Deadlift & Romanian Deadlift Tempo: Evidence-Based Prescriptions

Tempo prescription on the deadlift family is fundamentally different for the conventional deadlift than for the Romanian deadlift. The same four-digit notation reads differently on each lift, and the published evidence supports nearly opposite tempo strategies. This page summarizes what the research says, where the evidence runs out, and what advanced lifters should and should not prescribe.

• Conventional deadlift (CDL): maximal-intent concentric paired with reset or controlled (not slow) eccentric. Cumulative spinal load, lumbar flexion creep through fatigue, and grip become rate-limiting before muscular fatigue does. Slow eccentrics on heavy CDL extend exactly the wrong variable.
• Romanian deadlift (RDL): slow eccentric (4-0-1-0 or 5-0-1-0) is the precise mechanism the lift exists for. Maeo et al. (2024) and Kassiano et al. (2025) confirm hamstring hypertrophy responds preferentially to slow eccentrics in the lengthened position, and the RDL is built around that position.
• Critical evidence gap: zero peer-reviewed RCTs directly compare tempo conditions on the conventional deadlift or the Romanian deadlift. Every prescription on this page is an extrapolation from the squat/bench tempo literature, deadlift biomechanics studies, and coach convention. We flag this honestly throughout.

Why the conventional deadlift and the RDL need different tempo strategies

The conventional deadlift is not a back-loaded squat with the bar in your hands. It is a fundamentally different mechanical event from the Romanian deadlift, even though they share notation. Tempo prescription that treats them the same misses what the research has been showing for decades.

Cholewicki, McGill and Norman (1991) studied 13 female and 44 male competitors at the Canadian Powerlifting Championships and estimated peak L4/L5 compressive forces of up to 17,192 N at competition loads, with averages of 14,487 N at 259 kg loads in the 90-kg male class. Granhed, Jonson and Hansson (1987) reported even higher peaks of up to 36,400 N at the World Powerlifting Championships, though their model is generally considered to overestimate. Either way, heavy conventional deadlifts routinely exceed cadaver injury thresholds for compression. The relevant variable for tempo prescription is not just peak force but cumulative load-time integral, impulse. Slow eccentrics increase impulse without reducing peak force.

Boocock et al. (2019) showed that during repetitive lifting to failure, percent lumbosacral flexion increased from 71.7% on the first minute to 98.4% on the final minute. The MDPI Applied Sciences 2024 study on intra-set biomechanics during 3RM deadlifts in strength-trained women, the first study to measure within-strength-rep-range fatigue kinematics, reported that lower thoracic and upper lumbar flexion angles increased measurably from rep one to rep three even at strength-spec loading. Korakakis et al. (2024) added that repetition-to-failure deadlifting increases variability of lumbo-pelvic coordination. These findings converge on a single conclusion: every additional second under load on a heavy CDL increases the probability that the spine creeps further into flexion. That is the strongest evidence-based argument against five-second-plus eccentrics on the conventional deadlift.

The RDL escapes this for three reasons. The bar never returns to the floor, so the spine does not enter deep flexion. The lift typically stops at mid-shin. And RDL loads run roughly 60-80% of CDL loads in coaches' reports, so cumulative spinal stress per second under load is meaningfully lower. Slow eccentrics on the RDL train a productive position; slow eccentrics on the CDL extend exposure to a deteriorating position.

Lee et al. (2018) compared the two lifts directly at 70% of RDL 1RM in 21 trained men. The conventional deadlift produced significantly greater rectus femoris and gluteus maximus EMG and net joint torque; the RDL biased the work toward biceps femoris and gave hamstrings longer time under tension. This corrects a popular myth: at matched relative load, the CDL hits the gluteus maximus more than the RDL does. The RDL's superiority is in the hamstring, not the glute. Martín-Fuentes, Oliva-Lozano and Muyor (2020) systematic review confirms the broader pattern, erectors and quadriceps drive the deadlift family except in the RDL, where biceps femoris and semitendinosus exceed erector spinae activation. Slow eccentric CDL therefore extends erector and quad TUT (and adds spinal flexion risk), not hamstring TUT.

Tempo notation on this page uses the four-digit eccentric–pause–concentric–pause convention Charles Poliquin codified, extending Ian King's earlier three-digit system. The reading order is fixed across all variations; for the conceptual background see the four-digit tempo notation guide.

Conventional deadlift tempo prescriptions

The CDL prescriptions below are biased toward concentric intent paired with reset or controlled (not slow) eccentrics. Confidence reflects how directly the evidence applies, high when an RCT or biomechanics study supports the prescribed conditions, moderate when extrapolated from related lifts, low when the row rests on coaching convention rather than peer-reviewed data.

