PRACTICE GUIDE JANE STREET
Jane Street Stochastic Processes Practice Test
Jane Street does not publish the shape of its stochastic processes screen, and candidates describe it as 10 to 15 questions in 30 to 60 minutes, roughly two to five minutes per question, answered by picking one of the options.
Those are ranges across firms running this format rather than Jane Street's own figures, so treat them as the shape to train against and check the instruction screen on the day for the marking rule.
Outcry is not affiliated with Jane Street and has no access to their assessment content. This guide describes an assessment format that candidates report publicly; the questions here are generated by Outcry and are not Jane Street’s own.
What it screens
A quantitative trading firm active across equities, ETFs, bonds, options and currencies, known for OCaml and a strong internship pipeline.
- ✓Martingale arguments and optional stopping applied to games
- ✓Gambler's ruin: absorption probabilities and expected duration
- ✓Recognising when a process is a martingale (and when the argument breaks)
- ✓Random-walk intuition: recurrence, drift, and scaling
Where it sits at Jane Street
A quantitative trading firm active across equities, ETFs, bonds, options and currencies, known for OCaml and a strong internship pipeline. What that means for the screen is that the questions tend to be drawn from the work rather than from a textbook, so the format below is the shape to train against rather than a syllabus.
Candidates also report 3 other screens at Jane Street, covered separately on this site. Where a firm runs several, they usually sit in one round rather than spread across the process, so the pacing of the whole set matters more than any single section.
The format
Jane Street does not publish this screen's shape, and it varies between firms, so these are the ranges candidates report rather than exact figures.
| Questions | 10 to 15 |
|---|---|
| Time | 30 to 60 minutes |
| Per question | two to five minutes |
| Negative marking | No |
| Answer style | Multiple choice or short typed answer |
| Where it sits | Research-track screen, before the technical interviews |
What it tests, with a worked example
Every example below is generated by Outcry, drawn from the same question generators the timed drills run. None of them is Jane Street’s.
Inference and research integrity
What a result means, and what it would take for it to mean nothing.
Example
A test of a strategy's edge runs at the 5% level and has 70% power against the effect size that matters. Which statement follows?
- A false positive turns up 30% of the time
- A real effect of that size is missed 30% of the time
- Loosening the level to 10% would lower the power
- 70% of the significant results are real effects
Answer A real effect of that size is missed 30% of the time
Power is 1 − β against a stated alternative, so the type II error rate here is 30%. The type I error rate is the level, 5%, and it is chosen separately. Loosening the level to 10% raises power rather than lowering it - at a fixed sample size the two error rates trade against each other. The share of significant results that are real also depends on how many tested strategies had an edge, which no power calculation knows.
Applied statistics
Regression identities, standard errors, and reading a coefficient correctly.
Example
A new signal shows an edge of 20 basis points a trade, with a standard error of 12 basis points. Across signals of this kind the true edge has mean zero and standard deviation 5 basis points. What edge should be used for sizing? Write the argument out.
Answer 2.96
The posterior mean under a normal prior is the measurement scaled by τ²/(τ² + s²) = 25/(25 + 144) = 0.15. So 20 × 0.15 = 2.96 basis points. The argument: 20 is what was measured, not what is there. The measurement carries 12 basis points of noise against a population whose true edges only spread 5 basis points, so most of the 20 is noise and the estimate has to be pulled most of the way to zero. The weight is the ratio of signal variance to total variance, which is the same shrinkage that appears in regression to the mean and in ridge. Sizing on the raw 20 would put on roughly 6.8 times the position the evidence supports, and the loss from that shows up as realised volatility rather than as a missing return.
Stating assumptions
Research screens mark the assumption you named as much as the number you produced.
Conditional expectation
Tower property and the expectation of a stopped process, which is most of what these rounds ask.
Example
An 8-sided die is rolled 20 times. What is the expected number of rolls that match the roll immediately before them?
- 2.5
- 0.297
- 0.125
- 2.375
Answer 2.375
One indicator per adjacent pair. There are 20 − 1 = 19 pairs, each matching with probability 1/8, so the expectation is 19/8 = 2.375. The indicators are not independent, since neighbouring pairs share a roll, but linearity of expectation never needed independence. Counting 20 pairs instead of 19 is the usual slip.
Overfitting and multiple testing
Why a result at the five per cent level means very little after the twentieth test.
Example
A backtest uses the closing price to decide a trade placed at that same close. Which flaw is this?
- Look-ahead bias
- Multicollinearity
- Heteroskedasticity
- Survivorship bias
Answer Look-ahead bias
The rule consumes information not available at decision time. Live, you could not have known the close before trading it.
Variance decomposition
Signal against noise, and how much of a measured edge should be believed.
Example
A research group tests signals of which 10% are genuinely predictive. The tests run at the 1% level and have 60% power against the effects that are real. Of the signals that come back significant, what percentage are genuinely predictive? Answer as a percentage.
Answer 87
Take 1,000 signals. 100 are real and 60 of those are detected. 900 are null and 9 of those are flagged anyway. The significant pile holds 69 signals, of which 60 are real: 87%. The argument: the 1% level controls P(flag | null), which is not what anyone wants to know. What matters is P(real | flag), and turning the conditional round needs the base rate. When only 10% of the hypotheses are true, the null pool is 9 times larger, so even a small false positive rate on a large pool competes with a good detection rate on a small one. Raising power helps the numerator; lowering the level or testing better hypotheses helps more, because both attack the denominator.
What a good score looks like
Research screens are marked by a person more often than a machine, so a numeric cutoff is rarely visible. Candidates consistently report that a wrong answer with a stated assumption scores better than a right answer with none.
How to train for it
- 01Rebuild the standard results rather than memorising them - a two-state chain, gambler's ruin, the regression slope from correlation and standard deviations.
- 02Work under time, but slower than a trading screen. These reward a correct setup, not speed.
- 03Say what would falsify the result. Research screens are looking for the person who reaches for that first, and it is the cheapest habit to build.
TRAIN IT HERE
The drills that match each section
Martingale Mutiny
Optional stopping played as a game. Try to find a betting strategy that beats a fair coin.
Ruin Walker
Gambler's ruin with live absorption probabilities.
Distribution Lab
Recognise a distribution from how its samples behave.
SIT THE FULL BATTERY
All the sections back to back on one clock, marked the way the real screen marks them, with a by-skill breakdown at the end. Included with any pass.
Also reported at Jane Street
Common questions
- Is the Jane Street stochastic processes test multiple choice?
- Reported as multiple choice or short typed answer. Formats move, so treat this as the shape rather than a guarantee.
- How long is the Jane Street stochastic processes test?
- Candidates report 10 to 15 questions in 30 to 60 minutes, roughly two to five minutes per question.
- Is there negative marking on the Jane Street stochastic processes test?
- No. A wrong answer costs nothing beyond the mark you would have earned, so leaving an item blank is never better than guessing at it.
- How do I practise for it free?
- Every drill linked on this page is free to play, with no account, inside a daily run cap. Questions are generated fresh each run, so there is nothing to memorise between attempts.
Hiring now: current Jane Street quant openings, refreshed daily.