/biz/ - Business & Finance

>>401728
>make money from alcoholics??

>>401744
the inefficiency in sports betting markets that makes betting profitably possible is directly due to people betting

>on their favourite team
>to make up for yesterday's losses
>to make watching sports interesting
or even
>while drunk

it's true

I appreciate questions 3 and 4, since we're definitely not lying and all winners here. No families destroyed or pesky debt collectors right?

Right?

>>401778
>addictive personality disorder and/or worrywort detected

>>401728
depending on what you want to talk about
>>>/sp/49964363

>>402258
sorry, wrong thread... >>>/sp/49949278

>>401728
-NASCAR

-Average finish at style of track that they are racing on that weekend/how they did during practice

-after initial costs of 1 dollar per race and 1 dollar for chase races (final ten races of the year) for a total of $138 over 3 years, I made about $500

-no, it's their choice to drink or make up for their previous loss, then why can't I make a choice. I could still be wrong and they could still be right.

I bet $1K Saints fuck everybody's daughter on Spring Break after a SB victory. Who dat

>>402432

You have a 1/43 chance of winning. Any team sport gives you 50%.

>>402484
Does anyone else do UK horse racing?

>>402484
>posting on /biz/
>not understanding expected value

>>402432
>style of track
>nascar

>>402667
>sure anon
>retarded degenerate alcoholic britfags do all the time
>like you!

>>>/killyourselves/

felt like being being 3edgy9me for a second there

in seriousness though
>>402667
do you use anything more than tips? consensus? lane #?

>>402432
what is the combined rake % like in nascar? I know in things like soccer goal total markets, where you've got options like
>0 goals
>1 goal
>2 goals
>3 goals
>4+ goals
you can see the house taking upwards of 20-30% in margins
with so many possible winners, surely this is even more noticeable in nascar, or am i wrong?
the market that interests me as far as nascar (and F1) is concerned is the head-to-heads. ie. "Earnhardt vs. Danica Patrick"
ever give that a go? I presume the vigorish isn't nearly as bad there

I was reluctant to post this until after other people shared because I don't want to be the only focus of the thread

>last year's NHL season algorithm performance
>bankroll in blue
>line of estimated performance from 2012-2013 in red
>starting bankroll $100
>bet size: average of $10 and 10% of current bankroll
>pic related

>>403468
correction: net gain/loss in blue, not bankroll

>>403470
>correction: net gain/loss in blue, not bankroll
It's still effectively a bankroll graph then, just offset by a bit. Either way, nice work.

I'm anticipating betting on college football for the first time this year. I really regret not betting last year as I would have made a killing. When last year's season began, though, I was still learning to code, and was hand-munging data in my spare time, and couldn't really keep up effectively. So I just kept track of picks - 72% against the spread during the regular season, 86% during the postseason, including all the bcs upsets.

I've spent the offseason learning how to automate everything so I won't have to do much of anything by hand, and backtesting my model against previous years. Last year definitely would not have represented a typical year, but results from backtesting still imply that it should be consistently profitable. Can't wait.

>>403522
right you are, just offset a bit

that's sick! very good numbers. even with a 10% dip you'll still be beating any high-risk stock portfolio by miles.

do you allow for any manual inputs? ie. injuries and trades, or sanity checks before placing bets? or are your picks 100% automated from the stats. that's one of the dilemmas I toy with now and then

I'd be excited to pick your brain offline. we could both benefit from a second opinion

>>403568
I allow for manual inputs only to the extent of negating certain picks, but never generating them. So for example, I may choose to negate a pick if the team originally predicted to beat the spread has an injury to a key player. However I don't use injury as a reason to elevate a non-pick into a pick. So basically I generate all picks statistically and cull a few manually. I try to be disciplined about culling based off predefined explicit rules (such as "never bet against Gus Malzahn") so as to avoid subjective skew. Individual instances still pop up though.

>>404246
well put. if I were to implement along those lines I gotta say you've done it perfectly. adjustments by omission rather than inclusion, and certainly never switching sides altogether

so you only bet ATS? what is the output of your models? a predicted Home vs. Away differential and if it is off from the official line by a certain number of points you bet?

>>404350
I only bet ATS so far.
The models to which I've referred don't output scores. I realized somewhere along the way that you don't really have to predict scores if you can just find an individual factor or two that you evaluate better than the public. In this case I've written my own algorithms for evaluating offenses and defenses according to strength of schedule. I also have an algorithm for evaluating on an individual basis the degree to which being at home or on the road affects a team's performance. I then look at the difference between the teams' apparent performance and their adjusted performances. Teams with the greatest gap are the teams I expect to most likely have been under or over-rated by oddsmakers or other bettors. So I actually make those picks regardless of what the line actually is.
In short, betting ATS you only have to determine which direction the line is more likely to err in.
I still work on models that predict scores, of course, because I'm that kind of person. They're coming along, but so far none of them match the performance of what I've described.

