled all stack on top of the base formula disagreement. None of this is a defect in either formula; it's simply a field that never converged on one shared standard the way, say, a temperature conversion did.
The BU:GU angle, worked through both formulas
Bittering units to gravity units (BU:GU — total IBU divided by gravity points) is often a more useful number than raw IBU for judging whether a beer will taste balanced, and it inherits the same Tinseth/Rager disagreement, scaled the same way. The moderate 1.050 example above works out to a BU:GU of 0.80 under Tinseth and 1.07 under Rager — a genuinely different-sounding balance verdict (0.80 reads as hop-forward-but-reasonable; 1.07 reads as aggressively bitter) depending purely on which formula you're trusting for the numerator. This is one more reason not to lean on either formula's output as a single, precise verdict — see what is a good bitterness-to-gravity ratio for how to read a BU:GU figure sensibly given that it carries the same formula-choice uncertainty IBU itself does.
So which one should you actually trust
Neither one in isolation, and that's not a dodge — it's the accurate description of what IBU estimation actually is. Both formulas are decades-old, widely reproduced, genuinely useful approximations of a real chemical process, built from different underlying assumptions about how utilization responds to time and gravity. The practical move isn't picking a winner; it's treating the spread between them as your honest uncertainty band. A moderate beer where both formulas land within 15-20% of each other is telling you the estimate is reasonably tight. A big, high-gravity beer where they're 50-60% apart is telling you, honestly, that the "true" IBU of that specific recipe is genuinely less certain — not because anyone's calculator is broken, but because bitterness estimation itself gets less reliable exactly where wort gets stronger and hop schedules get more aggressive.
Where your own tasting notes beat either formula
Repeat a style enough times, sticking to whichever formula you happen to check by habit, and a personal calibration builds up whether you set out to build one or not — a felt sense of what "this formula's 45" or "this formula's 60" actually delivers in the glass, which is worth more than chasing a formula switch in search of a more accurate abstract number. Your water, your yeast strain's own attenuation, and your specific hop-schedule habits all shape perceived bitterness in ways neither formula was ever built to model — see what is a good bitterness-to-gravity ratio for a related figure that tracks perceived balance more closely than a raw IBU number does by itself, and treat both formulas together as a range to plan within rather than searching either one for a single correct figure, especially on a big beer.