Weekly leaderboard · Week 20 results (May 11 – May 17, 2026) · published Week 21

47 predictors placed 647 settled wagers this week (Week 20 results (May 11 – May 17, 2026) · published Week 21). R0Kk-_ led on accuracy at 67.6% Wilson, while tier0_s netted +$701.30 in profit. Average accuracy across the field: 26.5%.

🎯 Win Rate Champion

Highest accuracy by Wilson lower bound (3+ settled wagers required to qualify).

🥇

R0Kk-_

Win rate champion · 8 settled wagers

67.6%

Wilson WR

8

Settled

8

Correct

0

Wrong

1. 🥇 R0Kk-_ — 67.6% Wilson (8W / 0L)
2. 🥈 tatazin — 57.8% Wilson (11W / 2L)
3. 🥉 seberya — 52.1% Wilson (21W / 9L)

💵 Profit Champion

Highest net USD return (winnings minus stakes). Green bars are winners, red are net losers.

🥇

tier0_s

Profit champion · 30 settled wagers

+$701.30

Net profit

39.2%

Win rate

17

Won

13

Lost

1. 🥇 tier0_s — +$701.30 (30 bets)
2. 🥈 k1to — +$388.22 (17 bets)
3. 🥉 souv — +$276.51 (13 bets)

Weekly digest · Top-tier CS2 matches · Week 21 · May 11 – May 17, 2026

This week (Week 21 · May 11 – May 17, 2026) we analyzed 57 top-tier CS2 matches across 25 prediction models. STRONGHOLD dominated with 81.8% accuracy (9/11 correct), ahead of TIER at 80.0%. STRONGHOLD achieved a Brier Score of 0.1542 and Log Loss of 0.4943.

🏆 Model of the Week: STRONGHOLD

STRONGHOLD had an outstanding week, correctly predicting 9 out of 11 matches.

🥇

STRONGHOLD

ELO + SOS both >60% same direction

STRONGHOLD showed excellent performance with 81.8% accuracy across 11 predictions (9 correct, 2 wrong).

81.8%

Accuracy

11

Predictions

9

Correct

2

Wrong

Weekly breakdown

W20: 81.8% (11)

📊 Full Rankings This Week

Complete model performance for the week. The leader is highlighted.

Rank	Model	Accuracy	Predictions	Correct	Wrong
#1	STRONGHOLD	81.8%	11	9	2
#2	TIER	80.0%	25	20	5
#3	TIER3	80.0%	20	16	4
#4	TIER2	77.8%	27	21	6
#5	WR10	75.0%	44	33	11
#6	MIND	70.2%	47	33	14
#7	WR10C	69.8%	53	37	16
#8	ELOHQ	69.4%	36	25	11
#9	OPEN	68.1%	47	32	15
#10	WR10P	66.7%	51	34	17
#11	TRIBUNAL	66.7%	9	6	3
#12	SQUAD	66.0%	53	35	18
#13	FORM	65.3%	49	32	17
#14	XGPRO	64.5%	31	20	11
#15	MAJOR	64.5%	31	20	11
#16	STREAK	63.6%	33	21	12
#17	ELO	63.5%	52	33	19
#18	ELOFACE	62.7%	51	32	19
#19	SOS	55.1%	49	27	22
#20	APEX	52.8%	53	28	25
#21	PHANTOM	52.8%	53	28	25
#22	HIGH-ORACLE	52.6%	19	10	9
#23	NEXUS	50.0%	16	8	8
#24	ORACLE	47.1%	51	24	27
#25	EAGLE	40.4%	47	19	28

Weekly digest · Top-tier CS2 matches · Week 20 · May 04 – May 10, 2026

This week (Week 20 · May 04 – May 10, 2026) we analyzed 16 top-tier CS2 matches across 25 prediction models. STRONGHOLD dominated with 100.0% accuracy (8/8 correct), ahead of HIGH-ORACLE at 100.0%. STRONGHOLD achieved a Brier Score of 0.0677 and Log Loss of 0.2985.

🏆 Model of the Week: STRONGHOLD

STRONGHOLD had an outstanding week, correctly predicting 8 out of 8 matches.

🥇

STRONGHOLD

ELO + SOS both >60% same direction

STRONGHOLD showed excellent performance with 100.0% accuracy across 8 predictions (8 correct, 0 wrong).

