Smartgambler
Pro-Punter

Go Back   OZmium Sports Betting and Horse Racing Forums > Public Forums > Sports and Gambling
User Name
Password
Register FAQ Search Today's Posts Mark all topics as read

To advertise on these
forums, e-mail us.

Reply
 
Thread Tools Search this Thread Display Modes
  #31  
Old 16th April 2015, 08:57 AM
evajb001 evajb001 is offline
Member
 
Join Date: Oct 2009
Posts: 1,298
Default

walkermac,

When your attempting to predict a teams score each round I think home/away advantage needs to be taken into account. My ratings system simply determines which team to back at the line so I don't take home/away advantage into account because the line already does that for me to some extent.

My view from betting in and watching plenty of sports is that the home advantage is very team and geographically specific. Here are some examples:

In the AFL the swans play home games at both the SCG and ANZ Stadium. They have a much higher home advantage at the SCG. Some people may question why? Is it more fans attend the SCG? Is it because they train at/near the SCG? Either way it remains something to be aware of, you can't simply give swans an advantage on all home games, but have to take into account the venue and possibly who they are playing against.

Next example, in the NBA you have the Denver Nuggets who play/train 1,600 metre's above sea level. I've heard of simple betting systems that you bet on the nuggets whenever a team is playing the second game of a back to back in Denver simply because teams struggle with the thinner air. So once again its a distinct home advantage based on geography/venue.

Same goes for Fremantle/West Coast in the AFL, they travel every second week every year and become accustomed to flying all around australia for games. Victorian teams actually leave victoria maybe 5-6 times a year and even then it might only be to adelaide or sydney once or twice etc. So you'll find Freo/eagles have a distinct home advantage.

I'm not as familiar with NRL but certainly seems to me like you can find where there are distinct home advantages like warriors home games, or north qld at home vs the storm maybe given the travel distance etc.

Anyway just some food for thought, really enjoying this thread walkermac.
Reply With Quote
  #32  
Old 16th April 2015, 01:04 PM
walkermac walkermac is offline
Member
 
Join Date: Nov 2013
Posts: 995
Default

Quote:
Originally Posted by evajb001
When your attempting to predict a teams score each round I think home/away advantage needs to be taken into account.

Actually, I made an error with the reporting in my last post. The figures stated *did* take into account a team-specific home/away advantage - I forgot to re-run the test process without it. When considering simply the league-wide advantage it's all much of a muchness though:

2013-14: 63% vs 64.1%
2012: 64.1% vs 63%
2015 ('til Round 5): 52.5% vs 50% (i.e. 1 extra successful tip)

The team-specific home advantages I found (which were added to the league-wide advantage) are as follows:
Code:
Brisbane Broncos 0.462873286 Canberra Raiders -0.853890103 Canterbury-Bankstown Bulldogs -0.203583072 Cronulla-Sutherland Sharks -2.063045349 Gold Coast Titans -2.234931114 Manly-Warringah Sea Eagles 1.032411063 Melbourne Storm -0.6696935 New Zealand Warriors 0.603496144 Newcastle Knights 1.369869481 North Queensland Cowboys -0.375467596 Parramatta Eels 4.033649859 Penrith Panthers -0.771080275 South Sydney Rabbitohs -3.254271538 St. George Illawarra Dragons 1.068966332 Sydney Roosters -1.018238516 Wests Tigers -0.110045575

Between 41 and 28 home games were considered for each team, which is a statistically significant sample size (the smaller figure due to several teams sharing home grounds, and hence more games of theirs being at a neutral venue, when they met each other).

Problem 1: for most teams, the noted effects are quite small. The paper I referenced last post took into account factors such as distance travelled, how long the home team had been at home, how long the away team had been on the road, how long each had off prior to "this" game, etc - i.e. in far more detail than I'm capable of. They also took into account each team's quality and discovered that the home/away/travel factors accounted for about 1% of outcome variability. That is, 99% of results were better attributed to performance factors. These results weren't unique to these paper writers, they were mostly using a different approach to check earlier conclusions. (I also remember reading a blog post where a guy discovered that every 1000km that an NFL away-team travels adds 0.1 to the home team's score - only he couldn't say with any statistical confidence that it wasn't actually 0! )

Problem 2: NRL has a discrete point-scoring distribution (i.e. scores largely go up in 2's, 4's or 6's) so any of the bonuses noted above that are less than abs(of around 1) could be simply due to quirks in the scoring system.

