AFL Trade Value: Pricing Players and Draft Picks
What is trade value?
Trade value puts a player and a draft pick on the same scale, so a trade can be graded as what a club received minus what it gave up. The unit is the TORP value a player or pick is expected to produce over the next eight seasons, measured against a player a club could pick up for nothing.
It answers “who got more on-field value”, not “who made the smarter decision”. A club can lose on value and still be right to trade, for example to clear salary or to move picks into a stronger draft.
What goes into it?
For a player, four forecasts, each made fresh for every season ahead:
- His rating. Start from his current TORP. Ratings drift toward average from one season to the next, and age moves them a little: about +0.1 a season in the early 20s and about -0.1 a season from 29.
- Whether he is still on a list. At an average rating, 99% of 21-year-olds are on a list the next season, 89% of 31-year-olds and 47% of 35-year-olds. A higher rating and more games raise the chance.
- His games and time on ground. How many games he plays and what share of each game he is on the field, based on his age, rating and this season’s numbers.
- Where he was drafted, for his first three seasons only. At the same rating, a pick 1 is expected to rate about 0.8 points higher the next season than a pick 20.
For a draft pick, the average outcome of players taken at that pick: what the 2020 and 2021 draftees have actually produced so far, plus a forecast of their remaining seasons on the same basis as for players. A pick has no rating yet, so the record of past picks is its forecast. Busts count: a draftee who never plays adds nothing, and a below-replacement season counts against.
How is it calculated?
Step 1: value one season. A season is worth games played × time on ground × (TORP − replacement level). TORP is a per-full-game rate, so a game at 78% time on ground counts for 0.78 of it. Replacement level is −1.38, the average first-season rating of mature players picked up off the scrap heap: what a club gets if it fills the spot for free.
Step 2: simulate the next eight seasons. For each player we run 2,000 possible careers. Each season he stays on a list with the forecast probability, his rating moves by the age curve plus a random amount the size of a typical season-to-season change, and he plays the forecast games at the forecast time on ground.
Step 3: discount later seasons. Each season counts 10% less than the one before, so value in 2027 counts in full and value in 2034 counts about half.
Step 4: price picks the same way. The 2026 picks use the curve for players taken at that pick. A future pick is expected to land where its club finished on the 2026 ladder, pulled halfway to mid-round for 2027 and fully to mid-round for 2028, then moved back to allow for the 17-year-olds Tasmania has already signed out of those draft classes. It is discounted one more season for each year it waits.
Step 5: count only the seasons at the club. A club only gets value while the player is on its list. Players who change clubs stay a median three seasons, and draftees stay longer the earlier they were picked, so each season is also weighted by the share of similar players still at the club by then.
What does a pick come out at?
On the AFL’s own scale, where pick 1 is worth 3,000 points, our pick values land close to the AFL’s Draft Value Index for the picks we could check: pick 2 is 2,469 against the AFL’s 2,481, pick 5 is 1,768 against 1,795, and pick 25 is 536 against 590. Values fall away quickly after the first round, and picks after about 48 are worth close to nothing on average over eight seasons.
How long do players stay?
Players who changed clubs between 2000 and 2020 (787 players) stayed a median three seasons at the new club. Players who arrived aged 31 or older stayed a median one season. Most stays end with the player leaving the league, not moving again: only 112 of the 787 moved on to another club.
Draftees who played for the club that picked them (1,771 players drafted 1991 to 2020) stayed a median nine seasons when taken in the top five picks, and three seasons when taken after pick 45. These medians allow for the players still at their club, whose stays are not over yet; counting only finished stays, top-five picks stayed a median seven.
What are the limits?
- Salary is not counted. Players cost salary-cap space and draft picks are cheap, so trading for a player looks better here than it is. We have not found public contract data to fix this.
- Contract length is not counted. The time-at-club figures above stand in for it.
- The pick values rest on two drafts. The 2020 and 2021 drafts are the only ones whose first five seasons fall inside the TORP ratings, which start in 2021. That is 124 picks, and their later seasons are forecast, not seen.
- Future pick positions are estimates. Ladder positions only partly carry over from one season to the next, and the effect of Tasmania’s signings on the 2027 and 2028 drafts is our reading of the rules, not an AFL figure.
- The draft-pick effect is not settled. Adding where a player was drafted improved the forecasts for held-out seasons, but the improvement is small enough that it could be chance.
- Injury years count against a player. A short season lowers his forecast games.
How accurate is it?
Not yet tested end to end: the TORP ratings only go back to 2021, so no eight-season forecast can be checked against what happened. The one-season rating forecast is tested. Predicting 2026 ratings from 2025 for 631 players, it missed by 1.17 points on a typical player, against 1.22 for assuming no change.