100.0%

Accuracy

8

Predictions

8

Correct

0

Wrong

Weekly breakdown

W19: 100.0% (8)

📊 Full Rankings This Week

Complete model performance for the week. The leader is highlighted.

Rank	Model	Accuracy	Predictions	Correct	Wrong
#1	STRONGHOLD	100.0%	8	8	0
#2	HIGH-ORACLE	100.0%	7	7	0
#3	NEXUS	100.0%	7	7	0
#4	SENTINEL	100.0%	5	5	0
#5	ELO	81.2%	16	13	3
#6	ELOFACE	81.2%	16	13	3
#7	WR10C	81.2%	16	13	3
#8	APEX	81.2%	16	13	3
#9	PHANTOM	81.2%	16	13	3
#10	ORACLE	80.0%	15	12	3
#11	TIER2	80.0%	15	12	3
#12	ELOHQ	80.0%	15	12	3
#13	TIER	78.6%	14	11	3
#14	EAGLE	75.0%	16	12	4
#15	TIER3	75.0%	12	9	3
#16	XGPRO	71.4%	7	5	2
#17	MAJOR	71.4%	7	5	2
#18	SQUAD	68.8%	16	11	5
#19	SOS	66.7%	15	10	5
#20	FORM	66.7%	12	8	4
#21	STREAK	66.7%	3	2	1
#22	OPEN	60.0%	10	6	4
#23	WR10	54.5%	11	6	5
#24	WR10P	50.0%	16	8	8
#25	MIND	23.1%	13	3	10

Weekly leaderboard · Week 19 results (May 04 – May 10, 2026) · published Week 20

30 predictors placed 505 settled wagers this week (Week 19 results (May 04 – May 10, 2026) · published Week 20). Tw1nk1e17 led on accuracy at 49.1% Wilson, while Goofy netted +$413.28 in profit. Average accuracy across the field: 28.0%.

🎯 Win Rate Champion

Highest accuracy by Wilson lower bound (3+ settled wagers required to qualify).

🥇

Tw1nk1e17

Win rate champion · 18 settled wagers

49.1%

Wilson WR

18

Settled

13

Correct

5

Wrong

1. 🥇 Tw1nk1e17 — 49.1% Wilson (13W / 5L)
2. 🥈 sliimey — 47.3% Wilson (15W / 7L)
3. 🥉 AZIDOT — 45.4% Wilson (14W / 7L)

💵 Profit Champion

Highest net USD return (winnings minus stakes). Green bars are winners, red are net losers.

🥇

Goofy

Profit champion · 13 settled wagers

+$413.28

Net profit

29.1%

Win rate

7

Won

6

Lost

1. 🥇 Goofy — +$413.28 (13 bets)
2. 🥈 sliimey — +$271.94 (22 bets)
3. 🥉 lenci — +$236.79 (13 bets)

Monthly leaderboard · April 2026 results · published May 2026

30 predictors placed 404 settled wagers this month (April 2026 results · published May 2026). Viresse led on accuracy at 56.5% Wilson, while lenci netted +$316.68 in profit. Average accuracy across the field: 27.5%.

🎯 Win Rate Champion

Highest accuracy by Wilson lower bound (3+ settled wagers required to qualify).

🥇

Viresse

Win rate champion · 9 settled wagers

56.5%

Wilson WR

9

Settled

8

Correct

1

Wrong

1. 🥇 Viresse — 56.5% Wilson (8W / 1L)
2. 🥈 d1maje — 52.7% Wilson (13W / 4L)
3. 🥉 AZIDOT — 48.1% Wilson (14W / 6L)

💵 Profit Champion

Highest net USD return (winnings minus stakes). Green bars are winners, red are net losers.

🥇

lenci

Profit champion · 16 settled wagers

+$316.68

Net profit

28.0%

Win rate

8

Won

8

Lost

1. 🥇 lenci — +$316.68 (16 bets)
2. 🥈 Viresse — +$243.92 (9 bets)
3. 🥉 tripex17 — +$192.32 (16 bets)

Weekly leaderboard · Week 18 results (Apr 27 – May 03, 2026) · published Week 19

30 predictors placed 653 settled wagers this week (Week 18 results (Apr 27 – May 03, 2026) · published Week 19). AZIDOT led on accuracy at 55.3% Wilson, while lenci netted +$417.81 in profit. Average accuracy across the field: 31.2%.

🎯 Win Rate Champion

Highest accuracy by Wilson lower bound (3+ settled wagers required to qualify).