Problem 3: Not really a factor in NRL/AFL, but also noted in the paper is an issue in this age of free agency: some players don't even live where there home team is. Their family and home base is in an entirely different city and they are essentially always on the road; just playing at one particular stadium more than others. Familiarity with a venue may still have an effect though (teams moving stadiums have been noted to have a lesser HGA in the USA; in some European soccer leagues teams in capital cities are noted to have a lesser HGA than provincial ones - presumably due to more players having played at the grounds in capital cities)

Problem 4: strength of schedule is not necessarily consistent between seasons. Teams don't play each other twice in a season and they don't reverse the home and away schedule from one year to the next. Low crowd-drawing opponents in the NRL (often of lessor quality) are sometimes instead played at neutral grounds like Gosford, Darwin, Perth, etc. Playing higher quality/drawing teams at home would negatively impact the HGA figures.

Problem 5: people with far more time, skills and incentive than I have found little evidence of a significant effect. From the closing of "New Insights Involving the Home Team Advantage" by Swartz and Arce:
Quote:
With respect to differential home team advantages, there appears to be little evidence of the widely held belief that some teams in a league have extraordinary home team advantages. An exception to this rule may involve teams whose home games are played at high elevations where a systematic benefit is conferred. Although this observation appears to be generally appreciated for the "Mile high city" Denver, the finding is less well-known for major cities such as Salt Lake City and Calgary.


Problem 6: the above paper considered 30 years worth of data. I don't have 30 years of data - some teams haven't even existed that long. What is a reasonable amount of data to examine for HGA, to override the normal deviations in scoring that are otherwise 99% accounted for by performance factors?

Problem 7: what can matter doesn't often get measured. Soccer players were measured to have higher testosterone levels prior to playing in their home stadium than away. They were higher still if the home team perceived their opposition as a fierce rival rather than a moderate one (http://news.discovery.com/human/why...tter-131004.htm). Does the opposition become a fierce/r rival by being from the next suburb over (immutable) vs beating them in the grand final last year (changeable) ? Does a new player to the team have the same testosterone production as a one-club veteran? How do I know if they received a stirring speech or tongue-lashing from the coach prior to running out?

Problem 8: home ground advantage is all down to officiating bias anyway; isn't it? Each 10,000 fans equal 0.1 to the home team's score in soccer. Depending on how many games the referee has officiated. And whether there's a running track around the field or not. And whether the field is enclosed so that what crowd there is is louder.

Problem 9: people believe the darndest things. Players in a black uniform play more aggressively and are penalised more often than those who do not! ("The Dark Side of Self- And Social Perception: Black Uniforms and Aggression in Professional Sports", Frank and Gilovich). Imagine if players actually believed they should have a home field advantage! Until they didn't. Have a run of losses against an opponent and suddenly there's a curse!

Problem 10: rookie players have been noted to play poorer at home, perhaps because of anxiety about meeting the crowd's expectations. For star players, they improve at home.

Problem 11: confirmation bias.

Quote:
Originally Posted by evajb001
Next example, in the NBA you have the Denver Nuggets who play/train 1,600 metre's above sea level. I've heard of simple betting systems that you bet on the nuggets whenever a team is playing the second game of a back to back in Denver simply because teams struggle with the thinner air. So once again its a distinct home advantage based on geography/venue.
Yep, studies seem to bear this one out. Seemingly as it's due to physiological reasons and is noted across sports and countries (soccer teams in La Paz seem to go alright at home too).

Early season NRL games in Queensland get trotted out as similarly advantageous (due to the heat and humidity), but there's little support for the effect of game temperature vs away team's "home" temperature.

Quote:
Originally Posted by evajb001
Same goes for Fremantle/West Coast in the AFL, they travel every second week every year and become accustomed to flying all around australia for games. Victorian teams actually leave victoria maybe 5-6 times a year and even then it might only be to adelaide or sydney once or twice etc. So you'll find Freo/eagles have a distinct home advantage.
Travel is more interesting. Actual distance travelled seems to have little to zero effect. Some NHL hockey teams were actually found to whittle away home advantages the longer they'd been on the road! And surely teams on the road would do better the more rest they have, right? In baseball, the more rest the road tripping team has, the worse they do...

Time zones would be applicable to Freo and West Coast. From "Measuring Circadian Advantage in Major League Baseball: A 10-Year Retrospective Study" by Winter, et al
Quote:
The team with the circadian advantage won 2,620 games (52.0%), a winning percentage that exceeded chance but was a smaller effect than home field advantage (53.7%). When teams held a 1-h circadian advantage, winning percentage was 51.7%. Winning percentage with a 2-h advantage was 51.8% but increased to 60.6% with a 3-h advantage.

