BetWatch

BatWatch is a sports betting risk management solution.

1. Betwatcn

2. 20th-21st May 2016 9 La Térmica, Málaga, Spain

Bringing developers and DevOps together around Big Data

Get Tickets

3. BetWatch™

4. Areas that we wanted to address:
+ Risk Management
- Customer Engagement
- Fraud/Problematic Gamblers

vse were also hoping to:

5. - Frau ropiematic Gamblers

We were also hoping to:
- Increase cross-sell & Up-sell opportunities
« Reduce customer churn
- Increase revenue per customer
- Identify the true value of the customer
- Understand the customer's buying journey
+ Novelty detection

6. startup

7. ‘focus

9. EL

SIC EE

10. UU

2

Account ID | Segment) BETDATE

16Jan2009

11. DN Event:
ES Profit: (5641)

azo v Cagliari

$736

w
Athletic v
Liverpool

Tottenham Stoke City

Hotspur v v Liverpoo
Portsmouth E

Manchester
United v
Tottenham
Hotspur

Reading v
Watford

OF

Panionios

12. Powered by

The input metrics for clustering customers

+ LEAD_TIME: the average length of time (in h
that ome wager before the

starts.

13. The input metrics for clustering customers were:

« LEAD_TIME: the average length of time (in hours)
that a customer places a wager before the event
starts.

« MID_STAKE: the median stake that a customer
wagers (average not used due to right skewed
nature of this metric).

+ MID_ODDS: the median odds that a customer
takes when wagering (average not used due to right
skewed nature of this metric).

+ TURNOVER: Total amount wagered by customer
in AUD.

« RETURN: Total amount won or lost by customer in
AUD.

« ROI: Return on investment for customer as a
percentage = RETURN/TURNOVER.

+ BETTYPE_COUNT: Number of unique bet type
customer has wagered on.

14. * TURNOVER: Total amount wagered by customer

in AUD,
+ RETURN: Total amount won or lost by customer in
AUD.

+ ROI: Return on investment for customer as a
percentage = RETURN/TURNOVER,

* BETTYPE_COUNT: Number of unique bet type
customer has wagered on.

sl os | : |
.3

E
E)
9
4
al

im
7
ee IEEE INES 2770770 [a DEEE

15. Profit (millions)

‘ny ters van oe

Son ‘a Cluster 1 -- Moderate
‘atom ing Sized, early betors Cluster 3 ~ Biggest

Cluster 4 = Small,

Sime" wor shat ho ber on- al bet > rn betors, _ specalzing. … apparent
a ves ‘ypes ol offering and i ana few bet ypes and recreational punters,
EA Oi mon spot pes A DD or win à long vito bet on anything
a ee teadsime on wagers. and most bet types.
> potential arbtrages.
0 1 2 3 4 Final
Segment

O1Jan1 900 2

Segment

17. A …
Cluster 1 -- Moderate
sized, early bettors,
-who- bet on- all- bet-
types of offering and
on most sport types.
These _ are _ your
potential arbitrages.

19. Cluster 3 -- Biggest
_bettors, _ specializing.
on a few bet types and
sports, with a long
lead-time on wagers.

20. Cluster 4 -- Small,
apparently _____
recreational punters,
who bet on anything
and most bet types.

0 0 Ee A +

21. € > © D wire S8S3hnapm44 ap-southeast-1 compute internal

DIE oy =

L L „v2
AS [ee | Cm were: ri

22. Beiwatch
= 2.0 vato

23. @OnlineBettingTM OONLINEBETTING em

HOW US SPORTSBOOKS

AN INVESTIGATION INTO US SPORTSBOOKS AND PROFESSIONAL HANDICAPPERS
1. HOW US SPORTSBOOKS MAKE MONEY:

Sportsbooks make money by keeping the player's losses

and charging a Vig on all winning bets.
man sense
-250 +/-0 +200

MONEY LINE

24. "BETTER HR

fe H TENDENCY
= i E

a i Shane ss

Sortsbooks often shade the lines by 2-3% on the favorites
in order to charge more to the public:?

es ws

MONEY LINE

“Sportsbooks set the odds based on the likelihood of the outcome,
Prezhowever they also have to tailor the lines to their own pool of players.”