🥇

AZIDOT

Win rate champion · 27 settled wagers

55.3%

Wilson WR

27

Settled

20

Correct

7

Wrong

1. 🥇 AZIDOT — 55.3% Wilson (20W / 7L)
2. 🥈 tripex17 — 48.4% Wilson (17W / 8L)
3. 🥉 Viresse — 48.1% Wilson (14W / 6L)

💵 Profit Champion

Highest net USD return (winnings minus stakes). Green bars are winners, red are net losers.

🥇

lenci

Profit champion · 28 settled wagers

+$417.81

Net profit

32.6%

Win rate

14

Won

14

Lost

1. 🥇 lenci — +$417.81 (28 bets)
2. 🥈 tripex17 — +$365.94 (25 bets)
3. 🥉 Viresse — +$329.79 (20 bets)

Monthly digest · Top-tier CS2 matches · May 2026

Based on April 2026 data, we tracked 174 top-tier CS2 matches across 27 prediction models with a total of 2520 predictions. The average model accuracy was 72.3%, with a spread of 53.0% between the best and worst performing models.

🏆 Top 3 Models of the Month

TRIBUNAL takes the top spot with 100.0% accuracy, followed by NEXUS at 97.1% and SENTINEL at 92.3%.

🥇

TRIBUNAL

OPEN + PRO + ORACLE all >55% same direction

TRIBUNAL showed excellent performance with 100.0% accuracy across 20 predictions (20 correct, 0 wrong).

100.0%

Accuracy

20

Predictions

20

Correct

0

Wrong

Weekly breakdown

W14: 100.0% (9)W15: 100.0% (5)W16: 100.0% (4)W18: 100.0% (2)

🥈

NEXUS

APEX + ORACLE both >65% same direction

NEXUS showed excellent performance with 97.1% accuracy across 35 predictions (34 correct, 1 wrong). Best week: W14 at 100.0%, weakest: W18 at 75.0%.

97.1%

Accuracy

35

Predictions

34

Correct

1

Wrong

Weekly breakdown

W14: 100.0% (10)W15: 100.0% (2)W16: 100.0% (10)W17: 100.0% (9)W18: 75.0% (4)

🥉

SENTINEL

Anti-signal: SOS >60% confirmed by Mind anti-prediction

SENTINEL showed excellent performance with 92.3% accuracy across 26 predictions (24 correct, 2 wrong). Best week: W15 at 100.0%, weakest: W14 at 90.5%.

92.3%

Accuracy

26

Predictions

24

Correct

2

Wrong

Weekly breakdown

W14: 90.5% (21)W15: 100.0% (1)W16: 100.0% (1)W17: 100.0% (3)

📊 Full Monthly Leaderboard

Below is the complete ranking of all models. Models highlighted in blue are in the top 3. The bar chart shows relative performance at a glance.

Rank	Model	Accuracy	Predictions	Correct	Wrong
#1	TRIBUNAL	100.0%	20	20	0
#2	NEXUS	97.1%	35	34	1
#3	SENTINEL	92.3%	26	24	2
#4	STRONGHOLD	91.1%	45	41	4
#5	TIER3	86.4%	44	38	6
#6	VANGUARD	85.7%	14	12	2
#7	HIGH-ORACLE	85.3%	34	29	5
#8	TIER	82.7%	75	62	13
#9	ELOHQ	79.8%	94	75	19
#10	TIER2	78.3%	83	65	18
#11	ORACLE	74.7%	91	68	23
#12	APEX	73.9%	92	68	24
#13	SOS	73.5%	136	100	36
#14	PHANTOM	73.3%	90	66	24
#15	ELOFACE	70.0%	130	91	39
#16	ELO	68.5%	143	98	45
#17	STREAK	65.6%	90	59	31
#18	WR10	63.7%	113	72	41
#19	EAGLE	62.6%	91	57	34
#20	MAJOR	61.2%	116	71	45
#21	XGPRO	60.9%	115	70	45
#22	WR10C	60.3%	156	94	62
#23	OPEN	60.3%	131	79	52
#24	SQUAD	57.2%	166	95	71
#25	FORM	50.4%	133	67	66
#26	WR10P	49.6%	125	62	63
#27	MIND	47.0%	132	62	70

Weekly digest · Top-tier CS2 matches · Week 19 · Apr 27 – May 03, 2026

This week (Week 19 · Apr 27 – May 03, 2026) we analyzed 20 top-tier CS2 matches across 24 prediction models. STRONGHOLD dominated with 100.0% accuracy (7/7 correct), ahead of STREAK at 91.7%. STRONGHOLD achieved a Brier Score of 0.0822 and Log Loss of 0.334.