Note that team quality was not taken into account with this study. It appears pretty negligible to me up until that 3 hour difference, which would only incorporate WA/NZ teams playing Qld teams during daylight savings time (in AFL/NRL). This dude, however (http://phdfootball.blogspot.com.au/...nce-doesnt.html) says that teams travelling over more than 2 time zones from west to east, are disadvantaged - otherwise there's little measurable difference. (That is, West Coast and Freo aren't especially good at home, they're just worse when they play on the eastern seaboard).

In summary, as borne out by my (comparitively tiny) experiment, a team-specific home advantage appears to have very little bearing on game outcome - unless played at significant altitude (N/A in NRL/AFL) and perhaps for teams travelling greater than 2 time zones eastward (and possibly teams returning from a western game with a short turnaround, by inference).
Reply With Quote
  #33  
Old 16th April 2015, 01:18 PM
evajb001 evajb001 is offline
Member
 
Join Date: Oct 2009
Posts: 1,298
Default

very very interesting stuff walkermac, thanks for taking the time to post it all. I don't have time at the moment to comment further but just wanted you to know that at least one person is reading this and taking it in with interest. There are so many variables to HGA that as you say its likely to be negligible over time. But then if you dig deeper into those variables like distance traveled, quality of accommodation when traveling, or taking it to even further extremes, you can start to not have enough data, particularly with your favored sport NRL.

great post and great study, as i've said before really enjoying the discussion.
Reply With Quote
  #34  
Old 16th April 2015, 07:11 PM
woof43 woof43 is offline
Member
 
Join Date: Jan 1970
Posts: 862
Default

Yes, I second that post as well. It is very interesting.
My only question is, have you given up using MCsims.
Reply With Quote
  #35  
Old 16th April 2015, 08:21 PM
walkermac walkermac is offline
Member
 
Join Date: Nov 2013
Posts: 995
Default

Massey was my guiding influence with this method. Here's what he wrote regarding his use of a Universal (i.e. League-wide) Home Ground Advantage in his original paper: "Universal refers to the fact that each team experiences the same advantage when playing at home. It has been shown in one study that such a null hypothesis cannot be rejected at standard statistical levels of significance. Furthermore any indication of team to team differences in the homefield advantage is relatively unimportant (Harville 1994)"

Harville's paper can be downloaded here for "only" $14: http://www.jstor.org/discover/10.23...=21106023524141. No thanks....

Using a practical example, here is my work for...

NRL Round 7

League-wide Home Ground Advantage

Tips:
Canterbury-Bankstown Bulldogs
St. George Illawarra Dragons
Penrith Panthers
North Queensland Cowboys
Melbourne Storm
Wests Tigers
Newcastle Knights
South Sydney Rabbitohs

Value Bets (bets are $25.60):
St. George Illawarra Dragons @ $2.22
Penrith Panthers @ $2.00
New Zealand Warriors @ $3.30
Canberra Raiders @ $2.95

(if Parramatta Eels and the Cronulla Sharks drift from $2.95 and $2.30, by a matter of 2 or 3c, respectively; they would qualify as a value bet)


Applying a team-based HGA
Same tips as above, save for Canberra Raiders in lieu of Wests Tigers - the rating system has this game as a coinflip; only a tiny change was necessary to bump it from one team to the other. Same value bets as above (if Eels and Sharks drift by 10c, they qualify as a bet).


I see matterofstats writes in favour of team/ground-specific HGAs here: http://www.matterofstats.com/mafl-s...antage-usi.html. To be honest, I have less confidence in his results. He's one guy, not subject to oversight or peer review (unlike the thesis and journal article writers); he himself terms the rating system used as the "Very Simple Ratings System" - perhaps it's simply not very good; he also wrote "the process of determining the optimal parameters as recorded in this blog has taken a couple of weeks ... and there is no guarantee at all that they're globally optimal" saying that he used "hand-optimisation" to do it. ...but the main reason is: 'cause I've already developed a belief due to the other material and his work is contrary It's human nature to be in a denial
Reply With Quote
  #36  
Old 16th April 2015, 08:46 PM
walkermac walkermac is offline
Member
 
Join Date: Nov 2013
Posts: 995
Default

Quote:
Originally Posted by woof43
Yes, I second that post as well. It is very interesting.
My only question is, have you given up using MCsims.


Hi woof43,

Still using Monte Carlo Simulations and will probably post an example for both of Friday night's games, tomorrow. The sims use this method's score estimates as a starting point, so that's why I've been concentrating on it.

I've made one change so far regarding my initial talk of Monte Carlo. To determine the likely score distributions, I was pulling data from any source available. As a result, I think I was high-balling it a little with points, as sources included lower quality leagues. With more blow-outs and higher scores, they likely had less bearing on a tough, evenly-matched competition such as the NRL. So, less data now; but likely more representative.

