Probabilistic programming allows users to express their models in a more declarative way, making the code more readable and maintainable. Here's an example:

```python

import numpy as np

import pymc3 as pm

Define the model

model = pm.Model()

Define the variables

x = pm.Normal("x", mu=0, sd=1)

y = pm.Normal("y", mu=0, sd=1)

z = pm.Normal("z", mu=x + y, sd=1)

Define the observations

observations = np.array([1, 2, 3])

Fit the model to the observations

trace = model.sample(draws=1000, chains=4)

Print the results

print(trace)

```

This code defines a simple probabilistic model with three variables, `x`, `y`, and `z`. The variables `x` and `y` are defined as independent normally distributed random variables, and `z` is defined as the sum of `x` and `y`. The model is then fitted to three observations using Markov chain Monte Carlo (MCMC) sampling, and the results are printed.

This code is a lot more concise than traditional programming approaches for statistical models, which would involve manually writing out the likelihood function and the MCMC sampling algorithm. Probabilistic programming makes it easier to write complex statistical models and focus on the modeling task rather than the implementation details.