🏆 Model of the Week: STRONGHOLD

STRONGHOLD had an outstanding week, correctly predicting 7 out of 7 matches.

🥇

STRONGHOLD

ELO + SOS both >60% same direction

STRONGHOLD showed excellent performance with 100.0% accuracy across 7 predictions (7 correct, 0 wrong).

100.0%

Accuracy

7

Predictions

7

Correct

0

Wrong

Weekly breakdown

W18: 100.0% (7)

📊 Full Rankings This Week

Complete model performance for the week. The leader is highlighted.

Rank	Model	Accuracy	Predictions	Correct	Wrong
#1	STRONGHOLD	100.0%	7	7	0
#2	STREAK	91.7%	12	11	1
#3	TIER3	77.8%	9	7	2
#4	XGPRO	77.8%	9	7	2
#5	MAJOR	77.8%	9	7	2
#6	NEXUS	77.8%	9	7	2
#7	ELO	76.5%	17	13	4
#8	ELOFACE	75.0%	16	12	4
#9	TIER	75.0%	12	9	3
#10	TIER2	75.0%	12	9	3
#11	SOS	73.7%	19	14	5
#12	APEX	72.2%	18	13	5
#13	ELOHQ	71.4%	14	10	4
#14	WR10	71.4%	14	10	4
#15	SQUAD	70.0%	20	14	6
#16	HIGH-ORACLE	70.0%	10	7	3
#17	OPEN	68.4%	19	13	6
#18	MIND	66.7%	18	12	6
#19	PHANTOM	66.7%	18	12	6
#20	ORACLE	63.2%	19	12	7
#21	WR10C	60.0%	20	12	8
#22	WR10P	60.0%	15	9	6
#23	FORM	57.9%	19	11	8
#24	EAGLE	41.2%	17	7	10

Weekly digest · Top-tier CS2 matches · Week 18 · Apr 20 – Apr 27, 2026

This week (Week 18 · Apr 20 – Apr 27, 2026) we analyzed 13 top-tier CS2 matches across 24 prediction models. NEXUS dominated with 100.0% accuracy (9/9 correct), ahead of HIGH-ORACLE at 100.0%. NEXUS achieved a Brier Score of 0.0664 and Log Loss of 0.2942.

🏆 Model of the Week: NEXUS

NEXUS had an outstanding week, correctly predicting 9 out of 9 matches.

🥇

NEXUS

APEX + ORACLE both >65% same direction

NEXUS showed excellent performance with 100.0% accuracy across 9 predictions (9 correct, 0 wrong).

100.0%

Accuracy

9

Predictions

9

Correct

0

Wrong

Weekly breakdown

W17: 100.0% (9)

📊 Full Rankings This Week

Complete model performance for the week. The leader is highlighted.

Rank	Model	Accuracy	Predictions	Correct	Wrong
#1	NEXUS	100.0%	9	9	0
#2	HIGH-ORACLE	100.0%	8	8	0
#3	ELOHQ	100.0%	4	4	0
#4	TIER	100.0%	4	4	0
#5	TIER2	100.0%	4	4	0
#6	SENTINEL	100.0%	3	3	0
#7	STRONGHOLD	100.0%	3	3	0
#8	PHANTOM	90.0%	10	9	1
#9	ORACLE	81.8%	11	9	2
#10	EAGLE	81.8%	11	9	2
#11	ELOFACE	77.8%	9	7	2
#12	SQUAD	75.0%	12	9	3
#13	SOS	75.0%	12	9	3
#14	APEX	75.0%	12	9	3
#15	ELO	72.7%	11	8	3
#16	WR10C	46.2%	13	6	7
#17	FORM	36.4%	11	4	7
#18	OPEN	30.0%	10	3	7
#19	XGPRO	25.0%	8	2	6
#20	MAJOR	25.0%	8	2	6
#21	WR10	25.0%	8	2	6
#22	WR10P	22.2%	9	2	7
#23	MIND	16.7%	12	2	10
#24	STREAK	0.0%	3	0	3

Weekly digest · Top-tier CS2 matches · Week 17 · Apr 13 – Apr 20, 2026

This week (Week 17 · Apr 13 – Apr 20, 2026) we analyzed 30 top-tier CS2 matches across 25 prediction models. NEXUS dominated with 100.0% accuracy (10/10 correct), ahead of STRONGHOLD at 100.0%. NEXUS achieved a Brier Score of 0.0884 and Log Loss of 0.3524.