To come - and speaking of team-specific home ground advantages - I want to use team-specific standard deviations for scoring. To a regular viewer it seems that some teams have more capacity than others to score points. On investigation this does appear to be borne out by the data. Wests Tigers, for example, have a 9.7 standard deviation in the error, whereas the Warriors are at 12.7.

....but that data is from 3-4 years of games. How relevant are those early figures to today's team? I'm not sure whether I should just pay attention to their last however-many, or use a weighted standard deviation, or .... ?
Reply With Quote
  #37  
Old 17th April 2015, 01:10 PM
walkermac walkermac is offline
Member
 
Join Date: Nov 2013
Posts: 995
Default

Here's a Monte Carlo Simulation run for the Bulldogs-Sea Eagles game tonight. 10,000 repetitions which - as large as that sounds - probably isn't high enough to accurately model the unlikelier of events. Anything priced greater than $10 is likely worth taking with a grain of salt. Indeed, there are quite a number of outcomes that this simulation deemed impossible (the '#DIV/0!' errors); clearly this is not the case.

Initial team scores were seeded using the adapted-Massey rating system that was described prior in this thread. Full-time scores (i.e. *not* 80 minute scores) from the past 15 years or so of the NRL were used to model likely points distributions. Consult your bookmaker's terms and conditions to see whether this has repercussions for your bet.

(Semi-related note: there were 3 ties post-golden point over the last few seasons so, my model picks the winner - when there is a winner - at an ever-so-slightly higher rate than last reported )

Last post I queried how I would handle standard deviation (a Monte Carlo simulation randomly selects from a given distribution according to frequency; i.e. for a bell curve, values from the middle of the curve get selected more often than those at the tails; the standard deviation indicates how wide a range to select from for a given "confidence"). I went with a weighted standard deviation, using the same time-based weighting system that I calculated team ratings from. I haven't tested how successful this approach is yet.

With teams listed as per Tuesday and using a 95% confidence interval (though I don't have 95% confidence in the figures!), here are the estimates. I highlighted where Centrebet currently has prices >15% than what I calculated. Some of the lines aren't offered by Centrebet, I just use a standard spread of values.