The maximal-strength row reflects competition specificity. A 1-0-X-0 or X-0-X-0 reset deadlift is what the lifter does on the platform. Cholewicki's spinal force data argues for minimizing time under peak load, and the MDPI 2024 intra-set kinematics study shows that even at strength-spec rep ranges, lumbar flexion is creeping by the third rep. There is no productive role for a slow eccentric here, all it adds is impulse and fatigue.

The strength row at 3-5 reps shifts to a 2-second eccentric or a controlled touchdown. Wilk, Zajác and Tufano (2021) are explicit on the underlying point: “Deliberate manipulation of movement tempo may not be possible when exercises are performed with heavy loads.” At 80%+ 1RM the eccentric will be controlled because it has to be, the lifter doesn't have to write “slow” into the prescription. What matters is that the bar is not dropped from lockout, and the lifter intends to lift the next rep with maximal acceleration.

The hypertrophy row uses controlled touch-and-go in the 5-8 rep band. Schoenfeld, Ogborn and Krieger's (2015) meta-analysis caps the equivalence band at roughly 8 seconds per rep when sets approach failure, a 2-0-X-0 touch-and-go fits inside that window. A 3-0-X-0 reset works too, with the trade-off that reset cuts the SSC contribution and reduces reps achievable at a given load. For broader context on controlled eccentric work in heavy compound lifts, see the eccentric training guide.

Posterior-chain hypertrophy is the rare row where the prescription is to switch lifts entirely. Lee et al. (2018) is unambiguous: the RDL hits hamstrings more selectively at safer load ranges. Adding more eccentric duration to the CDL does not redirect the work to the hamstrings, it extends the erector and quadriceps TUT and exposes the spine to more flexion-under-load time. The honest answer is that the CDL is the wrong tool for hamstring hypertrophy and a slow CDL eccentric is a worse version of the wrong tool.

Pause variants on the CDL

Pause variants are a different question from straight slow-eccentric prescriptions. A pause-below-knee CDL at 1-2-X-0 uses a one-second eccentric to a pause height (usually mid-shin), a two-second isometric pause to train start-strength and lat tightness, and a maximal-intent concentric. A pause-at-knee CDL places the pause through the sticking-region transition. A deficit CDL at 1-0-X-0 reset increases the start-position range of motion at conservative load. None of these are slow-eccentric prescriptions, they target specific phases of the lift. Snatch-grip and stiff-leg variants follow the same principle: the pause or position, not the eccentric duration, is what's being trained.

Romanian deadlift tempo prescriptions

The RDL is where slow eccentrics actually work. The lift loads the hamstring in the lengthened position throughout the descent, the spine does not pass through deep flexion, and loads run conservatively enough that grip and lumbar fatigue are not the limiters. The hypertrophy literature converges directly on this lift.

The hypertrophy primary row is the strongest evidence-anchored prescription on the page. Maeo et al. (2024) compared lengthened-state eccentric training (LSET) against the Nordic hamstring exercise and found LSET produced greater overall hamstring hypertrophy, particularly biceps femoris long head and semimembranosus. Their earlier (2021) study showed long-muscle-length training produced greater hamstring growth than short-muscle-length training. Kassiano et al. (2025) directly compared a six-week RDL versus Nordic hamstring intervention and found both produced hypertrophy, with the RDL producing greater biceps femoris long head growth. Bourne et al. (2018) adds that lengthened-state and eccentric hamstring loading reduces strain injury risk and produces architectural changes such as longer fascicles. The 4-0-1-0 prescription matches the eccentric duration that has been validated in this literature.

The stretch-emphasis row at 3-2-X-0 exploits a two-second pause at the deepest hamstring stretch. Tom Hibbert reproduces this from a Poliquin snatch-grip RDL camp protocol. The mechanism is direct: a 2-second isometric hold at long muscle length doubles down on the lengthened-state stimulus that Maeo's work documents, and forces the lifter to break out of the bottom without elastic recoil. Slow eccentric loading in the lengthened position is the broader principle this row instantiates.

The TUT-max row reflects Pearson et al. (2022), who tested varying repetition tempo in single-joint lower-body work and found tempo variation did not augment outcomes versus moderate tempo. The 4-1-2-1 prescription (4 s eccentric, 1 s pause, 2 s concentric, 1 s top pause) sums to 8 seconds, right at Schoenfeld's documented hypertrophy ceiling, not above it. Useful as a variation block, not as a permanent default. For the convention on what each digit represents in execution order, see tempo notation explained.

