What Predicts RASP Success? Inside the Most Accurate Study on RASP Attrition

If you’re preparing for the U.S. Army’s Ranger Assessment & Selection Program (RASP), you’ve probably already heard of the arduous journey that awaits ahead. Some say you can achieve the tan beret with pure mental toughness. Most focus on crushing physical fitness standards.

Yet every cycle, the most motivated and sharp candidates arrive at the RASP only for over 50% to vanish within days. So, what do we do? How can we improve our odds? I dived deep into research to answer those very questions.

A landmark 2024 study on Predicting Ranger Attrition was published by military psychologists, answering these questions with unprecedented precision. Coombs et. al. analyzed data from 1,321 RASP candidates across 13 separate cohorts, building the most comprehensive predictive model ever created for special operations selection success.

In this article, I’ll walk you through a comprehensive breakdown of the top predictors of success, why week-one attrition matters most, the three predictors, and how you can outsmart the data. Let’s find out what science actually says.

1. Understanding the study, sample design, and methods

During this research, analysts examined a total of 1,321 candidates from the 75th Ranger Regiment who participated in the intensive 8-week RASP program over two consecutive years.

Spoiler: Results indicate that your performance on a few simple physical tests, combined with certain personality traits, can indicate a significant portion of your odds of completing the course. Even before you show up!

A. Sample Design

The study employed a two-sample cross-validation design to make sure they had separate training and validation cohorts.

B. Method of Analysis

There were two selection outcomes: “Success” (graduated RASP) or “Attrition” (failed to complete). And the results were coded in binary (1 for success and 0 for attrition) with all recorded reasons for course failure counted as attrition due to concerns about the reliability of these records.

Two control variables were included:

• Age
• Time on ground (7-189 days in the 75th Ranger Regiment before starting RASP)

You can refer to the study’s Table 3 for a detailed breakdown of the descriptive statistics for controls and predictors, as well as training and validation samples.

Data from both sample sets were collected based on the following predictors:

• Physical Predictors: Push-ups (2 min), sit-ups (2 min), and 2-mile run. All scores were calculated using the male 22-26 age scale for consistency and  Army Physical Fitness Test (APFT) scores were used.
• Cognitive Predictors: Verbal IQ and Performance IQ from the Multidimensional Aptitude Battery-II (MAB-II).
• Personality Predictors: Measured with the Jackson Personality Inventory-Revised (JPI-R)
  • Composite scores were scored on Conscientiousness (organization, energy level, traditional values) and Openness (complexity, breadth of interest, tolerance).
  • Individual facets such as responsibility, sociability, social confidence, and cooperativeness also played a significant role.

* While demographic data was unavailable, historically, candidates have predominantly been white males.

The research methodology warranted that three predictive models be created and studied separately.

• Week 1 – Survival
• Weeks 2-8 – Completion
• Full 8-week – Graduation

C. Early vs Late Attrition

With the framework set, the study first addressed a critical question: Does it matter when, in the course, someone fails? The answer is yes.

Table 2. Breakdown of recorded reasons for attrition by each time frame, training sample.

The decision to build a separate predictive model for week-one attrition was a fruitful endeavor. Researchers found that week one is a beast of its own.

Around 77% of week-one attrition was voluntary withdrawal, often linked to mindset and expectations. Later in the course, attrition was far more likely to be due to physical test failures or medical issues. In conclusion, survival in week one is largely dependent on your mindset and grit.

2. Predictors Of Success: Physical, Cognitive, Personality

Let’s start with a quick overview of individual predictors and their impact on each of our three models using bivariate correlations. Do note that these are “one‑at‑a‑time” relationships that may disappear once predictors are tested together in the regressions.

Table 5. Significant individual predictors (point-biserial correlations) of candidate success in at least one timeframe, training sample. | *p < .05, **p < .01

Note: r_pb = point‑biserial correlation measures the strength and direction of the relationship between a continuous score (like push‑ups) and a binary outcome (pass/fail).

A. Physical Abilities

Physical fitness is the strongest predictor of attrition across all time frames. While unsurprising, it was an interesting finding that physical fitness held such a high predictive score even though RASP candidates are already among the top in terms of physical capabilities.

The 2-mile run emerged as the single most consistent predictor, while push-ups showed their greatest impact during the grueling middle weeks when upper-body endurance became critical.

B. Personality

Contrary to the conventional belief, the study found that personality traits explained more variance in success than cognitive ability. This doesn’t mean intelligence doesn’t matter; it highlights that the right personality traits are more important.

• Conscientiousness was the strongest personality predictor for overall and Weeks 2-8 success.
• Openness was a risk factor in the “Week 1” model. Higher scores here were linked to more early voluntary dropouts.

Other traits worth noting are Responsibility and Sociability.

C. Cognitive Ability

Performance IQ (spatial reasoning) demonstrated moderate predictive power, but only after the first week. Verbal IQ had minimal impact throughout the course.

This suggests that once you have survived the initial shock period, spatial intelligence aids navigation and tactical problem-solving, but it’s not the deciding factor many assumed it to be.

3. Timeline Analysis

Since no candidate is one-dimensional, researchers ran logistic regression models for the three timeframes to better understand how the three abilities interact. The results were reported in standardized odds ratios (OR), which show how much the odds of success change when a predictor increases by 1 standard deviation.

Week 1: Psychological Survival

The first week is when we typically see the highest attrition, with almost all of it being voluntary withdrawal.

Table 6. Model 2. Prediction of Attrition during Week One * p<.05, ** p<.01.