🏆 Model of the Week: NEXUS

NEXUS had an outstanding week, correctly predicting 10 out of 10 matches.

🥇

NEXUS

APEX + ORACLE both >65% same direction

NEXUS showed excellent performance with 100.0% accuracy across 10 predictions (10 correct, 0 wrong).

100.0%

Accuracy

10

Predictions

10

Correct

0

Wrong

Weekly breakdown

W16: 100.0% (10)

📊 Full Rankings This Week

Complete model performance for the week. The leader is highlighted.

Rank	Model	Accuracy	Predictions	Correct	Wrong
#1	NEXUS	100.0%	10	10	0
#2	STRONGHOLD	100.0%	7	7	0
#3	TRIBUNAL	100.0%	4	4	0
#4	TIER3	93.8%	16	15	1
#5	ELOHQ	90.9%	22	20	2
#6	TIER2	89.5%	19	17	2
#7	TIER	88.9%	18	16	2
#8	HIGH-ORACLE	87.5%	8	7	1
#9	APEX	85.2%	27	23	4
#10	SOS	80.0%	25	20	5
#11	ELO	79.3%	29	23	6
#12	ELOFACE	78.6%	28	22	6
#13	PHANTOM	78.6%	28	22	6
#14	ORACLE	78.6%	28	22	6
#15	WR10C	73.3%	30	22	8
#16	WR10P	73.1%	26	19	7
#17	EAGLE	72.0%	25	18	7
#18	STREAK	71.4%	14	10	4
#19	WR10	70.8%	24	17	7
#20	SQUAD	70.4%	27	19	8
#21	MAJOR	64.7%	17	11	6
#22	FORM	62.5%	24	15	9
#23	XGPRO	62.5%	16	10	6
#24	OPEN	60.0%	25	15	10
#25	MIND	56.5%	23	13	10

Weekly digest · Top-tier CS2 matches · Week 16 · Apr 06 – Apr 13, 2026

This week (Week 16 · Apr 06 – Apr 13, 2026) we analyzed 50 top-tier CS2 matches across 25 prediction models. NEXUS dominated with 100.0% accuracy (12/12 correct), ahead of STRONGHOLD at 100.0%. NEXUS achieved a Brier Score of 0.0893 and Log Loss of 0.3546.

🏆 Model of the Week: NEXUS

NEXUS had an outstanding week, correctly predicting 12 out of 12 matches.

🥇

NEXUS

APEX + ORACLE both >65% same direction

NEXUS showed excellent performance with 100.0% accuracy across 12 predictions (12 correct, 0 wrong).

100.0%

Accuracy

12

Predictions

12

Correct

0

Wrong

Weekly breakdown

W15: 100.0% (2)W16: 100.0% (10)

📊 Full Rankings This Week

Complete model performance for the week. The leader is highlighted.

Rank	Model	Accuracy	Predictions	Correct	Wrong
#1	NEXUS	100.0%	12	12	0
#2	STRONGHOLD	100.0%	9	9	0
#3	TRIBUNAL	100.0%	9	9	0
#4	TIER3	95.2%	21	20	1
#5	TIER2	86.2%	29	25	4
#6	TIER	85.7%	28	24	4
#7	HIGH-ORACLE	83.3%	12	10	2
#8	ELOHQ	82.9%	35	29	6
#9	APEX	77.8%	45	35	10
#10	ORACLE	77.3%	44	34	10
#11	PHANTOM	76.7%	43	33	10
#12	SOS	76.2%	42	32	10
#13	ELO	76.1%	46	35	11
#14	ELOFACE	74.4%	43	32	11
#15	WR10	71.1%	38	27	11
#16	STREAK	70.4%	27	19	8
#17	WR10C	67.3%	49	33	16
#18	WR10P	65.9%	44	29	15
#19	EAGLE	65.1%	43	28	15
#20	MAJOR	64.7%	34	22	12
#21	FORM	64.3%	42	27	15
#22	XGPRO	63.6%	33	21	12
#23	OPEN	61.0%	41	25	16
#24	SQUAD	59.6%	47	28	19
#25	MIND	59.5%	42	25	17

We've added 2 new consensus models to CS2PREDICT: NEXUS and TRIBUNAL. Both were discovered through live data analysis — optimizing over real pre-match predictions with zero data leakage.

