Methodology

The same approach for Rock Werchter, Pukkelpop, Best Kept Secret and Pinkpop, based on public sources. Below: what the numbers mean, where the data comes from and how well the chances held up afterwards.

The numbers

Points
How big an act is today, measured with today's figures, only for the current edition. De standaardvolgorde op de pagina van de lopende editie.

Historical score (0 to 100)
How big an act was at the time of its edition, measured in the twelve months before each edition, so 2017 and 2027 can sit side by side.

Day strength
How strong a festival day is: the acts of one day together, by their level (A, B, C, D) relative to the festival.

Sources

All sources, with what we use from each source and a link, are on the page Sources.

Confirmations and contenders: the chances and how accurate they are

Every act not yet confirmed gets a chance, based on the festival's booking history and what is known about the act (concerts, records, radio, other festivals). No rumours: what circulates on forums does not count. Every previous edition was predicted without knowing that edition; the table shows per class how often an act was then actually booked. Explained in plain words on How good are the chances?

Rock Werchter

Class	Acts in the test	Actually played	Share	Average chance
Very likely	101	70	69%	69%
Likely	201	87	43%	43%
Possible	906	226	25%	25%
Unlikely	804	84	10%	11%
Very unlikely	2.012	95	5%	5%
(Nearly) impossible	76.444	195	0%	0%

Pukkelpop

Class	Acts in the test	Actually played	Share	Average chance
Very likely	101	59	58%	58%
Likely	401	159	40%	40%
Possible	1.104	243	22%	22%
Unlikely	1.523	184	12%	12%
Very unlikely	4.900	222	5%	5%
(Nearly) impossible	72.258	484	1%	1%

Best Kept Secret

Class	Acts in the test	Actually played	Share	Average chance
Likely	100	36	36%	36%
Possible	400	89	22%	22%
Unlikely	700	74	11%	11%
Very unlikely	2.000	100	5%	5%
(Nearly) impossible	76.800	398	1%	1%

Pinkpop

Class	Acts in the test	Actually played	Share	Average chance
Likely	101	41	41%	41%
Possible	100	31	31%	31%
Unlikely	1.002	114	11%	11%
Very unlikely	1.203	64	5%	5%
(Nearly) impossible	77.763	200	0%	0%

Toets over de edities 2017, 2018, 2019, 2022, 2023, 2024, 2025, 2026; laatst gedraaid 04/11/2026. Een goed geijkt model heeft een aandeel dat dicht bij de gemiddelde kans ligt.

Pulling power: expected audience and stage

How much audience did a festival expect from an act? Learned from the real stages and time slots since 2016 and applied to the new acts. The test predicts every edition without knowing that edition.

Edition	Acts	Correlatie model	Correlatie score	Correct level	Within one level
BKS2016	76	0.61	0.67	62%	99%
PKP2016	214	0.51	0.58	84%	99%
PP2016	52	0.69	0.67	50%	96%
RW2016	74	0.58	0.63	57%	93%
BKS2017	77	0.62	0.65	66%	96%
PKP2017	160	0.67	0.64	82%	97%
PP2017	51	0.62	0.52	51%	98%
BKS2018	97	0.58	0.61	56%	99%
PKP2018	154	0.67	0.72	82%	98%
PP2018	48	0.71	0.8	62%	96%
RW2018	93	0.68	0.66	65%	99%
BKS2019	111	0.55	0.69	47%	100%
PKP2019	175	0.64	0.66	83%	97%
PP2019	52	0.79	0.73	62%	98%
RW2019	95	0.68	0.76	66%	97%
BKS2022	72	0.68	0.53	65%	99%
PKP2022	214	0.67	0.64	87%	98%
PP2022	56	0.77	0.72	79%	96%
RW2022	100	0.8	0.84	71%	100%
BKS2023	92	0.58	0.36	60%	96%
PKP2023	158	0.7	0.66	87%	97%
PP2023	58	0.84	0.84	72%	100%
RW2023	98	0.74	0.71	66%	98%
BKS2024	99	0.47	0.43	45%	95%
PP2024	68	0.74	0.71	60%	97%
RW2024	100	0.79	0.79	67%	100%
BKS2025	96	0.58	0.65	47%	100%
PKP2025	165	0.48	0.41	80%	97%
RW2025	101	0.76	0.7	68%	99%
BKS2026	96	0.56	0.66	54%	100%
PKP2026	169	0.53	0.49	85%	99%
PP2026	63	0.76	0.83	70%	97%
RW2026	100	0.81	0.8	68%	99%