Deadlift variants: strength differentials across the major pulls

Trap-bar pulls run 8% heavier than straight-bar at the same lifter (Swinton 2011), sumo and conventional split on muscle recruitment without a 1RM advantage either direction (Escamilla 2000/2002), and RDL caps at 70-85% because it's hamstring/back-limited rather than floor-pull-limited. Default to conventional unless you've specifically chosen to compete sumo; trap-bar is the better default for general strength; RDL is an accessory; deficit fixes start-position weakness; paused fixes dead-bottom restart. Two of the differentials sit on peer-reviewed primary sources; the rest are practitioner consensus and labeled as such.

Most variant articles either lump every pull into a single table or pick one variant and pretend the others don't exist. This page is structured one variant per H2: a small comparison table, what the peer-reviewed literature actually says, and a one-line decision rule. The non-sourced differentials (RDL, deficit, paused) are flagged StrengthMath methodology in-line, not borrowed from JSCR / NSCA authority. For where these variant numbers land on the bodyweight-multiplier ladder, read deadlift standards by age and bodyweight; for the grip question that sits behind every heavy pull, see deadlift grip strategy.

Conventional vs sumo: not a 1RM gap — an EMG split

The conventional-versus-sumo question is asked as if one stance carries more weight. Two Escamilla papers say it doesn't. The first (Med Sci Sports Exerc 32(7):1265-75, 2000) is a three-dimensional biomechanical analysis of 24 national-level powerlifters competing in either style; the second (Med Sci Sports Exerc 34(4):682-8, 2002) is a 13-subject EMG study comparing the same stance pair. The two pulls differ on stance, joint angles, and muscle recruitment — but neither paper found a 1RM advantage either direction.

Escamilla 2000 (n=24, national-level powerlifters in competition, stance + joint kinematics) and Escamilla 2002 (n=13, collegiate football players, EMG comparison). The 25-40% shorter vertical bar distance is Escamilla 2000's headline mechanical-work finding; conventional pulls require correspondingly more mechanical work and energy expenditure for the same load.

The headline none of those numbers gives you: which one is heavier for the same lifter at the same depth. Escamilla 2000 sampled competing powerlifters who picked a style and trained that style exclusively, so the 1RMs in each style reflect lifters built for that style — not a within-subject comparison. The within-subject comparison in the strength-coaching literature lands on a 95-105% range either direction: lifters with hip-mobility and short femurs relative to torso typically sumo 0-5% above their conventional; lifters with longer femurs and tighter hip mobility typically conventional 0-5% above their sumo. The 95-105% range is StrengthMath methodology, not a published number; the underlying biomechanics (Escamilla 2000) are what justify the range being narrow.

Decision rule: default to conventional unless you have specifically chosen to compete sumo. Both stances are legal in every major powerlifting federation, but conventional is the more transferable lift for non-powerlifters — it teaches a hip-hinge from a narrow stance that maps cleanly to picking heavy things up off the ground. Sumo is the right style if your hip mobility / leverage profile favors it AND you compete; the shorter ROM (25-40% per Escamilla 2000) is a real efficiency advantage on competition day. The EMG split (sumo = quad-and-shin dominant, conventional = calf-dominant on the medial gastroc) is information about training transfer, not about which pull is mechanically “better.”

Trap-bar runs ~8% higher than straight-bar 1RM (Swinton 2011)

Swinton, Stewart, Agouris, Keogh, and Lloyd (J Strength Cond Res 25(7):2000-9, 2011) tested 19 male powerlifters across straight-bar and hexagonal-bar deadlifts at submaximal loads, then established 1RM in both. The headline number is the cleanest deadlift-variant differential in the peer-reviewed literature:

Swinton 2011, n=19 male powerlifters, within-subject comparison across both implements. The trap-bar geometry — handles at the lifter's sides rather than in front of the shins — shortens the moment arm at the hip and lumbar spine and lengthens it at the knee. The 8.2% absolute-load advantage and 11% peak-power advantage are consequences of that leverage shift.

