On Sunday, March 19, a 12-3 record could yield vastly different outcomes depending on which Grand Prix you participate in. At the 2,719-player event in Shizuoka, two 12-3 players didn’t cash. At the 570-player tournament in Porto Alegre, two 12-3 players made Top 8!

The reason for this striking disparity is the current payout structure: Prize money (at individual events) is paid out to 64th place if there are fewer than 3,000 players and to 180th place if there are at least 3,000 players. In other words, there is only a single threshold, which doesn’t allow for much granularity. What’s more, it can lead to perceived unfairness and could discourage players from attending events estimated to be just below 3,000 players.

It wasn’t always this way. From January 2014 – July 2014, prize money (at individual events) was paid out to:

  • 64th place if Grand Prix attendance was 1,199 or fewer participants.
  • 100th place if Grand Prix attendance was 1,200-2,399 participants.
  • 150th place if Grand Prix attendance was 2,400 or more participants.

Then, due to the growth of Grand Prix attendance, a change was announced. From August 2014-December 2015, prize money (at individual events) was paid out to:

  • 64th place if Grand Prix attendance was 1,199 or fewer participants.
  • 100th place if Grand Prix attendance was 1,200-2,399 participants.
  • 150th place if Grand Prix attendance was 2,400-3,599 participants.
  • 230th place if Grand Prix attendance was 3,600 or more participants.

Over a year later, the prize structure was revised again to award more money to the Top 4. In the process, we went from having three thresholds to a single threshold.

Since January 2016, prize money (at individual events) is paid out to:

  • 64th place if Grand Prix attendance is 2,999 or fewer participants.
  • 180th place if Grand Prix attendance is 3,000 or more participants.

There are advantages to having a single threshold at 3,000 people. For example, the payout structure is easy to understand, has less variance, and creates fewer “feel-bad” GPs that get close to a threshold but don’t actually meet it. But in my view, these advantages are outweighed by the disadvantages corresponding to the huge payout disparity that we saw in Shizuoka and Porto Alegre.

To make a concrete proposal for change, I went over all individual Grand Prix events from January 1, 2015 until March 19, 2017 (ignoring team GPs for simplicity) and noted down the final attendance number. In the following picture, each dot represents one GP.

GP size distribution

Out of these 103 Grand Prix tournaments, 13 had 2,400 or more participants. (The dots of a 2,566-player GP and a 2,571-player GP overlap in the picture.) Out of those 13 large events, five had 3,000 or more participants. Taking this distribution for granted, the expected prize money paid out at an arbitrary GP under the current prize structure is $51,335.

Assuming that the above GP attendance distribution is representative, I can now propose several budget-neutral alternatives. As far as I understand, WotC cannot provide a payout by record, possibly for legal or budgetary reasons, so I didn’t consider any such solutions. Instead, inspired by payout structures from 2014-2015, I tweaked the number and placement of thresholds, and then I adjusted payouts to keep the same expected payout.

Place Current system Proposal 1 Proposal 2 Proposal 3
1 $10,000 $10,000 $10,000 $10,000
2 $5,000 $5,000 $5,000 $5,000
3-4 $2,500 $2,500 $2,500 $2,500
5-8 $1,500 $1,350 $1,350 $1,500
9-16 $1,000 $1000 $1000 $1000
17-32 $500 $420 $500 $500
33-64 $250 $250 $250 $325 (if ≥1,200)
65-100 $250 (if ≥3,000) $200 (if ≥2,400) $200 (if ≥2,400) $260 (if ≥2,400)
101-180 $200 (if ≥3,000) $200 (if ≥2,400) $200 (if ≥3,000) $200 (if ≥3,600)
Expected payout $51,300 $51,300 $51,300 $51,300

The proposals differ in the number of thresholds. But all of them retain the same prizes for the Top 4, all of them pay out $200 to the 180th-place finisher at a GP with 3,600 or more participants, and all of them require the same expected budget. Most importantly, all of them would have awarded a prize to every 12-3 player at Grand Prix Shizuoka.

My preference would lie with proposal 3—I place a lot of value on a smooth progression of payouts—but I would be happy with all three, or any tweak. Ultimately, it is possible to increase the “fairness” of payouts in a budget-neutral way via relatively small changes to the current system.