The single-leg RDL row caps the eccentric at three seconds. This is a coaching convention rather than RCT-supported, but the rationale is mechanical: balance fails before the hamstring does. A 5-second eccentric on a single-leg RDL is essentially impossible to execute cleanly without compensatory hip rotation or knee bend. Three seconds is a defensible upper limit; longer eccentrics on the bilateral RDL with the same intent.

Touch-and-go vs reset vs full eccentric control: the deadlift tempo controversy

This is the question that gets the most heat in lifting commentary and the least light from the research. The honest synthesis: the popular positions are confidently stated, the evidence base is weak, and a clean decision rule based on goal is the only intellectually defensible output.

Mechanically, four variants are distinguishable:

The honest disclosure on evidence: there is no peer-reviewed RCT directly comparing dead-stop versus touch-and-go versus controlled-eccentric deadlift training on hypertrophy or 1RM outcomes. Every confident claim in the lifting blogosphere is an extrapolation from related lifts, biomechanics, or coaching tradition. State this explicitly to the reader; do not paper it over.

What is established by the evidence: touch-and-go captures elastic recoil. Tsoukos and Bogdanis (2025) measured flywheel deadlifts and found the first repetition (no SSC, like a reset) had 25-30% lower mean concentric velocity and 7-11% lower mean force compared to subsequent reps. Reset eliminates this contribution and trains start-strength as a discrete quality. Dropping the bar forfeits the eccentric stimulus entirely, Roig et al. (2009) and Schoenfeld's (2017) contraction-type meta-analysis converge on the conclusion that eccentric actions produce equal or greater hypertrophy than concentric-only at matched volume. And slow eccentrics on heavy CDL are explicitly contraindicated by Wilk, Zajác and Tufano (2021): heavy loads cannot be deliberately slowed without reducing the load.

The decision rule that follows from this evidence base, applied honestly:

• Strength specificity (1RM, competition prep) → reset, low rep, maximal concentric intent. Reset matches the platform.
• Hypertrophy on the conventional deadlift → touch-and-go with a controlled (never bounced) eccentric, 5-8 reps at 70-80% 1RM.
• Posterior-chain hypertrophy specifically → switch to RDL with a slow eccentric. Stop trying to redirect work in the CDL.
• Eccentric stimulus on the deadlift family → use the RDL or deficit RDL. Slow CDL eccentrics extend exactly the wrong variable.

Time under tension and the Schoenfeld ceiling, applied to the deadlift

Schoenfeld, Ogborn and Krieger's (2015) systematic review and meta-analysis remains the upper-bound reference for tempo claims. Repetition durations between 0.5 and 8 seconds produced statistically equivalent hypertrophy when sets approached failure. Within that window, faster, slower, and intermediate cadences produced similar muscle growth. Effort, not cadence, drives the response.

Wilk, Zajác and Tufano (2021) made the statement that anchors this section explicitly: “Deliberate manipulation of movement tempo may not be possible when exercises are performed with heavy loads.” That is the strongest current review of tempo research telling readers, in plain words, that the heavy-CDL low-rep range sits outside the evidence base for tempo prescription. You cannot deliberately slow a 90% 1RM deadlift, you can only fail to lift it. Practical takeaway: stop trying to prescribe a slow eccentric in that range. Use reset and concentric intent instead.

Maeo et al. (2024) provides the specific mechanism that makes slow RDL eccentrics work where slow CDL eccentrics fail. Lengthened-state hypertrophy, not generic time-under-tension, is the active variable. The hamstring grows preferentially when loaded at long muscle lengths under controlled descent, that is what a 4-0-1-0 RDL trains, and it is what a 4-0-1-0 CDL does not train (because the hamstring is not the limiting muscle in the CDL).

Kojic, Mandic and Duric (2025) provides the closest squat-specific data point: a 7-week RCT in untrained subjects compared 4-0-1-0 against 1-0-1-0 on the back squat and found the slow-eccentric group produced greater 1RM gains (effect size 1.60 vs 0.99) and greater vastus lateralis hypertrophy (1.74 vs 1.37). This is one squat-specific study in untrained subjects. It is suggestive evidence that controlled eccentrics matter on closed-chain compound lower-body lifts. It does not generalize cleanly to the conventional deadlift, which has a dead-stop start, no top-loaded back position, and progressive lumbar flexion under fatigue. Cite it as suggestive; refuse to overstate it. The broader TUT and slow eccentric research covers this in more detail.