Worked example: a lifter with a tested 200 kg straight-bar conventional 1RM will typically pull ~216 kg with a trap bar at the same training state (200 × 1.082 = 216.4 kg). The 8% advantage is robust at the group level; individual lifters land between roughly +5% and +12% based on how trained the trap-bar pattern is and how much the straight-bar 1RM is back-limited rather than leg-limited. A lifter whose straight-bar 1RM is capped by lumbar fatigue will see a bigger trap-bar advantage than a lifter whose straight-bar 1RM is capped by quad strength.

Decision rule: trap-bar is the better default for general strength; the lumbar peak-moment reduction is real, and the 1RM advantage doesn't come at the cost of training transfer. The exception is competitive powerlifters — competition is straight bar, and the 8% load advantage on the trap bar is exactly the kind of leverage gift that doesn't carry to meet day. For general-strength athletes, lifters with cranky lumbar spines, and field-sport athletes who pick up odd objects, trap-bar reads as the safer first heavy-pull pattern. The knee-moment increase is the tradeoff to watch — lifters with cranky knees may prefer straight bar for the same reason.

RDL: hamstring-dominant; caps at 70-85% of conventional

The Romanian deadlift is a top-down hinge. The bar starts at the rack (or at the lockout of a clean) and descends to mid-shin without touching the floor. The lifter never gets the bar to the floor and never gets to use a pull-from-floor leverage advantage. Lee, Schultz, Timgren, Staelgraeve, Miller, and Liu (J Exerc Sci Fit 16(3):87-93, 2018) compared conventional and Romanian deadlifts at 70% RDL 1RM and measured the joint-angle / EMG split:

Lee 2018, EMG comparison at 70% RDL 1RM. Conventional produced significantly higher rectus femoris and gluteus maximus activation; biceps femoris activation did not differ significantly between the lifts. The hamstring-dominance of the RDL is a joint-angle property (the knees stay nearly straight, the hip flexes deeply), not an absolute hamstring-EMG difference per Lee's measurements.

The RDL's 1RM is capped by hamstring length / lower-back tolerance / grip, not by floor-pull leverage. Most trained lifters land at 70-85% of their conventional 1RM in a strict RDL — a 200 kg conventional puller typically RDLs 140-170 kg. That range is StrengthMath methodology, synthesized from training-population practice across powerlifting and weightlifting coaching references. There is no peer-reviewed primary source that publishes a within-subject conventional-to-RDL 1RM ratio; Lee 2018 measured EMG, not absolute load capacity.

Decision rule: program RDL as a posterior-chain accessory, not as a primary strength lift. Its job is hamstring length and hip-hinge pattern at moderate load, not maximum absolute weight. Lifters who try to PR an RDL at 90%+ of their conventional are usually pulling with a rounded back and getting Lee's rectus-femoris recruitment because they've quietly turned the lift into a stiff-legged conventional deadlift. Keep RDL working sets at 70-80% of conventional 1RM for sets of 6-10 and the joint-angle posterior-chain stimulus comes through cleanly.

Deficit: 1-4 inches of extra ROM costs you 10-20%

A deficit deadlift is a conventional or sumo pull from a 1-4 inch elevated platform. The platform raises the lifter relative to the bar, which means the bar starts lower in the lift's coordinate-frame: deeper hip flexion, deeper knee flexion, longer range of motion. The cost is exactly where you'd predict — starting leverage gets worse, and the lift weighs less.

The 80-90% band reflects practitioner consensus across powerlifting coaching references. There is no peer-reviewed primary source that publishes a deficit-vs-floor 1RM differential; the band is StrengthMath methodology, not a sourced ratio. Sumo lifters typically take a smaller deficit hit than conventional lifters because the sumo starting position already absorbs some of the leverage cost.

Worked example: a lifter with a tested 200 kg conventional 1RM will typically deficit-pull 160-180 kgfrom a 2-inch platform — call it 175 kg as the working figure for a moderate deficit. A 4-inch deficit drops the working figure to ~160-170 kg. The lower number isn't the lift being “harder” in a general sense; it's the leverage tax for adding ROM at the worst possible spot in the pull.