What Are Consensus Models?

Unlike base models that compute predictions independently from raw team stats, consensus models act as filters on top of existing models. They only make a prediction when multiple models agree with high confidence. The result: fewer predictions, but significantly higher accuracy.

The New Models

NEXUS — APEX + ORACLE both >65% same direction

The sharpest double confirmation on the site. NEXUS only activates when both APEX and ORACLE are strongly aligned — meaning two independent multi-factor models see the same outcome with high conviction. In backtesting on live top-tier matches: 100% accuracy on 12 predictions.

TRIBUNAL — OPEN + PRO + ORACLE all >55% same direction

A three-model consensus built around ORACLE as the anchor, confirmed by both the OPEN and PRO predictors. These three models use different feature sets — when all three agree, the signal is remarkably consistent. In backtesting on live top-tier matches: 100% accuracy on 11 predictions.

Important Context

The 100% accuracy figures are from a small sample of top-tier matches — treat them as promising early signals, not guarantees. Both models will fire rarely by design: NEXUS requires two models at 65%+ confidence, TRIBUNAL requires three at 55%+. Expect 1–3 predictions per week on top-tier matches.

As live data accumulates over the coming weeks, you'll see real accuracy statistics on each model's page.

NEXUS and TRIBUNAL are available for Pro subscribers.

Weekly digest · Top-tier CS2 matches · Week 15 · Mar 30 – Apr 06, 2026

This week (Week 15 · Mar 30 – Apr 06, 2026) we analyzed 193 top-tier CS2 matches across 27 prediction models. NEXUS dominated with 100.0% accuracy (22/22 correct), ahead of TRIBUNAL at 100.0%. NEXUS achieved a Brier Score of 0.0861 and Log Loss of 0.3457.

🏆 Model of the Week: NEXUS

NEXUS had an outstanding week, correctly predicting 22 out of 22 matches.

🥇

NEXUS

APEX + ORACLE both >65% same direction

NEXUS showed excellent performance with 100.0% accuracy across 22 predictions (22 correct, 0 wrong).

100.0%

Accuracy

22

Predictions

22

Correct

0

Wrong

Weekly breakdown

W14: 100.0% (10)W15: 100.0% (2)W16: 100.0% (10)

📊 Full Rankings This Week

Complete model performance for the week. The leader is highlighted.

Rank	Model	Accuracy	Predictions	Correct	Wrong
#1	NEXUS	100.0%	22	22	0
#2	TRIBUNAL	100.0%	18	18	0
#3	HIGH-ORACLE	86.4%	22	19	3
#4	TIER3	84.1%	44	37	7
#5	VANGUARD	83.3%	24	20	4
#6	SENTINEL	82.1%	39	32	7
#7	STRONGHOLD	81.0%	58	47	11
#8	TIER	79.7%	79	63	16
#9	ORACLE	77.1%	70	54	16
#10	ELOHQ	76.9%	104	80	24
#11	TIER2	75.0%	88	66	22
#12	APEX	74.6%	71	53	18
#13	PHANTOM	72.9%	70	51	19
#14	SOS	67.5%	151	102	49
#15	ELOFACE	65.3%	144	94	50
#16	STREAK	64.5%	110	71	39
#17	EAGLE	64.3%	70	45	25
#18	ELO	64.1%	156	100	56
#19	WR10	64.0%	125	80	45
#20	WR10C	60.5%	167	101	66
#21	MAJOR	60.1%	138	83	55
#22	OPEN	60.0%	145	87	58
#23	XGPRO	59.9%	137	82	55
#24	SQUAD	54.3%	186	101	85
#25	FORM	50.3%	145	73	72
#26	WR10P	48.1%	133	64	69
#27	MIND	45.1%	142	64	78

What Are Predictions on CS2PREDICT?

CS2PREDICT lets you test your CS2 knowledge by predicting match outcomes. Use your account balance to place predictions on upcoming matches and earn payouts when you're right.