Velocity-based interpretation of “X” on the deadlift

The “X” in deadlift tempo notation often confuses lifters because at heavy loads the bar visibly moves slowly. The deadlift load-velocity relationship is also noisier than the squat or bench press, Lake, Naworynsky, Duncan and Jackson (2017) tested predicted versus actual deadlift 1RM from individualized load-velocity profiles and found predicted 1RM was significantly less than actual (effect sizes between 1.03 and 1.75). The deadlift's dead-stop start, grip variability, and fatigue patterns make it less amenable to the precise velocity-based percentage targets that work on bench press.

Lopes dos Santos et al. (2023) tested the hexagonal-bar deadlift and found mean concentric velocity predicted 1RM with R² = 0.85, a slightly cleaner relationship than the conventional deadlift's. Pérez-Castilla et al. (2020) confirmed the load-velocity relationship is reliable within an individual (R² greater than 0.95) but variable across individuals (around 20% coefficient of variation). The minimum velocity threshold at conventional deadlift 1RM is approximately 0.15-0.20 m/s for males.

The conceptual point is the same as on the front squat: X is intent, not realized velocity. At 90% 1RM the bar will move at strength-velocity speeds regardless of intent, but the neural drive that produces the effort is what trains the strength and power adaptation. A 1-0-X-0 reset CDL at 90% with maximum acceleration intent trains the same quality whether the bar moves at 0.30 m/s or 0.45 m/s. The lifter cannot make the bar faster by trying harder past a certain load. The lifter can fail to try hard, and that's the variable the X protects against.

Common errors specific to the deadlift family

Six prescription errors recur in commercial fitness writing about deadlift tempo. Each is grounded in a specific evidence-based limit.

Slow eccentric on heavy CDL

The dominant error. Above roughly 75% 1RM, slow eccentrics extend exposure to lumbar flexion creep (Boocock 2019; MDPI 2024) and grip becomes the limiting factor before the posterior chain does. Wilk-Zajác-Tufano 2021 is explicit that heavy loads cannot be deliberately slowed. Use reset or controlled-touchdown instead.

Bar drop with no eccentric

Forfeits the eccentric stimulus that produces equal or greater hypertrophy than concentric-only work at matched volume (Roig 2009; Schoenfeld 2017). Reserve the bar drop for Olympic-weightlifting contexts where the catch position makes controlled lowering dangerous. Not a general strength-training prescription.

Inconsistent touch-and-go (real bouncing)

Uncontrolled bouncing uses ground-reaction force from the plates rebounding, not the stretch reflex. This degrades the training stimulus and increases lumbar load if the lifter loses position in transit. Touch-and-go must be controlled, lifter maintains tension, bar contacts briefly, lifter does not lose neutral spine.

Resetting without releasing tension

Many lifters “reset” only the bar position, not their setup. Functionally this is a poorly executed touch-and-go with worse rhythm. A real reset releases the breath, breaks tension, and re-establishes the start position. If the lifter is not doing that, they're not getting the start-strength training the reset is supposed to deliver.

Long bottom pause on RDL with rounded back

The bottom-pause RDL is high-yield only if the spine remains neutral. A 3-2-X-0 RDL with a hyperflexed lumbar spine trains flexion under load, the opposite of the lengthened-state hamstring stimulus the prescription is supposed to deliver, and a known mechanism for posterior ligamentous strain.

Treating RDL like a CDL

The bar starts at lockout, never returns to the floor, knees stay nearly fixed throughout. If the lifter re-bends the knees and the bar touches the floor, the lift becomes a stiff-leg deadlift, a different exercise with different mechanics. Tempo prescription must specify which lift; the same notation produces different training outcomes on RDL versus stiff-leg deadlift versus CDL.

Frequently asked questions

What's the best tempo for deadlift?

There isn't one, the conventional deadlift and Romanian deadlift need different tempo strategies. CDL favors maximal concentric intent with a controlled or reset eccentric because cumulative spinal load and grip become limiting before the muscles do (Wilk, Zajác & Tufano 2021). RDL is where slow eccentrics work, 4-0-1-0 or 5-0-1-0, because the lift loads the hamstring in the lengthened position where stretch-mediated hypertrophy is strongest (Maeo et al. 2024).

Should I use touch-and-go or reset deadlifts?