25. 5. AVERAGE WINNING MARGIN% FOR VEGAS SPORTSBOOK
FROM 1992 - 20111

SPORTSBOOK WIN PERCENTAGE ON...

aa =
y =

o

"Baseball has the smallest win percentage, however because there
are so many games it has the highest value for sportsbooks”

Gk 18

MILLION"

26. Risk Management

28. @OnlineBettingTM OONLINEBETTING

HOW US SPORTSBOOKS

AN INVESTIGATION INTO US SPORTSBOOKS AND PROFESSIONAL HANDICAPPERS
1. HOW US SPORTSBOOKS MAKE MONEY:

Sportsbooks make money by keeping the player's losses

and charging a Vig on all winning bets.
mens casinos
-250 +/-0 +200

PR PRES PEPA PP

MONEY LINE

29. Real-time betting/rejecting

J

31. Multi-leg Parlays

Time to hedge your position:
Cezar Ferreira -200 odds
Edson Barboza -150 odds
Antonia Silva - 150 odds

32. UFC costs sportsbook nearly $1M

Darren Rovell Feb 23, 2015
ESPN Senior Writer

Aremarkable number of upsets at UFC Fight Night 61 resulted in a historic
beating for the sportsbook Bovada.LV.

Frank Mir's victory over Antonio Silva in the main event in Porto Alegre,
Brazil, on Sunday night marked the 10th win by an underdog, costing
Bovada nearly $1 million, according to the sportsbook's manager.

200058

Underdogs went 10-1, and as a result, parlay bets, in which gamblers bet
between $1 and $5 on all underdogs, hit big.

According to Bovada sportsbook manager Kevin Bradley, one bettor had a
$4 all-underdog parlay that paid $85,000. Another had four $1 parlays that
paid $25,000 each.

"We had at least a dozen of these bets that paid out $10,000 or more,"
Bradley said.

33. Leg Fight Favorite | Long shot | Decimal odds [Fractional odds] Payout
1|Cody Gibson (-210) vs. Douglas Silva (+160) $ 1.60
2| Tiago dos Santos (-190) vs. Mike de la Torre (+150) $ 5.50
3| William Macario (-260) vs. Matt Dwyer (+180) $ 17.20
4| Jessica Andrade (-245) vs. Marion Reneau (+175) 5 49.05
5|Sean Strickland (-210) vs. Santiago Ponzinibbio (+160) $ 12913
6] luri Alcantara (-315) vs. Frankie Saenz (+235) $ 434.94
7|Rustam Khabilov (-300) vs. Adriano Martins (+220) $ 1,393.99
8|Cezar Ferreira (-265) vs. Sam Alvey (+185) $ 3,973.73
9] Edson Barboza (-175) vs. Michael Johnson (+135) $ 9,341.97

10] Antonio Silva (-240) vs. Frank Mir (+200) $28,028.91

34. Time to hedge your position:
Cezar Ferreira -200 odds
Edson Barboza -150 odds
Antonia Silva - 150 odds

36. 4 l facebook © o

38. Most important channels

89% of all the mentions across the
sportsbook industry are tweets

© 3% Facebook
Twitter © 3% News sites
e 3% Forums
89% e 1% Blogs

@ 1% General websites

All the brands that we analysed had Twitter as
their largest and most impactful channel

42. The offer was extended to anyone following the
company’s official Snapchat account, betfairoffcial
Betfair sent images that contained a custom URL,
giving users just a few seconds to read it and access
the offer, The first 500 customers to place a bet on the
URL benefited from Bettair's enhanced odds.

43. _ facebook.

45. @ Message Us

© Send to Messenger

48. =

49. Presented by | . whe 3
Ey) amazoncoins 74% <n
and d YA >
RM : y.

, \

be. 4

51. Games

M Cramos MASGame
CR Porta la gente quese mete con
1 demas por su sio

pas Toe nc con ei cs
ia uen en user canal oe out

Mamma eso)

onto: Grenada ane paar enon yes
propos.

Lento: Garner,"

cundo ss han jgado eta parts?

54. Problem/
fraudulent
Gambling

55. 20th-21st May 2016 9 La Térmica, Málaga, Spain

Bringing developers and DevOps together around Big Data

Get Tickets