Code:
Canterbury-Bankstown Bulldogs 1.66 Manly-Warringah Sea Eagles 2.59 Canterbury-Bankstown Bulldogs (-2) 2.00 Manly-Warringah Sea Eagles (+2) 2.33 Canterbury-Bankstown Bulldogs (1-12) 2.45 Canterbury-Bankstown Bulldogs (13+) 5.12 Manly-Warringah Sea Eagles (1-12) 2.98 Manly-Warringah Sea Eagles (13+) 20.28 Draw 90.91 Canterbury-Bankstown Bulldogs (+12.5) 1.05 Manly-Warringah Sea Eagles (+12.5) 1.24 Game Total Over 37.5 1.80 Game Total Under 37.5 2.25 Game Total: 0-10 #DIV/0! Game Total: 11-20 37.17 Game Total: 21-30 5.47 Game Total: 31-40 2.59 Game Total: 41-50 3.18 Game Total: 51-60 11.59 Game Total: 61-70 270.27 Game Total: 70+ #DIV/0! Both Score >= 6 1.05 Both Don't Score >= 6 19.61 Both Score >= 8 1.12 Both Don't Score >= 8 9.48 Both Score >= 10 1.16 Both Don't Score >= 10 7.36 Both Score >= 12 1.25 Both Don't Score >= 12 5.07 Both Score >= 18 2.76 Both Don't Score >= 18 1.57 Both Score >= 20 4.38 Both Don't Score >= 20 1.30 Both Score >= 24 11.71 Both Don't Score >= 24 1.09 Canterbury-Bankstown Bulldogs by > 6.5 5.82 Margin < 6.5 2.16 Manly-Warringah Sea Eagles by > 6.5 2.74 Canterbury-Bankstown Bulldogs (by 19+) #DIV/0! Canterbury-Bankstown Bulldogs (by 13-18) 20.28 Canterbury-Bankstown Bulldogs (by 7-12) 8.16 Canterbury-Bankstown Bulldogs (by 1-6) 4.68 Draw 90.91 Manly-Warringah Sea Eagles (by 1-6) 4.20 Manly-Warringah Sea Eagles (by 7-12) 5.88 Manly-Warringah Sea Eagles (by 13-18) 8.69 Manly-Warringah Sea Eagles (by 19+) 12.45 Either team by 13+ 4.09 Either team by < 12.5 1.32 Canterbury-Bankstown Bulldogs (-26.5) 105.26 Canterbury-Bankstown Bulldogs (-25.5) 55.25 Canterbury-Bankstown Bulldogs (-24.5) 52.08 Canterbury-Bankstown Bulldogs (-23.5) 34.01 Canterbury-Bankstown Bulldogs (-22.5) 32.26 Canterbury-Bankstown Bulldogs (-21.5) 21.69 Canterbury-Bankstown Bulldogs (-20.5) 21.23 Canterbury-Bankstown Bulldogs (-19.5) 13.64 Canterbury-Bankstown Bulldogs (-18.5) 12.45 Canterbury-Bankstown Bulldogs (-17.5) 9.77 Canterbury-Bankstown Bulldogs (-16.5) 9.44 Canterbury-Bankstown Bulldogs (-15.5) 7.37 Canterbury-Bankstown Bulldogs (-14.5) 7.28 Canterbury-Bankstown Bulldogs (-13.5) 5.49 Canterbury-Bankstown Bulldogs (-12.5) 5.12 Canterbury-Bankstown Bulldogs (-11.5) 4.37 Canterbury-Bankstown Bulldogs (-10.5) 4.25 Canterbury-Bankstown Bulldogs (-9.5) 3.55 Canterbury-Bankstown Bulldogs (-8.5) 3.47 Canterbury-Bankstown Bulldogs (-7.5) 2.89 Canterbury-Bankstown Bulldogs (-6.5) 2.74 Canterbury-Bankstown Bulldogs (-5.5) 2.38 Canterbury-Bankstown Bulldogs (-4.5) 2.35 Canterbury-Bankstown Bulldogs (-3.5) 2.03 Canterbury-Bankstown Bulldogs (-2.5) 2.00 Canterbury-Bankstown Bulldogs (-1.5) 1.75 Canterbury-Bankstown Bulldogs (+1.5) 1.55 Canterbury-Bankstown Bulldogs (+2.5) 1.39 Canterbury-Bankstown Bulldogs (+3.5) 1.38 Canterbury-Bankstown Bulldogs (+4.5) 1.28 Canterbury-Bankstown Bulldogs (+5.5) 1.27 Canterbury-Bankstown Bulldogs (+6.5) 1.21 Canterbury-Bankstown Bulldogs (+7.5) 