How to Use ADP and Projection Systems

By 4Baseball Staff · November 5, 2026 · 11 min read

Average draft position and projection systems explained: how ADP is calculated, how projections are built, how to turn them into draft values, and the common mistakes that cost fantasy managers a season.

Every fantasy baseball draft starts with two numbers next to a player's name. One is average draft position, usually shortened to ADP, which says where other managers have been taking him. The other is a projection, which says what a forecasting system expects him to do. Used well, the pair tells you what a player costs and what he is worth, and the gap between those two is where drafts are won. Used badly, they turn into a list that you read from top to bottom and a feeling that you have a plan.

This guide explains what each number is, how it is built, where it misleads, and how to combine them with your own league's settings. It does not recommend specific players, because rosters, injuries and playing time change every week; it explains the method, so you can apply it to whichever list you are looking at.

What ADP is, and what it is not

ADP is an average. For every draft in a sample, take the overall pick at which a player was selected, add those pick numbers together and divide by the number of drafts. If a player went 22nd, 25th, 31st and 28th in four drafts, his ADP is (22 + 25 + 31 + 28) divided by 4, or 26.5. An average of 26.5 means he is typically a third-round pick in a 12-team league.

That is all ADP is: a record of past behavior. It is not a ranking from an expert, and it is not a forecast of how the player will perform. It tells you the market price. Markets can be wrong, but they can also be right for reasons you do not know, such as a recent injury or a change in role that the drafters have heard about and you have not. Treat ADP as the first piece of evidence, not the last.

Three details determine how useful a given ADP list is:

• Sample size and source. An ADP built from thousands of drafts is steadier than one built from a few dozen. Some sites publish ADP from drafts on their own platform, and some publish it from leagues in which participants play for high stakes. The two groups of drafters behave differently.
• Date. ADP in February and ADP in late March can differ a lot for the same player, because injuries, trades and lineup news arrive in between. Use the most recent list, and look for lists that show a date range.
• League settings. An ADP list assumes a league size, a scoring format and a number of roster spots. A list built from 15-team leagues puts players at different overall numbers than one built from 10-team leagues.

Turning ADP into rounds and picks

ADP is an overall pick number, but you draft in rounds. To convert, divide by the number of teams in your league. In a 12-team league, picks 1 through 12 are round one, 13 through 24 are round two, and so on. The formulas are simple:

• Round = the pick number divided by the league size, rounded up.
• Pick within the round = the pick number minus the league size times (round minus 1).

Using the earlier example, an ADP of 26.5 in a 12-team league is round three (the picks run from 25 to 36), around the second pick of the round.

Most leagues use a snake draft, in which the order reverses every round. The manager who picks first in round one picks last in round two. If you hold pick number k in a league of n teams, your picks in the first two rounds are k and (2n - k + 1). In a 12-team league with the third pick, you also pick 22nd (24 - 3 + 1), so you should read ADP for players near 3 and near 22 rather than for the whole board. A manager at the end of round one has a different problem: picks 12 and 13 come back to back, and then there is a long wait. Writing down the overall numbers you hold before the draft turns a long ADP list into a short list of realistic targets.

In an auction draft, where managers bid on players with a fixed budget, the equivalent number is the average auction value. It is read the same way: the typical price paid. Your budget gets scaled by the league's roster rules, so a dollar figure from a list for a different format needs adjusting.

What projection systems are

A projection is a forecast of a player's statistics for a coming season. The systems that produce them are different from a columnist's opinion because they follow a consistent method, they are applied to every player, and their results can be tested after the season.

The simplest useful method is a weighted average of recent seasons with a correction. A well-known version, called Marcel, takes the last three seasons with the most recent weighted the heaviest, regresses the result toward the league average because small samples exaggerate both luck and skill, and then adjusts for age. It is deliberately basic, and it is often used as a benchmark: a more complicated system has to beat it to earn its complexity.

More elaborate systems add information. Some compare a player with historical players who had similar profiles and used their later careers as a guide. Some use batted-ball data, strikeout and walk rates, and minor-league performance for younger players. Several are published each winter by the people who build them, and each has a different emphasis. Their names are easy to find, but the lesson that matters here is that all of them are predicting an average outcome, not a specific one.

The key properties of a good projection are these:

1. It regresses to the mean. A player who hit .340 on a high rate of balls in play is not projected to hit .340, because part of that number was luck. Sample Size and Regression to the Mean in Baseball Statistics explains why.
2. It adjusts for age. Hitters and pitchers rise and fall on a recognizable curve, described in Aging Curves: When Hitters and Pitchers Peak and Decline. A projection for a 24-year-old should not look like one for a 36-year-old with the same history.
3. It accounts for the ballpark and the league. A hitter moving to a new park carries a new environment with him.
4. It shows a range, not a point. The honest version of a projection is a median with an uncertainty band. Many sources show only the median, and you should supply the uncertainty yourself.

Where projections are weakest

The weakest part of any projection is playing time. A system can estimate a hitter's rate of home runs per plate appearance with some confidence, but how many plate appearances he gets depends on injuries, lineup spots, platoons, a manager's decisions and the strength of the players behind him. Counting statistics such as runs, runs batted in and saves are the product of rate times opportunity, and the opportunity part is hard to forecast.

Pitchers are harder to project than hitters. Their performance is more volatile year to year, injuries are more frequent, and relievers' save totals depend on role, which can change in a week. A closer's projection is really a projection of whether he keeps the job.

