Best of Seven Jul 2022

Best of 7
Two teams face off in a best of 7 series.

I provide analysis and plots.

Click for Best of Seven Live Chart


Evenly Matched Teams

If the two teams are evenly matched, then we say the probability that Team A wins each game is 0.50. The probability that Team B wins each game is also 0.50, and the sum of these two probabilities is 1.0 as expected.

The question here is, what is the likelihood of each possible playoff outcome:

  • 4-0 Team A wins
  • 4-1 Team A wins
  • 4-2 Team A wins
  • 4-3 Team A wins and
  • 4-0 Team B wins
  • 4-1 Team B wins
  • 4-2 Team B wins
  • 4-3 Team B wins

Team A is Stronger

If the two teams are not evenly matched, then let’s say the probability that Team A wins each game is P. The probability that Team B wins each game is now therefore Q = 1 – P, and the sum of the probabilities P + Q = 1.0 as expected.

For example, if the probability that Team A wins each game is P = 0.81, then the probability that Team B wins each game is Q = 1 – P = 0.19.

Again, the question is, what is the likelihood of each possible playoff outcome:

  • 4-0 Team A wins,
  • 4-1 Team A wins,
  • 4-2 Team A wins,
  • 4-3 Team A wins, and
  • 4-0 Team B wins,
  • 4-1 Team B wins,
  • 4-2 Team B wins,
  • 4-3 Team B wins?

Discussion

This answer to this problem is relatively easy to discuss yet somewhat complex to calculate, Fortunately, it can be broken down into several similar parts.


Series Goes 4-0 or 0-4

An easy start: if Team A wins in 4 straight, then the probability is simply

    \[{P^4} = 0.81^4 = 0.430\]

And similarly: if Team B wins in 4 straight, then the probability is simply

    \[{Q^4} = 0.19^4 = 0.001\]

in this example very unlikely.


Series Goes 4-1 or 1-4

Now, if Team A is to win the series 4-1, then after 4 games, they must be up 3-1 just before the 5th game in the series. That means there is one loss in the 4 games played and there are only 4 ways that can happen.

Therefore the probability is

    \[\left[{P^4}\cdot{Q}\right]\cdot\left[{_4}{{C}_1}}\right] = ({0.81^4}\cdot {0.19})\cdot {4} = 0.327\]

Explanation:

  • ({P^4}\cdot{Q}) is the one-time probability of Team A winning 4 games and Team B winning one game.
  • {_4{{C}_1}} is the number of combinations in which one team wins 3 games and the other wins only one game.
    Click for explanation of combinations

Similarly, if Team B is to win the series 4-1, then after 4 games, they must be up 3-1 just before the 5th game in the series. Therefore the probability is

    \[\left[{Q^4}\cdot{P}\right]\cdot\left[{_4}{{C}_1}}\right] = ({0.19^4}\cdot {0.81})\cdot {4} = 0.004\]


Series Goes 4-2 or 2-4

If Team A is to win the series 4-2, then after 5 games, they must be up 3-2 just before the 6th game in the series. That means there are 2 losses in the 5 games played and there are only 10 ways that can happen.

So we have established the pattern, if Team A is to win the series 4-\mathbf{X}, the probability is

    \[\left[{P^4}\cdot{Q^\mathbf{X}}\right]\cdot\left[{_{(\mathbf{X}+3)}{{C}_\mathbf{X}}}\right]\]

Likewise, if Team B is to win the series \mathbf{X}-4, the probability is

    \[\left[{Q^4}\cdot{P^\mathbf{X}}\right]\cdot\left[{_{(\mathbf{X}+3)}{{C}_\mathbf{X}}}\right]\]


    \[_{(\mathbf{0}+3)}{{C}_\mathbf{0} = \prescript{_3}{{C}_{0}} = \frac {3!}{{3!}\cdot{0!}} = 1\]

    \[_{(\mathbf{1}+3)}{{C}_\mathbf{1} = \prescript{_4}{{C}_{1}} = \frac {4!}{{3!}\cdot{1!}} = 4\]

    \[_{(\mathbf{2}+3)}{{C}_\mathbf{2} = \prescript{_5}{{C}_{2}} = \frac {5!}{{3!}\cdot{2!}} = 10\]

    \[_{(\mathbf{3}+3)}{{C}_\mathbf{3} = \prescript{_6}{{C}_{3}} = \frac {6!}{{3!}\cdot{3!}} = 20\]


Observation

Notice that compared to a single game, a seven game series increases the probability that the strong team wins the series. For example, with P = 0.81, as shown above, the probability that Team A wins the series is calculated at 0.972. Note further that the probability that Team B wins the series decreases from 0.19 to only 0.028.

Using Odds:

The single game had odds of 81:19 or about 4:1 whereas the series has odds of 972:28 or about 35:1.


Plots

See the plots of all possible outcomes for all probabilities:

    \[P \mid 0 < P < 1\]

Example probabilities are given for the specific case where P = 0.81.

Live Plot

Click for Best of Seven Live Chart

Using PLOTLY live plots

Each chart series can be toggled off and on

by clicking its legend entry.

On a desktop device, reset the chart to full scale

by reloading the web page.

On a mobile device, reset the chart to full scale

by double tapping.

All Outcomes
Winning Outcomes
The live chart looks best on a laptop or desktop screen. If viewing on a mobile device select ‘Request Desktop Website’ and use landscape orientation of the device.

Static Plots

All Outcomes
Winning Outcomes


Reference

Best of Seven Outcome Probability Dec 2020


Bonus Calculations (added June 2024)

One game played

What if one game has already been played and Team A won that game; now what is the probability of each series arrangement?

If Team A is to win the series 4-\mathbf{X}, the probability is

    \[\left[{P^3}\cdot{Q^\mathbf{X}}\right]\cdot\left[{\prescript{_{(\mathbf{X}+2)}}{{C}_\mathbf{X}}}\right]\]

Two games played (imagine the 2024 Celtics in such a situation)

What if two games have already been played and Team A won both games; now what is the probability of each series arrangement?

If Team A is to win the series 4-\mathbf{X}, the probability is

    \[\left[{P^2}\cdot{Q^\mathbf{X}}\right]\cdot\left[{\prescript{_{(\mathbf{X}+1)}}{{C}_\mathbf{X}}}\right]\]

Closed form solution

Let’s add another variable. Let’s say \mathbf{Y} games have been played and that Team A won those games.
Now what is the probability of each series arrangement?
If Team A is to win the series 4-\mathbf{X}, the probability is

    \[\left[{P^{(3-\mathbf{Y})}\cdot{Q^\mathbf{X}}\right]\cdot\left[{_{(\mathbf{X}+3-\mathbf{Y})}{{C}_\mathbf{X}}}\right]\]

If Team B is to win the series \mathbf{X}-4, the probability is

    \[\left[{P^{(\mathbf{X}-\mathbf{Y})}\cdot{Q^4}\right]\cdot\left[{_{(\mathbf{X}+3-\mathbf{Y})}{{C}_{(\mathbf{X}-\mathbf{Y})}}}\right]\]

Sample prediction with Celtics up by two games
(P = 0.571, Y = 2)

Leave a Reply

Your email address will not be published. Required fields are marked *