Martingale Report

1. – QUESTION

The simulation results demonstrated that all 1000 episodes achieved the target winnings
of $80. Using python, I was able to calculate the proportion of successful episodes by
dividing the count of such episodes by the total number of episodes. This method en-
sures that the conclusion is made from data driven calculations without relying on vis-
ual inspection of the plots. The results showed that all 1000 episodes achieved the tar-
get winnings of $80, resulting in a calculated probability of 100%.

This means that in every episode, the gambler successfully reached the $80 target with-
in 1000 sequential bets. This outcome is consistent with the nature of the Martingale
betting strategy. With an unlimited bankroll, the strategy guarantees eventual success as
the gambler can just double the bet after a loss, ensuring recovery of losses and a net
gain on the next win.

In conclusion, the probability of achieving $80 within 1000 bets is 1.0, as the unlimited
bankroll condition ensures the effectiveness of the Martingale strategy in achieving the
target winnings.

1
2. – QUESTION

The estimated expected value of winnings after 1000 sequential bets is $80. This out-
come is consistent with the design of the Martingale strategy and the simulation’s early
termination condition. This ensures that every episode successfully achieves the target
winnings before reaching 1000 spins.

The expected value of winnings was computed as the mean of the final winnings from
all 1000 episodes. Since most episodes reached or exceeded the $80 target early in the
sequence and persisted this value for the remaining spins, the average winnings led to
$80.00. This result aligns with the structure of the Martingale strategy, which is de-
signed to incrementally achieve the target with consistent success when applied with an
unlimited bankroll.

3. – QUESTION

In Experiment 1, the upper standard deviation line (mean + standard deviation) and the
lower standard deviation line (mean - standard deviation) stabilize. This stabilization
occurs because the winnings for all episodes either reach the $80 target or persist at $80
for the remainder of the spins due to the stop-and-persist condition. This behavior is
evident in Figure 2, where the mean winnings stabilize at $80, and the shaded region
representing the standard deviation shrinks and eventually vanishes after approximately
200 spins.

Regarding whether the standard deviation lines converge, they do. The graph shows
that the standard deviation lines essentially converge to the mean after approximately
200 spins, as the winnings stabilize at $80 for all episodes. The lack of any visible
shaded area beyond 200 spins supports this observation—once every episode reaches
$80 and persists, there is no variability left in the cumulative winnings. As a result, the
standard deviation becomes zero, indicating perfect convergence.

2
4. – QUESTION

The probability of winning $80 within 1000 sequential bets in Experiment 2, based on
the simulation output, is calculated to be approximately 62.30%. This probability was
derived using the simulation results by counting the number of episodes where the
maximum cumulative winnings reached or exceeded $80 and dividing that by the total
number of episodes (1000).

The reduced probability compared to Experiment 1 (where the probability was 100%)
highlights the impact of the bankroll limitation. In Experiment 2, the strategy is more
sensitive to extended losing streaks, which can result in the gambler running out of
money before hitting the target winnings. Despite this, the probability of achieving $80
remains relatively high due to the doubling-bet strategy, which leverages the finite
bankroll effectively in many scenarios.

5. – QUESTION

The estimated expected value of winnings after 1000 sequential bets in Experiment 2 is
approximately -46.672. This result was calculated by averaging the final winnings
across all simulated episodes, considering the experiment’s constraints. While some
episodes successfully reach the $80 target, many others end in significant losses due to
the bankroll limitation. The persistence of losses once the bankroll is depleted strongly
influences the average, pulling the expected value into the negative range.

The Martingale strategy, which involves doubling the bet after each loss, becomes un-
sustainable under the constraints of a limited bankroll. This results in a higher likeli-
hood of episodes ending in losses rather than profits. Consequently, the expected value
reflects the inherent risk and limitations of the strategy in this scenario, underscoring
the financial risk involved when applying the Martingale betting system with con-
strained resources.

6. – QUESTION

In Experiment 2, the behavior of the upper and lower standard deviation lines (mean ±
standard deviation) provides critical insights into the volatility and distribution of cu-
mulative winnings. Observing Figure 4, it is evident that the standard deviation lines
initially expand due to the variability introduced by successive bets. However, after

3
approximately 200 spins, the standard deviation range stabilizes, as reflected in the
shaded region. This stabilization occurs due to the constraints imposed by the gambling
rules: players either achieve the target winnings of $80 or exhaust their bankroll. These
boundaries cap the variability in cumulative winnings, as no further gains or losses can
occur once these limits are reached.

Similarly, Figure 5 demonstrates a gradual stabilization of the median winnings at

higher spin numbers. While the associated standard deviation region initially widens, it
also stabilizes around the same point, aligning with the median line. This indicates that
the gambling rules create a bounded outcome space, reducing further variability and
preventing the standard deviation lines from continuing to grow indefinitely. The
graphs clearly illustrate this effect, as the mean and median winnings flatten, and the
standard deviation lines stabilize, reflecting the constrained dynamics of the betting
strategy and its reliance on bankroll limits and target winnings.

The standard deviation lines in both figures do not converge with each other as the
number of sequential bets increases. Instead, they reach a steady state where the spread
between the upper and lower bounds remains consistent after about 200 spins. This be-
havior arises because the gambler either reaches the target winnings or exhausts the
bankroll, creating a bounded outcome space. The graphs illustrate this effect clearly,
where the mean and median winnings flatten, and the standard deviation lines stabilize,
reflecting the constrained dynamics of the betting strategy.

4
7. – QUESTION

Using expected values in experiments offers several benefits over relying on a single
random episode. One key advantage is statistical reliability. Expected values, calculated
by averaging multiple trials, reduce the impact of outliers or extreme results. A single
episode may represent an unusually favorable or unfavorable outcome, leading to mis-
leading conclusions. Expected values provide a more robust and accurate representation
of the system’s overall behavior.

Another benefit is the reduction of randomness. In probabilistic systems like gambling,

individual episodes are heavily influenced by chance, which can obscure trends. Ex-
pected values smooth out these fluctuations, offering clearer insights into the system’s
long-term behavior. This reliability makes them useful for generalization, providing a
more accurate basis for predicting strategy performance.

Expected values also enable consistent comparisons across scenarios. For example,
comparing winnings under unlimited and limited bankrolls offers a standardized metric
for evaluation. Additionally, expected values often include measures of variability, like
standard deviation, which reveal outcome volatility, aiding better decision-making.