>>401728
These bets are even more pathetic than before. People stopped betting on Championship games years ago because there are statistical winners and losers. I haven't heard of a close Championship game, in past years, it's usually been one team just being plain better than the other. The difference might be small, but it shows and it is predicted.
Every bet, there were often far more winners than losers. So you make pocket change.
I started doing these random draws where people all pitch in money and randomly select names out of a jar, hat, box etc. You don't pick your team, your luck does.

>>404730
Hence the spread, totals, moneyline, prop bets, and so on.
>>404730
>I started doing these random draws where people all pitch in money and randomly select names out of a jar, hat, box etc. You don't pick your team, your luck does.
Sounds pointless then, but to each his own.

Learn matched betting, anything else is just luck.

Sites offer free money for sign up, bet using free money then use betfair or similar to bet against the bet you placed. You lose a little each time but because it is free money you win so $20 free ends up being $15 with no risk (you win either way)

>>404836
Sounds interesting, but does it scale? I'm not interested in making a risk-free $20 each week doing this. If I wanted to do that I'd become an arbitrager.

Surely there's a ceiling to these free bets? I used to dabble on bet365 (before I realized the value of statistics) and the deposit bonuses are a one-time shot with a required "rollover" of up to 5x in order to withdraw. more one-time bait than a legitimate long-term strategy AFAIK

>anything else is just luck
incorrect.

>>404728
that's cool. the way I approach moneyline is to assign (or deduct) each team "points" for their performance over 5-10 key stats. the difference between the home/away team's points is then fed into an exponential equation that returns a probability of a home win. for example, if my model assigns the home team "2.41" and away team "2.35", the home team is favoured by "0.06" points

the S-curve evaluates this advantage to home win probability of 59%, away win 41%
>pic related

"fair odds" are returned as 1.70 and 2.43

if the available odds' payouts exceed the fair odds by a required difference (ie 3%) then a bet is placed

>>405025
If you don't mind my asking, how did you come to that equation? Did you test the correlation of your performance indicators and then find a best-fit equation, or did you independently derive it, or something else entirely?

>>405211
sure, anon

it's all done through random optimization using Palisade's Evolver add-in for Excel, where the decision variables are

a) the slope of the Exp equation
b) the weights of each dependent variable

and the objective is to maximize a weighted formula of
a) Net Profit and
b) ROI

this of course is done on a training subset of game data and the algorithm is applied to a testing subset. a chi-square test is performed to check probability of the observed outcome occurring through pure chance given expected win-rate (50%) and expected ROI per bet (-3%)

>>405218
>>405211
ps the training/testing test is done mainly to guard against overfitting

>>405218
Thanks, I wasn't familiar with Evolver, I'm checking it out now. I might try using it to attempt to generate scores. I've actually just been writing my own functions into Excel with VBA as necessary.

>inb4
>>excel
>>bayes
>>costanzaface.jpg

>>405223
of course

Thanks for the response. I was just curious because I actually derived my offensive and defensive metrics from probability laws first, rather than coming up with an essentially arbitrary value and then finding a best fit. I'm not sure which method is more common.

>>405268
I was just curious because I actually derived my offensive and defensive metrics from probability laws first, rather than coming up with an essentially arbitrary value and then finding a best fit. I'm not sure which method is more common.

My viewpoint is there are two ways to do it
Estimated P(home win) as:
= 0.5 +/- net "points" assigned for home team

the problem with that is there really should be a degree of decelerating returns as far as estimating the probability of a home win goes. as an example:

If the LA Kings played a university level team, they would have a roughly 98% chance of winning

Against a high school team, they would still probably have a 99% chance of winning

However, between a high school team and university team is a huge ability gap (probably much greater than the relative gap between any two NHL teams) yet the effect on probability is minimal

that's why I use the formula before estimating probability, not the other way round

>>405281
>If the LA Kings played a university level team, they would have a roughly 98% chance of winning
>Against a high school team, they would still probably have a 99% chance of winning
>However, between a high school team and university team is a huge ability gap (probably much greater than the relative gap between any two NHL teams) yet the effect on probability is minimal
If I understand your last sentence correctly, that's exactly what I derived - how to effectively calculate the win probability in situations similar to what you gave. Admittedly it's only mathematically rigorous in a theoretical and effectively impossible set of circumstances - most importantly that a team's ability doesn't change from game to game, which obviously isn't the case. But even though it's not rigorous in actual application, it sure seems to be effective.
In the case you gave my model would have the university win against the HS team 66.9% of the time, but I also think you overestimated the chances of a high school team beating the Kings. I understand that you were just using example numbers though.
For the sake of argument, let's say the HS team will beat the Kings in 1000 times rather than 1 in 100. In that case, my model would project the university team to win against the HS team 95.3% of the time, which sounds more like what are looking for.