1.19 Canterbury-Bankstown Bulldogs (+8.5) 1.13 Canterbury-Bankstown Bulldogs (+9.5) 1.12 Canterbury-Bankstown Bulldogs (+10.5) 1.08 Canterbury-Bankstown Bulldogs (+11.5) 1.08 Canterbury-Bankstown Bulldogs (+12.5) 1.05 Canterbury-Bankstown Bulldogs (+13.5) 1.04 Canterbury-Bankstown Bulldogs (+14.5) 1.02 Canterbury-Bankstown Bulldogs (+15.5) 1.02 Canterbury-Bankstown Bulldogs (+16.5) 1.01 Canterbury-Bankstown Bulldogs (+17.5) 1.01 Canterbury-Bankstown Bulldogs (+18.5) 1.00 Canterbury-Bankstown Bulldogs (+19.5) 1.00 Canterbury-Bankstown Bulldogs (+20.5) 1.00 Canterbury-Bankstown Bulldogs (+21.5) 1.00 Canterbury-Bankstown Bulldogs (+22.5) 1.00 Canterbury-Bankstown Bulldogs (+23.5) 1.00 Manly-Warringah Sea Eagles (-23.5) #DIV/0! Manly-Warringah Sea Eagles (-22.5) #DIV/0! Manly-Warringah Sea Eagles (-21.5) #DIV/0! Manly-Warringah Sea Eagles (-20.5) #DIV/0! Manly-Warringah Sea Eagles (-19.5) #DIV/0! Manly-Warringah Sea Eagles (-18.5) #DIV/0! Manly-Warringah Sea Eagles (-17.5) 185.19 Manly-Warringah Sea Eagles (-16.5) 147.06 Manly-Warringah Sea Eagles (-15.5) 56.18 Manly-Warringah Sea Eagles (-14.5) 53.48 Manly-Warringah Sea Eagles (-13.5) 24.27 Manly-Warringah Sea Eagles (-12.5) 20.28 Manly-Warringah Sea Eagles (-11.5) 14.08 Manly-Warringah Sea Eagles (-10.5) 13.46 Manly-Warringah Sea Eagles (-9.5) 9.01 Manly-Warringah Sea Eagles (-8.5) 8.67 Manly-Warringah Sea Eagles (-7.5) 6.35 Manly-Warringah Sea Eagles (-6.5) 5.82 Manly-Warringah Sea Eagles (-5.5) 4.72 Manly-Warringah Sea Eagles (-4.5) 4.63 Manly-Warringah Sea Eagles (-3.5) 3.65 Manly-Warringah Sea Eagles (-2.5) 3.57 Manly-Warringah Sea Eagles (-1.5) 2.83 Manly-Warringah Sea Eagles (+1.5) 2.33 Manly-Warringah Sea Eagles (+2.5) 2.00 Manly-Warringah Sea Eagles (+3.5) 1.97 Manly-Warringah Sea Eagles (+4.5) 1.74 Manly-Warringah Sea Eagles (+5.5) 1.73 Manly-Warringah Sea Eagles (+6.5) 1.58 Manly-Warringah Sea Eagles (+7.5) 1.53 Manly-Warringah Sea Eagles (+8.5) 1.40 Manly-Warringah Sea Eagles (+9.5) 1.39 Manly-Warringah Sea Eagles (+10.5) 1.31 Manly-Warringah Sea Eagles (+11.5) 1.30 Manly-Warringah Sea Eagles (+12.5) 1.24 Manly-Warringah Sea Eagles (+13.5) 1.22 Manly-Warringah Sea Eagles (+14.5) 1.16 Manly-Warringah Sea Eagles (+15.5) 1.16 Manly-Warringah Sea Eagles (+16.5) 1.12 Manly-Warringah Sea Eagles (+17.5) 1.11 Manly-Warringah Sea Eagles (+18.5) 1.09 Manly-Warringah Sea Eagles (+19.5) 1.08 Manly-Warringah Sea Eagles (+20.5) 1.05 Manly-Warringah Sea Eagles (+21.5) 1.05 Manly-Warringah Sea Eagles (+22.5) 1.03 Manly-Warringah Sea Eagles (+23.5) 1.03 Manly-Warringah Sea Eagles (+24.5) 1.02 Manly-Warringah Sea Eagles (+25.5) 1.02 Manly-Warringah Sea Eagles (+26.5) 1.01 Game Total Over 41.5 2.60 Game Total Under 41.5 1.62 Canterbury-Bankstown Bulldogs Total Over 18.5 1.73 Canterbury-Bankstown Bulldogs Total Under 18.5 2.37 Manly-Warringah Sea Eagles Total Over 17.5 1.87 Manly-Warringah Sea Eagles Total Under 17.5 2.15