How It Works

1. Find a match — browse upcoming matches on the main page. Look for matches with the "PREDICT" buttons visible.
2. Choose your team — click PREDICT Team A or PREDICT Team B
3. Set your amount — enter how much you want to predict (min $1). Free tier: up to $50. Pro tier: up to $500.
4. Confirm — your prediction is locked in and your balance is deducted
5. Wait for results — when the match ends, payouts are calculated automatically

The Prediction Pool

All predictions on a match go into a shared pool. Below each match you'll see:

• Pool: $X — total amount predicted by all users
• X vs Y — how the pool is split between the two teams
• Odds (x1.5, x2.3, etc.) — your potential payout multiplier

How Odds Work

Odds are calculated from the pool distribution, not from AI predictions:

• If 80% of the pool is on Team A, Team A pays x1.25 (low payout, popular pick)
• Team B would pay x5.0 (high payout, underdog pick)
• The less popular the pick, the higher the potential payout

Odds update in real-time as more users place predictions.

Payouts

When the match completes:

• Correct prediction — you receive your stake multiplied by the odds. Example: $10 at x2.0 odds = $20 payout ($10 profit).
• Wrong prediction — you lose your stake. It goes to the winners' pool.
• Payouts are processed automatically within minutes of the match result.

Prediction Locking

Predictions lock when the match starts (goes LIVE). You cannot change or cancel a prediction after it's locked. Plan ahead — place your predictions before match start.

Your Balance

Check your balance in the top-right corner of the site. To add funds:

1. Go to Balance page
2. Send USDT (TRC-20) to the provided wallet address
3. Deposits are detected automatically (checked every 5 minutes)

Subscription Tiers

See Subscription page for current pricing.

Tips

• Check model consensus before predicting — when most AI models agree, the outcome is more likely
• Watch the odds — sometimes the crowd is wrong, and the underdog pays well
• Start small — learn the system with $1-5 predictions before going bigger
• Diversify — don't put everything on one match

Where to See Your Predictions

Your active predictions appear on each match card (below the pool info). Past results are visible on the Results page — check which of your predictions were correct.

How to Use Match Predictions

Every upcoming match on CS2PREDICT shows predictions from multiple AI models. This guide explains how to read the match page and make the most of the data.

The Match Card

Each match card on the main page shows:

• Teams — with logos, HLTV rankings, and clickable names leading to team profiles
• Lineups — 5 players per team with photos, each clickable to their player profile
• Time — match start time (converted to your local timezone)
• Event — tournament name with star rating (more stars = bigger event)
• Predictions — model outputs shown as percentage cards on the right

Reading the Prediction Cards

Each small card represents one AI model's prediction:

• Top label — model name (e.g., ORACL, APEX, TIER)
• Percentage — how confident the model is in the predicted winner (always shown as the higher side, e.g., 67%)
• Bottom label — which team the model predicts to win
• Lock icon — model requires Basic or Pro subscription

Color Coding

Prediction percentages are color-coded:

• Green (70%+) — high confidence, the model sees a clear advantage
• Yellow (60-70%) — moderate confidence, edge exists but upset possible
• Red/Gray (50-60%) — low confidence, close match

Predictions between 48-52% are hidden (shown as 50/50) because they carry no useful signal.

The Consensus Bar

Below the team names, a colored bar shows the overall consensus across all models:

• If the bar is mostly on one side (e.g., 70/30), most models agree — stronger signal
• If it's close to 50/50, models disagree — uncertain match, proceed with caution

Match Detail Page

Click any match to see the full detail page with:

• All model predictions — not just the top 15
• Model accuracy badges — how accurate each model has been historically
• Player lineups — clickable to individual player profiles with tactical radars
• Team links — clickable to team pages with roster analysis and match history
• Wager pool — community predictions with odds

After the Match

Once a match completes:

• The result appears on the Results page
• Each prediction is scored — correct (green check) or wrong (red cross)
• Model accuracy on the leaderboard updates immediately
• ELO ratings recalculate, improving future predictions

Tips for Best Results

1. Follow consensus — when 15+ models agree, they're usually right
2. Check the leaderboard — not all models are equal, some are consistently better
3. Watch for model disagreement — if TIER says 75% Team A but MIND says 60% Team B, the match is genuinely uncertain
4. Consider the tier — we only show predictions for Top & Major matches where accuracy is highest
5. Look at player profiles — check the tactical radar to understand each team's playstyle

Overview

CS2PREDICT uses multiple independent AI models to predict the outcome of professional CS2 matches. Each model analyzes different aspects of the game — from pure skill ratings to psychological momentum to team chemistry. No single model is perfect, but together they provide a comprehensive view of each matchup.