No peer-reviewed RCT has directly compared the two on hypertrophy or 1RM. Touch-and-go captures elastic recoil and typically permits 5-10% more weight or 1-3 more reps at the same load (Tsoukos & Bogdanis 2025 quantified the first-rep velocity deficit at 25-30%). Reset eliminates the stretch-shortening contribution and trains start-strength specifically. Choose by goal: strength specificity → reset; hypertrophy density → touch-and-go controlled, never bounced.

Is Romanian deadlift safer than conventional deadlift for slow eccentrics?

Yes, for three biomechanical reasons. The RDL never returns to the floor and the spine doesn't enter deep flexion. Loads typically run 60-80% of the CDL 1RM. And the working-muscle bias is hamstring-led rather than erector-led (Lee 2018; Martín-Fuentes 2020). The lumbar flexion creep that Boocock et al. (2019) documented during heavy CDL fatigue (71.7% to 98.4% lumbosacral flexion) does not apply to RDL at typical loads.

What tempo did Charles Poliquin recommend for the deadlift?

Poliquin's documented 4-0-1-0 was the structural-balance test condition, not a training prescription. For RDL specifically, his archived programs show 2-0-2-1 (Infinite Labs reproduction), 3-2-1-0 (Hibbert reproduction of his snatch-grip RDL camp protocol), and structural-balance 4-0-1-0 testing. For the conventional deadlift Poliquin published very little formal tempo prescription, his framework treated CDL as a power lift and shifted hypertrophy work to the RDL family.

How much weight should I use for tempo deadlifts?

For controlled touch-and-go CDL hypertrophy work, 70-80% 1RM is the typical productive range (Schoenfeld 2015 puts the relevant rep duration in the 0.5-8 second window). For RDL with 4-0-1-0, the typical RDL 1RM is reported by experienced coaches at 60-80% of the CDL 1RM (no normative dataset; phrase it as coach convention). Start at 50% of your CDL 1RM on RDL and adjust upward by execution quality, not by the calculator.

Can I build muscle with reset deadlifts only?

Yes. No RCT has shown touch-and-go superior to reset for hypertrophy. Both produce eccentric stimulus, controlled lowering on reset, touch-and-go mid-rep on continuous sets, both pull through the same range of motion, both load the same muscles. Reset trains start-strength specifically and matches powerlifting competition specificity. Touch-and-go simply permits higher density. Repko's tempo timer enforces all four phases without requiring you to count in your head. The deadlift family, especially RDL with 4-0-1-0 or 5-0-1-0, is one of the lifts where this matters most. Try Repko free.

The honest gap: what we don't know

The deadlift family is the most under-studied of the big three for tempo manipulation. Zero peer-reviewed RCTs directly compare tempo conditions on the conventional deadlift. Zero direct comparisons on the Romanian deadlift either. The Krzysztofik and Wilk Katowice group has published multiple tempo papers on bench press and back squat, not the deadlift.

What this means in practice: every prescription on this page is an extrapolation. Schoenfeld's 2015 meta-analysis and Wilk-Zajác-Tufano's 2021 review establish the general tempo-and-hypertrophy framework. Cholewicki's 1991 spinal force data and Boocock's 2019 lumbar flexion creep findings establish why heavy CDL behaves differently from squat or bench under tempo manipulation. Maeo's 2024 lengthened-state work and Kassiano's 2025 RDL-specific data establish the mechanism by which slow RDL eccentrics produce hamstring hypertrophy. The numbers in this page's tables, 3-second versus 4-second versus 5-second eccentrics, 70% versus 80% 1RM, sit inside ranges that the evidence supports, but their precise selection is coaching judgment, not RCT-derived precision.

The strongest evidence-anchored claim on this page is the direction of the prescription bias: different tempo strategies for CDL versus RDL. The specific tempo numbers are best practice within evidence-supported ranges, not RCT-validated optima. Honest content acknowledges this; commercial “ten best deadlift tempos” content does not. Repko's position on the deadlift family is the conservative, evidence-anchored one: prescribe by lift, by goal, and by load tolerance, and refuse to overstate what the literature actually shows.

Closing

This page is part of Repko's tempo training cluster: the four-phase tempo guide covers the framework, eccentric training covers the slow-lowering literature, tempo notation explained covers the language, and the sibling exercise deep-dives on front squat tempo prescriptions for the squat side, lat pulldown and pull-up tempo prescriptions for the open-chain pull pattern, bar muscle-up and negative tempo prescriptions for closed-chain advanced calisthenics, and side plank and McGill Big 3 tempo prescriptions for static-isometric spine endurance extend the tier-3 cluster. For background on why this app exists and who built it, see the about page. Repko's tempo timer enforces every digit of the prescription so you can train without counting in your head.