Manly by 1-6 is the best value bet, by my measure ($4.20 vs $6.50).

This is an interesting match as these two teams have pretty much the lowest standard deviation (in the estimation error) in the comp. Tonight's other match sees two far less predictable sides meeting each other.
Reply With Quote
  #38  
Old 17th April 2015, 02:28 PM
walkermac walkermac is offline
Member
 
Join Date: Nov 2013
Posts: 995
Default

And here's the estimates for the other match. Far fewer discrepancies here, likely due to their variability in scores.

The best value bet is Game Total 61-70 $24.63 vs $34 BUT - as noted earlier - given the unlikelihood of the event and that "only" 10000 matches were simulated, it's probably not a very accurate estimate in this instance.

Code:
St. George Illawarra Dragons 2.09 Brisbane Broncos 1.93 St. George Illawarra Dragons (+2.5) 1.81 Brisbane Broncos (-2.5) 2.24 St. George Illawarra Dragons (1-12) 3.50 St. George Illawarra Dragons (13+) 5.20 Brisbane Broncos (1-12) 3.51 Brisbane Broncos (13+) 4.31 Draw 181.82 St. George Illawarra Dragons (+12.5) 1.30 Brisbane Broncos (+12.5) 1.24 Game Total Over 36.5 1.63 Game Total Under 36.5 2.60 Game Total: 0-10 256.41 Game Total: 11-20 25.19 Game Total: 21-30 5.75 Game Total: 31-40 3.33 Game Total: 41-50 3.35 Game Total: 51-60 7.17 Game Total: 61-70 24.63 Game Total: 70+ 178.57 Both Score >= 6 1.09 Both Don't Score >= 6 11.53 Both Score >= 8 1.22 Both Don't Score >= 8 5.57 Both Score >= 10 1.28 Both Don't Score >= 10 4.55 Both Score >= 12 1.47 Both Don't Score >= 12 3.13 Both Score >= 18 2.97 Both Don't Score >= 18 1.51 Both Score >= 20 4.47 Both Don't Score >= 20 1.29 Both Score >= 24 10.58 Both Don't Score >= 24 1.10 St. George Illawarra Dragons by > 6.5 2.78 Margin < 6.5 3.10 Brisbane Broncos by > 6.5 3.15 St. George Illawarra Dragons (by 19+) 7.46 St. George Illawarra Dragons (by 13-18) 10.22 St. George Illawarra Dragons (by 7-12) 7.85 St. George Illawarra Dragons (by 1-6) 6.35 Draw 181.82 Brisbane Broncos (by 1-6) 6.25 Brisbane Broncos (by 7-12) 7.97 Brisbane Broncos (by 13-18) 10.29 Brisbane Broncos (by 19+) 10.53 Either team by 13+ 2.36 Either team by < 12.5 1.74 St. George Illawarra Dragons (-22.5) 19.65 St. George Illawarra Dragons (-21.5) 14.62 St. George Illawarra Dragons (-20.5) 14.47 St. George Illawarra Dragons (-19.5) 11.00 St. George Illawarra Dragons (-18.5) 10.53 St. George Illawarra Dragons (-17.5) 8.49 St. George Illawarra Dragons (-16.5) 8.28 St. George Illawarra Dragons (-15.5) 6.85 St. George Illawarra Dragons (-14.5) 6.79 St. George Illawarra Dragons (-13.5) 5.49 St. George Illawarra Dragons (-12.5) 5.20 St. George Illawarra Dragons (-11.5) 4.53 St. George Illawarra Dragons (-10.5) 4.44 St. George Illawarra Dragons (-9.5) 3.91 St. George Illawarra Dragons (-8.5) 3.87 St. George Illawarra Dragons (-7.5) 3.30 St. George Illawarra Dragons (-6.5) 3.15 St. George Illawarra Dragons (-5.5) 2.85 St. George Illawarra Dragons (-4.5) 2.82 St. George Illawarra Dragons (-3.5) 2.50 St. George Illawarra Dragons (-2.5) 2.47 St. George Illawarra Dragons (-1.5) 2.21 St. George Illawarra Dragons (+1.5) 1.98 St. George Illawarra Dragons (+2.5) 1.81 St. George Illawarra Dragons (+3.5) 1.79 St. George Illawarra Dragons (+4.5) 1.66 St. George Illawarra Dragons (+5.5) 1.65 St. George Illawarra Dragons (+6.5) 1.56 St. George Illawarra Dragons (+7.5) 1.53 St. George Illawarra Dragons (+8.5) 1.42 St. George Illawarra Dragons (+9.5) 1.41 St. George Illawarra Dragons (+10.5) 1.36 St. George Illawarra Dragons (+11.5) 1.35 St. George Illawarra Dragons (+12.5) 1.30 St. George Illawarra Dragons (+13.5) 1.29 St. George Illawarra Dragons (+14.5) 1.23 St. George Illawarra Dragons (+15.5) 1.23 St. George Illawarra Dragons (+16.5) 1.19 St. George Illawarra Dragons (+17.5) 1.19 St. George Illawarra Dragons (+18.5) 1.15 St. George Illawarra Dragons (+19.5) 1.15 St. George Illawarra Dragons (+20.5) 1.11 St. George Illawarra Dragons (+21.5) 1.11 St. George Illawarra Dragons (+22.5) 1.09 St. George Illawarra Dragons (+23.5) 1.09 St. George Illawarra Dragons (+24.5) 1.07 St. George Illawarra Dragons (+25.5) 1.07 St. George Illawarra Dragons (+26.5) 1.05 St. George Illawarra Dragons (+27.5) 1.05 Brisbane Broncos (-27.5) 20.41 Brisbane Broncos (-26.5) 19.96 Brisbane Broncos (-25.5) 15.38 Brisbane Broncos (-24.5) 14.95 Brisbane Broncos (-23.5) 12.24 Brisbane Broncos (-22.5) 12.09 Brisbane Broncos (-21.5) 9.90 Brisbane Broncos (-20.5) 9.79 Brisbane Broncos (-19.5) 7.75 Brisbane Broncos (-18.5) 7.46 Brisbane Broncos (-17.5) 6.39 Brisbane Broncos (-16.5) 6.28 Brisbane Broncos (-15.5) 5.35 Brisbane Broncos (-14.5) 5.31 Brisbane Broncos (-13.5) 4.49 Brisbane Broncos (-12.5) 4.31 Brisbane Broncos (-11.5) 3.90 Brisbane Broncos (-10.5) 3.81 Brisbane Broncos (-9.5) 3.42 Brisbane Broncos (-8.5) 3.37 Brisbane Broncos (-7.5) 2.90 Brisbane Broncos (-6.5) 2.78 Brisbane Broncos (-5.5) 2.53 Brisbane Broncos (-4.5) 2.51 Brisbane Broncos (-3.5) 2.26 Brisbane Broncos (-2.5) 2.24 Brisbane Broncos (-1.5) 2.02 Brisbane Broncos (+1.5) 1.83 Brisbane Broncos (+2.5) 1.68 Brisbane Broncos (+3.5) 1.67 Brisbane Broncos (+4.5) 1.55 Brisbane Broncos (+5.5) 1.54 Brisbane Broncos (+6.5) 1.47 Brisbane Broncos (+7.5) 1.44 Brisbane Broncos (+8.5) 1.35 Brisbane Broncos (+9.5) 1.34 Brisbane Broncos (+10.5) 1.29 Brisbane Broncos (+11.5) 1.28 Brisbane Broncos (+12.5) 1.24 Brisbane Broncos (+13.5) 1.22 Brisbane Broncos (+14.5) 1.17 Brisbane Broncos (+15.5) 1.17 Brisbane Broncos (+16.5) 1.14 Brisbane Broncos (+17.5) 1.13 Brisbane Broncos (+18.5) 1.10 Brisbane Broncos (+19.5) 1.10 Brisbane Broncos (+20.5) 1.07 Brisbane Broncos (+21.5) 1.07 Brisbane Broncos (+22.5) 1.05 Game Total Over 41.5 2.13 Game Total Under 41.5 1.89 St. George Illawarra Dragons Total Over 17 1.69 St. George Illawarra Dragons Total Under 17 2.50 Brisbane Broncos Total Over 19.5 1.90 Brisbane Broncos Total Under 19.5 2.11
Reply With Quote
  #39  
Old 18th April 2015, 01:35 PM
woof43 woof43 is offline
Member
 