Decision rule: program deficit pulls when your miss point is the floor break, not when you want a heavier-feeling lift. Lifters whose competition deadlift gets stuck off the floor — bar breaks ground, slows visibly in the first six inches, then either dies or rallies — benefit from deficit work. Lifters whose deadlift dies at lockout or mid-shin do not, and should program rack pulls or block pulls from the appropriate height instead. The deficit's value is starting-position-specific, not generic overload.

Paused: 5-15% off touch-and-go for the dead-bottom isometric

A paused deadlift holds the bar 2-3 inches off the floor for 1-3 seconds before continuing the concentric drive. The mechanical effects are exactly the paused-squat ones, applied to a pull instead of a press: any stretch reflex the bar had off the floor is gone (the brief isometric drains the elastic component), and the dead-bottom hold adds positional fatigue at the position where most deadlift misses happen.

The 85-95% band is StrengthMath methodology, synthesized from practitioner consensus across powerlifting coaching references. There is no peer-reviewed primary source publishing a paused-vs-touch-and-go deadlift differential. Most lifters lose more to the pause on a deadlift than on a squat because the bar truly resets between reps in a pause, while a paused squat is still the same continuous bar path.

Worked example: a 200 kg touch-and-go conventional 1RM, taken with a 2-second pause 2-3 inches off the floor, typically lands at 170-190 kg for the same lifter. A 1-second pause usually only costs 10-20 kg; the longer the pause, the bigger the cut, with diminishing returns past 3 seconds because by then the isometric is dominating. The lower paused number is the feature — the pause forces the lifter to drive concentrically out of the hardest position without the help of any rebound.

Decision rule: paused deadlifts off the floor address start-position weakness; paused deadlifts mid-shin or higher are not the right tool for a slow lockout. The pause height should match where the miss happens. If the competition deadlift dies at lockout, the right variant is a rack pull or block pull from above the knee, not a paused floor pull. The general rule across pause variants: pause the lift where the weakness lives.

Decision rule: which variant to default to

The variant ladder isn't about which deadlift is “the best.” Each variant trains a different limiter. Reading top to bottom by absolute-load potential at the same lifter, with conventional as the reference:

Sourced numbers: trap-bar +8.2% (Swinton 2011); sumo-vs-conventional stance and EMG split (Escamilla 2000/2002); RDL hamstring-dominance per joint angle (Lee 2018). All percentage differentials in the ladder are StrengthMath methodology— peer-reviewed biomechanics where the literature published the number, practitioner consensus where it didn't. Never blended.

The biggest mistake lifters make with this ladder is treating it as a hierarchy of difficulty. RDL isn't “harder than” conventional because the number is lower. The RDL number is lower because the lift caps on different tissue. A lifter who pulls 200 kg conventional and 130 kg RDL has a hamstring-length-limited RDL — the lower-body strength is closer to 200 than to 130, and the way to close the gap is mobility plus controlled tempo work, not by chasing a heavier RDL number with progressively rounder back. A lifter pulling 200 kg conventional and 170 kg RDL is closer to their hamstring/glute strength ceiling on both lifts.

For 1RM input from a submax-rep set in any of these variants, run it through the 1RM calculator first; the per-formula choice is covered in best 1RM formula (Epley for bench, any of the four for squat, any of the four +10% for deadlift). Deadlift-specific reps-to-1RM with the +10% adjustment lives at how to estimate deadlift 1RM from reps; for the band assignment your variant 1RM lands in, see deadlift standards by age and bodyweight.

Common questions

Is sumo deadlift easier than conventional?

Not at the 1RM. Escamilla 2000 (Med Sci Sports Exerc 32(7):1265-75) measured 24 national-level powerlifters and found sumo athletes pulled the bar over a 25-40% shorter vertical distance with a more upright torso, but the same lifters' competition 1RMs in either style sit in roughly the same range. Escamilla 2002 (Med Sci Sports Exerc 34(4):682-8) followed up with EMG: sumo recruited vastus medialis, vastus lateralis, and tibialis anterior more; conventional recruited medial gastrocnemius more. The two pulls split on muscle recruitment, not on 1RM. Whether sumo is 'easier' depends entirely on the lifter's hip-mobility and leverage profile, not on the lift itself.