>>405370
damnit
>1 in 1000
>what you are looking for
Getting sloppy.

>>405370
>>405371
true, those were just bs numbers for general effect. could you go into a little more detail about your probability derivation? are you using ELO as your basis?

>>405404
The theory is just a relatively simple variation on Bayesian conditional probability that makes the k-factor and curve aspect of elo completely unnecessary.
It's the application that's messy as it generates a probability matrix of the likely value of team's strength, rather than an absolute value. This is because, unlike the example we discussed in which we assumed specific prior probabilities of team A beating team B and C, and derived the probability of team B beating team C, the model assumes all teams are unknown from the beginning, which means that it never generates a single definite value for what I've been calling a teams "strength," but a curve of probable values. The chance of winning is ultimately calculated by combining the probabilities of every possible matchup.

>>405417
So just to clarify, a given teams strength is represented as a curve, and a matchup is calculated by building a matrix from the two quantized curves.

>>404730
could be the worst post i've ever read on 4chan. about 5 sentences but literally nothing is said
1/10 for making me reply. don't know why i didn't see that earlier...

>>405421
could I trouble you for a visual representation of what you mean?

in a way-way-way-old model of mine, I would plot
>Team A's average game totals on X
>Team B's average game totals on Y
>each intersection would be the average of the two percentiles
And my expected Game total would be the average of the entire matrix while the probability of a certain number of goals (ie. 5 or more would be calculated as the % of the matrix where that many goals were scored, in this case 44%)

is what you're doing along those lines?

>>405572
I never got around to evaluating the accuracy of that model though... decided optimization was the way to go over averaging things on gut feel

>>405572
Apologies in advance for how godawful this looks. It's old and was never intended for display - everything done by these tables is now done by custom functions that run faster and don't occupy miles of spreadsheet real estate. But it's got the gears and wheels out on display, so to speak.
The bold numbers are the quantized possible strength ratings for each team, adjacent to these are the probability of a team being at that rating. The top table calculates the probability of every possible pair of strength ratings. The next table calculates the chance of winning in the event of each possible case of the top table. The sum total of the table is the chance of winning. The next table calculates the chance of losing which is redundant but makes for a good double check as the total of the two tables should always add up to one. Next to the blue highlight is what the team's probabilities of each rating will be adjusted to in the event of a win (top table) or loss (bottom table).
In this case it's a particularly strong team facing a slightly below average team. I actually fucked up a bit labeling them home and away - these are not actually the right home and away teams, the team labeled home is the one who's chances are being calculated. On the opponent's sheet is a symmetrical table.

I'm not sure how much this will help, as again, it wasn't really built for looking at.

On another note I do have a defunct model similar to the one you posted. I wonder how many separate models people who get into this make and discard, on average.

>>405624
that's a great explanation of an interesting approach anon, and mother of fancy jesus that's gotta be one large file with that many tabs!
>ohlawd.jpg

question. for generating the strength ratings, do you use a trailing average of a certain number of key stats? or is it based purely on prior wins/losses and the strength of the teams those games were played against?

I'm curious because my approach is very much from the "key stats" school of thought. mostly because in the little time I've thought about it, it seems like the process of rating team strength would inevitably require dealing with a circular reference, which has sort of discouraged me (though I'm not sure it'd all that tough with some thought and research)

>On another note I do have a defunct model similar to the one you posted. I wonder how many separate models people who get into this make and discard, on average.
To be honest, I'm not sure there are many people out there doing quant analysis of sports. outside sites like 538 and sport-specific bloggers, I imagine that only a handful of every 1000 bettors do as much analysis as you or I...

>using all this math

That shit only works in baseball. If you want to be a good sports better you have to know the game, the athletes and coaches. It's a lot of studying and you have to be able to look at every game and come up with solid theories about how things will happen.