Inverse range ORs are reported for negative regression weights for a comfortable comparison with predictors that have positive regression weights. Physical fitness still matters a great deal, of course; however, personality signals are increasingly prominent early on.

The researchers noted that high openness increases voluntary withdrawal risk, and responsibility helps you push through the initial discomfort and uncertainty. Cooperativeness and social confidence also appear as meaningful Week 1 protectors.

Weeks 2-8: Completion

Next, the program transitions into sustained field operations and progressively more stringent standards. This is where critical pass/fail events happen.

Table 6. Model 3. Prediction of Attrition between Weeks Two and Eight | * p<.05, ** p<.01.

Inverse range ORs are reported for negative regression weights for a comfortable comparison with predictors that have positive regression weights.

Push-ups (1.63 STD OR), 2-mile run (1.41 STD OR), and sit-ups (1.23 STD OR) understandably maintained their top spots. Conscientiousness (1.33 STD OR) is important because it provides key protection against cumulative fatigue, helping you stay organized, consistent, and compliant day after day.

Finally, Performance IQ (1.23 STD OR) starts to matter, likely due to increased tactical complexity and problem-solving demands under stress.

Physical standards and land navigation failures account for 26.6% and 11% of attrition, respectively, during this window. The phase also experiences a steady stream of voluntary quits (~40%).

Full Course: The Complete Picture

Across the whole 8-week program, fitness keeps the heaviest weighting, with push-ups and running dominating across all models.

Table 6. Model 1. Prediction of Overall Eight-Week Attrition | * p<.05, ** p<.01.

Inverse range ORs are reported for negative regression weights for a comfortable comparison with predictors that have positive regression weights.

When each family of predictors was tested alone on the validation sample, the results demonstrated that physical ability outperforms both personality and cognitive ability. The scores were as follows:

• Physical r_pb = .322
• Personality r_pb = .124
• Cognitive r_pb = .098

4. Interpretation of Scores

Now that the predictive models are ready, researchers tested their accuracy using three key measures:

A. Key Metrics

• AUC (Area Under the Curve): Indicates how effectively the model distinguishes between graduates and dropouts.
• F1 Score: Balances precision (how often the “pass” prediction was correct) with recall (how many actual passes the model caught).
• MCC (Matthews Correlation Coefficient): An overall quality score that works well even if pass/fail numbers aren’t balanced.

B. Candidate Probability Score

Using the final full-course model, the researchers created a Candidate Probability Score (CPS). It is a number between 0 and 1 that estimates each candidate’s probability of graduating from RASP.

Table 7. A sample of candidate probability scores (CPS), validation sample
[PU = Push-ups, SU = Sit-ups, Perf_IQ = Performance IQ, Op = Openness, Co = Conscientiousness, So = Sociability]

When candidates in the training sample were split into quintiles based on their CPS, their pass rates showed a clear performance gap:

• Top quintile: 77% pass
• Lowest quintile: 29% pass

Even before day one, the CPS could identify groups with sharply different probabilities of success.

The CPS is compensatory, meaning strength in one predictor can offset weakness in another. In the study’s examples, two candidates each had a CPS of 0.48 (a 48% graduation probability), but achieved it through very different combinations of physical, personality, and cognitive scores.

This indicates that there is no single profile for success; different strengths can yield the same statistical outcome.

C. Uniqueness of this study:

The researchers trained their model on the 2019 classes and tested it on 2020 candidates. The goal was to determine if the same predictors would work just as well a year later with a different group.

Here’s what they found:

• Brier score (measures average prediction error) went from 0.20 to 0.23. Lower is better, and this is only a tiny increase. The model stayed nearly as accurate in year two.
• Point-biserial correlation (measures how well predictions match actual outcomes) dipped from 0.41 to 0.35. That’s still a solid relationship, especially when applied to a completely new set of candidates.

This means about 85% of predictive strength carried over into a different year’s population. It’s evident now that the traits linked to success (physical performance, conscientiousness, and certain personality factors) remained stable across years. 

5. Research Limitations

While the study was largely successful in providing the clearest prep roadmap (as of 2025), it has several constraints that you must be aware of:

1. Personality scores were self‑reported.
2. Physical and mental qualities were pre‑screened prior to RASP entry.
3. Analyses were based on archival, scale-level data, instead of specific moment-to-moment attrition decisions.
4. Results are specific to RASP’s combined training‑and‑selection format only. Not other SOF pipelines.

6. Tips For Aspiring Rangers

After reviewing the paper in depth and based on my personal experience with coaching SOF trainees for the Ranger Assessment and Selection Program (RASP), I have concluded a few points:

• Physical fitness remains the ultimate gatekeeper. Remember, you’ll be competing against the best of the best.
• Train yourself to expect monotony.
• Practice land navigation, map reading, and spatial reasoning tasks under physical load and time pressure, to improve Performance IQ.
• Practice working effectively under stress while maintaining attention to detail, and develop organizational systems that function efficiently under pressure.
• Success can be achieved through various strength combinations, so train according to your individual strengths.

Your best preparation strategy for RASP is to combine elite-level physical conditioning with deliberate development of key personality traits. While you can’t completely change your personality, you can build habits and systems that reflect the traits of successful Rangers.

If you focus and train for what really matters, you’re stacking the odds in your favour before you even step off the bus. If you want to optimize your selection prep further, you can reach out for 1 on 1 coaching with me.

Source: Predicting ranger attrition by Aaron K. Coombs and Neil M.A. Hauenstein (MILITARY PSYCHOLOGY, 2025, VOL. 37, NO. 1, 73–84, link)

7. Other Important Studies

Table 1. Overview of military attrition studies- Elite unit selection and training.