How much more can I lift with a trap bar than a straight bar?

About 8% more, per Swinton 2011 (J Strength Cond Res 25(7):2000-9). Nineteen male powerlifters tested for 1RM in both implements averaged 265 ± 41 kg with the hexagonal (trap) bar versus 245 ± 39 kg with the straight bar — an 8.2% advantage at p < 0.05. Peak power was 11% higher (4,872 W vs 4,388 W). Peak moments at the lumbar spine, hip, and ankle were lower; peak knee moment was higher. The bar geometry shifts load from the lower back toward the quad, which is why trap-bar 1RMs run higher and why trap-bar pulling reads as more knee-dominant.

What percent of conventional deadlift is a Romanian deadlift?

Typically 70-85% for the same lifter, capped by hamstring length and lower-back tolerance, not floor-pull mechanics. The RDL is a top-down hinge — the bar starts at the rack or at the lockout, descends to mid-shin, and reverses without touching the floor. Lee 2018 (J Exerc Sci Fit 16(3):87-93) compared CD and RDL EMG at 70% RDL 1RM and found conventional produced significantly higher rectus femoris activation (58.6% vs 25.3% peak) and higher gluteus maximus activation (51.5% vs 46.9%). The RDL is hamstring-dominant by joint angle. The 70-85% range is StrengthMath methodology, synthesized from training-population practice; it is not a published peer-reviewed ratio.

Why does a deficit deadlift weigh less than a regular deadlift?

Because adding 1-4 inches of range of motion at the bottom is exactly where you have the worst leverage. Standing on a 1-4 inch platform raises the lifter relative to the bar, which increases the moment arm at the start position and forces the lift to begin from a deeper hip and knee angle. The cost is roughly 10-20% off your same-depth competition deadlift — a 200 kg conventional puller will typically deficit-pull 160-180 kg. The 80-90% range is StrengthMath methodology, not a peer-reviewed ratio. Deficit pulling is a starting-position weakness builder, not a higher-load substitute.

Which deadlift variant should a beginner start with?

Conventional or trap-bar. Conventional is the more transferable skill — it teaches a hip-hinge with the bar over mid-foot and applies cleanly to picking heavy things up off the ground. Trap-bar lowers the lower-back moment (Swinton 2011) and adds about 8% to the absolute load, which makes it the safer first heavy-pull pattern for lifters with cranky lumbar spines or for general-strength athletes who don't compete in powerlifting. Skip RDL until you can hinge cleanly with a barbell and skip deficit/paused work until you have a stable competition deadlift to build off. Sumo is a competitive style choice, not a beginner default.

Can I use a paused deadlift to fix a slow lockout?

No — pause it where the weakness lives. A paused-off-the-floor deadlift (typically 2-3 inches up, held 1-3 seconds) overloads the start position and the dead-bottom isometric, which is the right tool for lifters who miss low. For a slow lockout, the right variant is a block pull or rack pull from above the knee, not a paused floor pull. The paused floor pull costs 5-15% off your touch-and-go competition 1RM (StrengthMath methodology) — a 200 kg lifter pauses 170-190 kg. The lower number is the feature, not the bug; the pause removes whatever stretch reflex the bar has off the floor and forces a pure concentric restart.

Where to next

Once you have a variant-specific 1RM that fits the band you train in, the next decision is how that pull compares to your other lifts. The deadlift-to-bench ratio is the most-asked cross-lift question for pullers — see deadlift to bench ratio: what's typical for how variant choice changes the ratio (a trap-bar puller and a straight-bar puller at the same bench will read different deadlift-to-bench ratios at the same lifter, and the gap is exactly the trap-bar advantage covered above). For the grip strategy that sits underneath every heavy pull, deadlift grip strategy covers mixed vs double-overhand vs hook plus when to wear straps.