Source: I made over 2k last year playing fantasy football

>>405672
I think your statement is more true of football than most sports. the possession-based rhythm of the game and the variety of ways in which a team can win games make it much more difficult for numbers alone to encapsulate a team's ability.

having said that, the only justifiable reason to not be using stats to your advantage is if you aren't putting money at risk or you are playing against people with vastly inferior expertise as you. I could be totally wrong, but I wonder if a not insignificant part of your 2k haul may have been due to the other people in your pool having no clue what they are doing?

have you ever "paper traded" real betting sites? you really ought to if you're as good as you say you are (not being an asshole, serious)

>>405671
>that's a great explanation of an interesting approach anon, and mother of fancy jesus that's gotta be one large file with that many tabs!
When my desktop broke down my laptop took four or five hours to recalculate and crashed about half the time. It was terrible, but also what spurred me to learn VBA and make it many times more efficient.

>is it based purely on prior wins/losses and the strength of the teams those games were played against?
This.
>it seems like the process of rating team strength would inevitably require dealing with a circular reference
You are absolutely correct, it's a recursive formula that converges on a rating over multiple iterations, hence why it originally took so long to run. Usually after about fifteen iterations most teams' ratings are stable out to about 4 decimal points.

The trouble I have had is in taking scoring into account. Because it's a straight win/loss zero-sum probability calculation, it has no real way to evaluate degree of victory, each game is a one or a zero. What I settled on, and what generated the results I mentioned at the beginning of the discussion, is retooling it to compare offenses to offenses and defenses to defenses rather than evaluate the actual games played, and correlate the resulting ranking with expected score. It works, but I'd still like to try and find a way to generate scores directly with the same method.

Thanks for your interest by the way. I don't get to discuss this much.

>>405672
On the other hand, the more difficult the analysis is, the more opportunity there is to outplay other bettors and oddsmakers. I went after NCAA football partly for the challenge, but also because it's the sport I know best. I caught errors and had ideas that I might not have had if I'd focused on baseball, which I don't keep up with or know much about.

>>405698
>four or five hours to recalculate and crashed about half the time. It was terrible
lol, Excel truly is a hulking behemoth for this kinda stuff. I've made the transition to a language called R but it's not as intuitive as Excel, or even VBA which was the first language I learned back in grade 9

>retooling it to compare offenses to offenses and defenses to defenses
That's crazy because I've actually found comparing Team A's offense to Team B's defense and Team A's defense to Team B's offense was more valuable. In simple speak:
If Team A routinely scores 1 more goal than their opponents defense typically allows, they are likely to do so against their next opponent. Obviously that's a very simplistic interpretation, but that's just how I saw it when I began juggling my stats

>Thanks for your interest by the way. I don't get to discuss this much.
Likewise man most people hear "sports model" and either laugh (because GAMBLING! anon!), or ask to find out how it works only to have their eyes glaze over 30 seconds into the explanation

send me an invite on the ol' facebook. I've got a few ideas about monetizing picks I want to talk to you about
>fb/liamgrue

>>405812
I'm kind of one of those reclusive types that's terrible with social media, and has no FB. Shocker seeing those on 4chan, I know.

I just followed you on my sparsely used Twitter, the best way to get ahold of me is to send messages to my gmail or live account by the same name (without the space between the words). Messages to the live account will go to my phone, which I'm fine with.

>most people hear "sports model" and either laugh (because GAMBLING! anon!), or ask to find out how it works only to have their eyes glaze over 30 seconds into the explanation
Story of my life.

btw I should clarify that I compare O to O and D to D to rank them, but when making predictions compare the difference in rankings between Team A's O and Team B's D and vice versa. Hope that makes sense, I'm getting pretty tired.

Anyone bet Formula 1?

I don't like to gamble much. I don't mind going to a Casino and playing blackjack or putting some bucks down on a footy game (im Australian) or go to one of the big horse racing events, get dressed up, drunk and place some bets (but always with money i can afford to lose)

My sister was a gambling addict (not that i can say much, i was a heroin addict at the time as well). But i just don't really care much for gambling. It's fun in small doses.

>>407237
sorry to hear that, anon
what was it that kept your sister going back? the rush? trying to win back losses? how much did she lose?

>>405679
What's the best to go about paper trading for sports do you think? Are there sites that keep a balance for you? A spreadsheet on your own?

>>408956
I'd go for Excel spreadsheet personally, it's most straightforward and interfaces well with graphing your gains/losses down the road

page 10 bump

>>401728
anyone have any good recommendations for professional sports handicappers online? there seem to be a lot of scammers who inflate their success rate but i would rather pay a legit guy with a 55%+ track record than trying to half-ass my way through learning to handicap on my own.