Join Date: Jan 1970
Posts: 862
Default

Hi Walkermac,
Can you eyeball your MCsims printout to find say the Max score achieved by either team?
Reason I ask if your using a Normal distribution and you have scores either side.
As an example If I ran a runner be it a horse or greyhound thru my generator with out putting high or low bounds and the Mcsim is using a normal distribution you may end up with race times that are quicker than the track record at the tails, we know a runner can run slow and slower, but to run race records no, so you don't really want that sort of calculation in your mix as it is unrealistic, the same may apply with unrealistic scores.

In regards to Standard Deviation, I'm not sure if it's applicable with sports, but in the three racing codes, it is beneficial to use barrier or Box specific stdev, so I would be looking at when I can use a narrower stdev (of course it has to have enough data to be valid).
I'd also be looking at doing a moving average calculation on the Stdev to find how many games you require for it to become stable, then run that calculated moving average on your past history and try and look for fluctuating patterns and why.
Reply With Quote
  #40  
Old 18th April 2015, 02:17 PM
walkermac walkermac is offline
Member
 
Join Date: Nov 2013
Posts: 995
Default

Quote:
Originally Posted by walkermac
Manly by 1-6 is the best value bet, by my measure ($4.20 vs $6.50).

and For the Dragons-Broncos match:
Quote:
Originally Posted by walkermac
The best value bet is Game Total 61-70 $24.63 vs $34 BUT - as noted earlier - given the unlikelihood of the event and that "only" 10000 matches were simulated, it's probably not a very accurate estimate in this instance.


My measure of best value above is pretty naive: "odds on offer" divided by "my odds" and biggest number wins.

Using Kelly Strategy and the formula for expected growth - and being conservative and presuming my estimate is 10% greater than the true odds - the best value bet for:

Bulldogs vs Manly
Either wins team by < 12.5 points WIN (just)
$1.32 by my estimate, $1.82 offered by Centrebet
Expected Growth of Bankroll: 3.6%

Dragons vs Broncos
Game Total over 36.5 LOSS
$1.63 by my estimate, $1.95 offered by Centrebet
Expected Growth of Bankroll: 0.3%

Titans vs Panthers
Penrith Panthers +2.5
$1.56 by my estimate, $1.91 offered by Centrebet
Expected Growth of Bankroll: 0.6%

Cowboys vs Warriors
Warriors +4.5
$1.67 by my estimate, $2.41 offered by Centrebet
Expected Growth of Bankroll: 3.2%

Storm vs Roosters
Melbourne Storm +4.5
$1.61 by my estimate, $1.99 offered by Centrebet
Expected Growth of Bankroll: 0.64%

Some of those bets seems worthwhile, others not so much.
Reply With Quote
Reply


Thread Tools Search this Thread
Search this Thread:

Advanced Search
Display Modes

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

vB code is On
Smilies are On
[IMG] code is On
HTML code is On
Forum Jump



All times are GMT +10. The time now is 08:54 PM.


Powered by: vBulletin Version 3.0.3
Copyright ©2000 - 2018, Jelsoft Enterprises Ltd.
2008 OZmium Pty. Ltd. All rights reserved . ACN 091184655