Projection systems also do not predict injuries well. Most apply a general reduction for risk, but they cannot know about the sore elbow reported in spring training. This is where your own research adds value: news after the projection date, changes in role, and playing-time battles.

Finally, projections are averages across many players, so some players will do much better and some much worse. If a system projects 25 home runs for ten hitters, you should not expect each to hit 25. You should expect a spread, with a few at 35 and a few at 15.

Combining systems and your own judgment

If several projections are available, consider averaging them. Each system makes different errors, and the average of several is usually at least as good as the best single one. It also tells you where there is disagreement. When two systems differ widely for a player, the cause is usually something specific, such as a change in role, a young player's development, or the treatment of a good year in a small sample. That is where to look hardest.

You should also adjust for your league. A projection by itself does not know whether your league counts batting average or on-base percentage, whether it rewards saves or holds, or whether it penalizes strikeouts. The same player can be worth very different amounts in different formats. The steps are below.

1. Find the projected statistics for the player.
2. Apply your league's scoring to them to get a total.
3. Compare the total to what a replacement-level player at the same position would score.

The difference in step 3 is the player's value. The idea is simple: what counts is not how many points a player produces, but how many more than the best player you could pick up for free.

A worked example in a points league

The following example is hypothetical, with invented players and an invented scoring system, to show the mechanics. The scoring awards 1 point per single, 2 per double, 3 per triple, 4 per home run, 1 per run, 1 per run batted in, 1 per walk, 1 per stolen base, and takes away 1 point per strikeout.

Player A has the better home run total and the better RBI total, so he looks like the star when scanning a list. But Player B scores 19 points more under this scoring, because he hits more singles, steals more bases, walks more and strikes out far less. If a manager drafted by home runs alone, he would take Player A first. If a manager drafted by projected points, he would take Player B first. That is the practical use of a projection: it replaces a headline stat with a total that matches the rules of your league.

Now position scarcity. Suppose the best shortstop projects to 400 points and the best shortstop likely to be available late projects to 300. The gap is 100. An outfielder who projects to 420 when a late outfielder projects to 380 has a gap of 40. The shortstop is the more valuable draft pick even though his raw total is lower. This logic is why the same projection list can lead to different draft orders in different leagues.

Rotisserie and head-to-head leagues

In a rotisserie (roto) league, you earn points in each statistical category based on where you rank among all teams. The usual approach is to convert each category into a common measure of how many standings points a unit of the statistic is worth, such as a home run or a point of batting average. Standard deviations are often used: a player's projected total in a category is compared with the average of the players likely to be drafted, in units of the spread. The sum across categories gives a single number.

In a head-to-head league, you win or lose each week, and the weekly variation in counting statistics matters more than the season totals. Some managers value consistency, such as hitters with a regular lineup spot and starting pitchers with a regular turn, a little higher for that reason. ADP lists built from one format are a rough guide in the other, so look for one that matches your format.

Using ADP and projections together

The two numbers answer different questions, and the useful move is to compare them. ADP asks what other managers are paying; the projection asks what the player should deliver. When a player's projection ranks higher than his ADP suggests, the market may be undervaluing him, and you can wait a little longer for him. When his ADP is well above his projected rank, you are being asked to pay a premium, and you should be sure the reason is one you accept.

A few rules of thumb help.

• Do not reach on a rank list. If you rank a player 40th and his ADP is 80, you can usually wait. The 40th rank is your opinion, but the ADP says what the room will do. Taking him early costs you the players who would have been your next choices.
• Know when the room moves. ADP is an average, not a schedule. Some players are drafted well before their ADP when a run on a position starts. Watch the board and note when a position tier is about to end.
• Draft in tiers. Group players with similar projected value into tiers. Within a tier, the difference between players is smaller than the uncertainty of the forecast, so you can choose by need, risk or preference. The moment to act is when a tier is nearly empty.
• Take value, then need. Early in the draft, take the best available value. Later, fill the positions that remain.

The ADP spread also matters. A player with an average of 60 who is taken between pick 40 and pick 90 is a very different case from one who always goes between 58 and 62. A wide spread means the room disagrees, which can mean an opportunity or a trap.

Mistakes to avoid

• Treating ADP as a ranking. It is a record of what other people did, and they make mistakes.
• Anchoring on last year's stats. A player's last season is one input, not the forecast. Regression and age matter.
• Using a single projection as truth. Compare more than one, and note the disagreements.
• Ignoring league settings. A list built for a different format will mislead you.
• Skipping the news. A projection is built before the injury report. Check the latest news before you draft.
• Forgetting that the draft is only the start. After the draft, ADP stops mattering; rest-of-season projections and the waiver wire take over. A good draft with no follow-up management loses to a decent draft with good management.

Common questions

Is ADP a good predictor of performance? It is a good predictor of cost. As a predictor of performance it is imperfect, because the market is made of people who share the same information and the same biases.

Should I draft strictly by ADP? No. ADP gives you a sense of where the market will be, but a draft is won by finding gaps between price and value.

Which projection system is best? There is no clear winner year after year. The safest approach is to use an average of several.

Do projections work for keeper leagues? They help, but a keeper league adds age and contract value. See Keeper and Dynasty League Strategy.

Where to go next

If you are new to the game, start with Fantasy Baseball for Beginners, then read First-Round Strategy in Fantasy Baseball for how to apply ADP at the top of the draft and Fantasy Baseball Trade Strategy for managing the season after it. Commissioners should read How to Run a Fantasy Baseball League as Commissioner. The site's fantasy tools and the fantasy baseball topic page collect more.