How predictions are generated

When a new match appears on HLTV, our system automatically:

1. Identifies both teams and their current rosters (5 players each)
2. Loads player history — recent match stats, ratings, form trends from our database
3. Computes 50+ features per match — ELO ratings, momentum, tilt, head-to-head record, roster stability, strength of schedule, and more
4. Runs all models — each model outputs a probability (e.g., "Team A has 67% chance to win")
5. Displays results — predictions refresh every 5 minutes as new data arrives

Model categories

What happens after a match

Results are checked every 10 minutes. When a match completes:

• Each model's prediction is scored as correct or incorrect
• ELO ratings update based on the result (winners gain, losers drop)
• Model accuracy on the leaderboard updates in real-time
• All upcoming matches are re-predicted with updated ratings

Prediction tiers

We classify matches into tiers based on team rankings and tournament importance:

• Top — matches between ranked teams at major events. Best prediction accuracy.
• Major — major qualifier and playoff matches.
• Other — lower-tier matches. Shown as schedule only, without predictions, because accuracy drops to ~50%.

Data sources

All data comes from HLTV.org — the authoritative source for professional CS2 statistics. We collect match results, player box scores, team rankings, and tactical profiles (Rating 3.0 radar). Data refreshes every 5 minutes for matches and daily for player profiles.

The Tactical Radar

Each player profile on CS2PREDICT shows a tactical radar with 7 metrics scored from 0 to 100. These come from HLTV's Rating 3.0 system and represent the player's style and strengths over the past 3 months.

Metrics explained

Reading team compositions

A well-rounded team typically has:

• 1-2 high Firepower players (star fraggers)
• 1 high Entrying player (entry fragger)
• 1 high Sniping player (AWPer)
• 1-2 high Utility players (support/IGL)

Teams with balanced radar across all players tend to be more consistent. Teams with one extremely high firepower player and low stats elsewhere are "star-dependent" — if the star underperforms, the team collapses.

Example

See a live player profile: ZywOo — one of the highest-rated players in CS2 history. Notice his extreme Firepower and Sniping scores.

Data freshness

Tactical profiles update daily at 02:00 UTC. The metrics reflect the last 3 months of competitive play on HLTV.

We've added 4 new prediction models to CS2PREDICT, bringing the total to 21 AI models. Each was discovered through automated optimization across 3,400+ feature combinations tested against 667 completed matches.

The New Models

APEX — ELO + Form + Streak + Strength of Schedule
Balanced model optimized for the best accuracy-to-coverage ratio. Combines team rating with current momentum, win/loss streaks, and opponent difficulty. 66% accuracy, 60% coverage.

ORACLE — ELO + ELOHQ + Form + Tilt + SOS
Five-factor model that adds psychological pressure (tilt) and dual ELO systems. Performs best when teams have clear form differences. 69% accuracy, 51% coverage.

EAGLE — Rank + ELO + ELOHQ + Form + WinRate
High-confidence model that combines three ranking signals with team form. Predicts less frequently but with strong conviction. 77% accuracy, 14% coverage.

PHANTOM — ELO + TierELO + WinRate + Tilt
Maximum accuracy model. Only makes predictions when all signals align strongly. When PHANTOM speaks, it's almost always right. 92% accuracy, 6% coverage.

Calibration Period

All four models are currently in calibration mode — they need at least 10 confident predictions before appearing in the model rankings. You can already see their predictions on individual match pages.

The new models are available exclusively for Pro subscribers.

What is the Model Leaderboard?

The Model Leaderboard ranks all AI models by their prediction accuracy on Top-tier matches. It's a live scoreboard — accuracy updates after every completed match.

Color coding

Models are colored on a gradient from green (best accuracy) to red (worst). The color updates dynamically as accuracy changes — if a model goes on a hot streak, it moves up and turns greener.

Tier filters

• Top — accuracy on top-tier matches only (most reliable signal)
• Major — accuracy on major event matches
• All — combined Top + Major accuracy

Calibrating models

New models show a pulsing orange dot and "calibrating" status until they have at least 10 predictions. This prevents ranking models based on too few data points.

Which model should I follow?

No single model is always right. The best strategy is to look at consensus — when most models agree, the prediction is usually reliable. When models disagree sharply, the match is genuinely uncertain.

Check the leaderboard regularly — the best model this week might not be the best next week. Form matters for models too.