PRACTICE GUIDE TOWER RESEARCH
Tower Research Quant Developer Practice Test
This guide offers quant developer preparation for readers interested in Tower Research Capital. The general practice ranges below are not a verified assessment specification for that employer.
Those are ranges across firms running this format rather than Tower Research Capital's own figures, so treat them as the shape to train against and check the instruction screen on the day for the marking rule.
These are Outcry preparation topics, not a verified current firm assessment specification. Check your role-specific invitation for the test, timing, tools and recruitment stage.
Outcry is not affiliated with Tower Research Capital 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 Tower Research Capital’s own.
What it screens
A quantitative trading firm whose public careers material covers trading, research and engineering.
- ✓Reading concurrent code and pointing at the defective line
- ✓Order-book data structures: what happens on a cross, a cancel, a partial fill
- ✓Algorithmic complexity chosen for the input sizes stated in the problem
- ✓Monte Carlo methods and when a simulation beats a closed form
Where it sits at Tower Research Capital
A quantitative trading firm whose public careers material covers trading, research and engineering. The topics here are Outcry preparation suggestions, not a verified firm syllabus or recruiting stage.
Outcry also has 2 other topic guides for readers interested in Tower Research Capital. Their presence does not mean applicants must sit that many tests or that they belong to one recruiting round.
The format
These are general practice ranges, not a verified current assessment specification for Tower Research Capital. Check the employer’s invitation for the actual format.
| Questions | 2 to 4 problems |
|---|---|
| Time | 60 to 120 minutes |
| Per question | 20 to 40 minutes |
| Negative marking | No |
| Answer style | Code editor, run against hidden tests |
| Where it sits | Preparation topics · employer round unconfirmed |
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 Tower Research Capital’s.
Complexity that is actually graded
Hidden tests sized so the naive solution times out. A correct answer that is too slow scores the same as a wrong one.
Reading someone else's code
A diff or a function with a bug in it, and the question is where.
Example 1
int count = 0;
for (int a = 0; a < 4; ++a)
for (int b = a; b < 4; ++b)
++count;What is the final value of count?
Answer 10
The inner loop runs 4, 3, … 1 times - that is 4·5/2 = 10.
Example 2
A lock-free SPSC ring buffer publishes its head index with memory_order_relaxed. What can go wrong?
- The index can wrap incorrectly
- The consumer may see the new index before the slot's data is visible
- The producer may block
- Nothing - SPSC needs no ordering
Answer The consumer may see the new index before the slot's data is visible
The publishing store needs release semantics to pair with the consumer's acquire load, or the buffer write can be reordered after it.
Example 3
Which change most reduces cache misses when summing a large matrix?
- Use a larger integer type
- Iterate in row-major order matching the memory layout
- Unroll the loop 2×
- Mark the accumulator volatile
Answer Iterate in row-major order matching the memory layout
Sequential access uses the whole cache line that was fetched. Striding across rows discards most of every line.
Numerical and data handling
Floating point, aggregation and joins on data that does not fit the obvious shape.
Example 1
Cov(X, Y) = 11. What is Cov(2X, 2Y + 2)?
Answer 44
Covariance is bilinear and a constant shift drops out: Cov(2X, 2Y + c) = 2 x 2 x Cov(X, Y) = 44. The added constant has no variance, so it cannot covary with anything.
Example 2
A fund has volatility 20% and the index has volatility 24%. Their correlation is 0.8. What is the fund's beta to the index?
Answer 0.6666666666666666
Beta = correlation x (fund vol / index vol) = 0.8 x 20/24 = 0.667. Correlation and beta only agree when the two volatilities match, which here they do not.
Example 3
A price walks on the whole numbers 0 to N. Each step it moves +1 with probability p and -1 with probability 1 - p, independently. It stops the first time it touches 0 or N. N = 8 start = 4 p = 0.6 One question, six parts. Each part uses the part before it.
Part 1 of 6. Take p = 0.5 for this part only. What is the probability the walk stops at N rather than at 0?
Answer 0.5
With no drift the stopping probability is linear in the starting point, so it is start/N = 4/8 = 0.5. That is the only linear function that is 0 at 0, 1 at N, and equal to the average of its neighbours everywhere between.
Edge cases
Empty input, one element, duplicates and overflow. The hidden tests always include them.
Statistics in code
Rolling windows, correlations and quantiles, implemented rather than imported.
Example
A model is made more flexible while the sample size is held fixed. What happens to its error decomposition?
- Bias falls and variance rises
- Bias falls and variance is unchanged
- Bias and variance both fall
- Bias rises and variance falls
Answer Bias falls and variance rises
Flexibility lets the fit track the true function more closely, which cuts bias, and lets it track the noise more closely, which raises variance. Test error is the sum of the two plus irreducible noise, so it falls and then turns back up. Constraining a model - fewer parameters, shrinkage, more regularisation - trades the other way.
Reading a spec precisely
Return type, ordering and tie-breaking are graded, and are where most silent failures come from.
What a good score looks like
No current employer cutoff is verified for these general practice ranges. Use the worked examples to identify what to study and check your role-specific invitation for the actual assessment rules.
How to train for it
- 01Implement the core structures from scratch once each. The screens ask you to build them, not use them.
- 02State the complexity before you write. If it is worse than the input size allows, the approach is already wrong.
- 03Get the function signature exactly right before anything else. Return type and ordering are graded, and a correct algorithm behind a wrong signature scores zero.
TRAIN IT HERE
The drills that match each section
Concurrency Clash
Find the data race, name the fix - real C++ defect patterns, five levels.
Order Book
Matching-engine mechanics: crosses, cancels and partial fills.
Algorithm Interview Track
DP tables, Monte Carlo estimation and speed rounds at four difficulty tiers.
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 Tower Research Capital
Common questions
- Is the Tower Research quant developer test multiple choice?
- This guide uses code editor, run against hidden tests for practice. The employer’s current answer format was not verified here.
- How long is the Tower Research quant developer test?
- This guide uses general practice ranges of 2 to 4 problems questions in 60 to 120 minutes. They are not a verified current employer format.
- Is there negative marking on the Tower Research quant developer 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?
- The warm-up drills are free with no account, and every pass opens the rest of the drills linked on this page. Questions are generated fresh each run, so there is nothing to memorise between attempts.