Should You Start Your Aces or Save Them?
As September closes and Fall greets us, nothing heralds the change of the seasons like the start of the MLB playoffs. Every year the wild card round features big-spending monoliths versus out-of-nowhere surprises. It features divisional rivalries. It features upsets due to the quick nature of the best of 3 series. And best of all, it features a strategic showdown that hardly resembles the slog of the regular season.
No strategic decision is in the spotlight than the choice of which pitchers to trot out and for how long. The common wisdom is to start your pitchers in decreasing order of quality. Ace against ace.
But if you’re the fan of the underdog and you’re down halfway through the game, a nagging thought creeps into your head. Did we just waste our Ace? By putting your best guy against theirs maybe you didn’t get the full value of your ace. What if we saved our Ace for Game 2 and practically guaranteed a win?
Let’s take a quick look at the common wisdom of starting your aces as early as possible. Is this strategy really optimal? Is there any reason to maybe save your ace? If so, when?
Model Description
In order to not get bogged down into the details, we looked at a very simplified model of win probability. We took two teams in the 2025 playoffs – the Cleveland Guardians and the Detroit Tigers – and grabbed each of their three starters in games 1-3 of the playoffs. We held the Detroit Tigers rotation constant going from best to worst with Skubal, Flaherty, and Mize in that order. Then we calculated the probability of the Guardians winning the series for each possible order of their three starters: Gavin Williams, Bibee, and Cecconi. Did they do best if they sent their Ace against Skubal? Did they do better if they saved him?
To calculate the probability of winning in a pitcher v. pitcher matchup, we used the ‘win probability added’ stat on baseball reference. By dividing total win probability added by number of appearances, we can get the win probability added per game for that pitcher. For example Tarik Skubal was 6th in the majors this year with a total of 400% total win probability added over 31 appearances, good for about 13% per game.
This means that, for equally matched offenses and going against a league-average starting pitcher, Skubal ups the Tigers’ win probability to 63% (63=50+13% from Skubal). But the Guardians pitcher needs to factor into the conversation too. The Guardians Gavin Williams added about 6% win probability to his team each game. This would take the Tigers win probability down from 63% to just 57%.
Repeating the same idea for games 2 and 3 and for each possible order of pitchers for the Guardians, we can analyze the various ways in which the Guardians can set their rotation.
Results
The Guardians and Tigers played a three game series where a winner of two or more takes the whole thing. Skubal, Flaherty, and Mize in that order had a per-game win probability added of +13%, +0%, and -4%. The Guardians three pitchers Williams, Bibee, and Cecconi boasted stats of +6%, -2%, +1%. The table below shows the game-by-game win probabilities for Detroit against whichever pitching order Cleveland could have used. (A row heading of “1, 2, 3” means Williams pitched, then Bibbee, then Cecconi while “3, 2, 1” means they pitched in the opposite order).
CLE Order | DET G1 W% | DET G2 W% | DET G3 W% |
|---|---|---|---|
1, 2, 3 | 57% | 52% | 45% |
1, 3, 2 | 57% | 49% | 48% |
2, 1, 3 | 65% | 44% | 45% |
2, 3, 1 | 65% | 49% | 40% |
3, 1, 2 | 62% | 44% | 48% |
3, 2, 1 | 62% | 52% | 40% |
The game-by-game probabilities change fairly dramatically based on the order that Cleveland uses their starters. Ace v. Ace usually gives close to equal odds in each matchup while saving your Ace tends to lead to more lopsided affairs.
What is fairly shocking, though, comes when we look at the odds of winning the entire series. The needle hardly moves based on when Cleveland uses their best pitchers.
Across these six different rotations for Cleveland, the overall chances Detroit wins the series were 52%, 52%, 51.9%, 52%, 52%, and 52%. So, while the game-by-game odds can change fairly dramatically based on the order Cleveland uses their pitchers, the series barely reacts.
What About Longer Series?
We ran the same type of simulations for longer series like we’ll see in the divisional round, championship series, and World Series. While it’s tempting to think that longer series open up more opportunity for strategy, we find largely the same thing.
In a hypothetical five- or seven-game series between the Guardians and Tigers, the probabilities did change based on the order of the pitchers. However, the effects were nearly totally attributable to certain pitchers going more than others. For example, with a three man rotation in a 7 game series it is clearly best to throw out your Ace first so they pitch three times while everyone else pitches twice.
Other than this effect though, there was very little interesting strategy associated with targeting certain matchups with your pitching rotation.
So What Should You Do?
These simulations show that there is very little to be gained by playing tactical games with starting lineups. Outside of maximizing the number of starts for your best pitchers, there is basically no advantage to be gained with pitching matchups.
There are two things unexplored by this very simple model. First, because there is no benefit to starting your Aces in certain games due to matchups alone, you can prioritize rest. For example, if you’re in a tough wildcard race and need to win to make the playoffs, it’s probably better to start your Ace in game 162 even though they’ll be unavailable until later in the wildcard round. Since it doesn’t matter when they pitch in the wildcard round, it’s better to use them when it does matter in the regular season. For pitching rotation strategy, it all comes down to rest and freshness. Targeting or avoiding matchup is likely a waste of effort.
This might not always be the case in other sports. In baseball, even the best pitchers move the needle only a very small amount – rarely more than 10%. But in other sports the win probabilities may sway more dramatically based on the chosen matchups. For example, there are Chess tournaments where the format is 2 teams of 5 squaring off in 5 individual games. Chess games are certain to be more lopsided than the 40-60% settings of baseball. Maybe the best strategy here is to put your worst against their best and ensure an advantage in the remaining games? In a followup article, we’ll explore the ‘Ace v. Ace’ question in a slightly more mathematical setting.
To receive email updates when new articles are posted, use the subscription form below!