>>401728

I bet on UFC fights.

I do very little statistical analysis. I was a wrestler for years, even post college, and I've been an avid MMA watcher since UFC 9 or so.

I am quite profitable. In the ~5 years of betting I've only hit my quit mark once (each year I set a loss limit of $3K) - however, keeping track of the best I would have made the rest of the year I would have ended ~$4K up). I average around $5K/year gains. Even given thought to doing it professionally, but one long bad run and it's curtains for early retirement.

I don't know what the numbers are, but I see professional gamblers placing >$100K every Sunday and Monday, and there are millions bet by professionals every UFC event. That's a lot of alcoholics and addicts placing $50 bets to make up for.

So no - because I'd wager (heh) that the vast majority of the betting pool comes from professionals, not addicts or drunks.

>>405812
>>405829

>tfw read this entire conversation and didnt understand any of it

do you actually win a lot? i think i do a smiliar thing but i dont have a code or anything i just do it manually before each game i wanna bet on

Used to do some dota 2 betting. You'd look at several things. For instance, the heroes the team uses, and the average comp gpm those heroes have against the top picked heroes for the opposing team. Average comp gpm includes ganks and towers in addition to creepscore, so it's the core factor in who wins.

>>412041
I should also mention it's not really solid income since most teams fuck around in low reward matches to try out new builds.

I bet $10 Mexico would lose.
I bet $10 Germany would win.

Girlfriend still hasn't paid me. Should I break up with her?

>>411978
UFC could be great to bet on if I had the expertise, but I've never been a fan of wrestling, boxing, or anything of the sort. good to hear you've been consistently in the black.

>>412021
haven't bet a dollar so far, only back-tested or "paper traded" as mentioned earlier. looks highly promising though, average of 5-7% profit per bet it looks like

>>412041
>>412045
sounds dicey, but you should always bet on whatever you're familiar with, be it sports, politics, vidya, etc. keep at 'er

>>412046
kek

>>411962
check out pickmonitor, seems to be the most legit of the handicapper-for-hire sites. you can see stats and profitability numbers. check em before you buy, you know? and it's a 3rd party website not just some fuck with a blog and "a proven track record of success!!"

>>412137

how do you go about creating your own formula like you have?

what sport do you mainly focus on?

>>412307
(Posting from phone so new ID)
exclusively NHL
>step 1 is to download and organize statistics in a trailing average way
>ie. if I'm trying to predict the winner of game 40, I would look at each team's average/median performance over a number of stats during games 30-39 for example. this number would vary depending on sport given how "streaky" the sport may be
>step 2 is to find a correlation between these trailing avg statistics (30-39 in our example) and the probability of the outcomes of the game you're predicting (40)
>if you haven't guessed, that's the hard part
You could use Excel, Matlab, anything else. I use a combination of R and Excel personally
>step 3 is to make your model actionable - this means automatically downloading game odds and comparing those to your derived probabilities to identify promising picks

Hope that helps let me know if you want more details

>>412382

did you post in /bet/ on /sp/ earlier about something to do with NHL? swear someone was talking about a similar thing.

>>412390
about a week or two ago? I tried but it seems /bet/ is literally filled with alcoholics and thrill seekers who do nothing besides pounce on 3rd tier Aussie rules football and are upset when Bet$3.65 limits their accounts to $10 wagers or refuses payouts

As you can see /biz/ has MANY more intellectual posters than that

>>412394

LEL

some people on /bet/ make a decent amount but a lot are like you said too haha

/biz/ is 2deep4me tbh i love sports and betting on it and do okay but id have no idea how to write an algorithm for it

>>412422
yeah, I'm sure it's a mix that's true. add me up on fb/liamgrue if you wanna talk more or compare picks if you do NHL at all

>>412428

Im a brit I mostly bet on Football (soccer) and tennis.

Out of curiosity, have any of you guys used Econometric modeling for sports betting?

>>412462
If you mean using economic or financial theory to make picks, then no. If you mean in terms of process, then yeah.
Most predictive modeling for sports could be described as a stochastic econometric model, except that selection and weight of variables is based on knowledge or theory of the sport rather than economic theory.

>>412462
What do you mean by econometric exactly? How would you anticipate it applying to a sport? Interested in alt theories so please do explain

>>403468
Great, now forward test it. Backtesting is too easy.

Where do you get all the data to put in from?

Like the previous sports games and stats.

>>413461
It is possible to deceive yourself with backtesting, if you do it wrong, true. OTOH it takes a lot less time to work through multiple years when back testing than when forward testing, which takes, well, multiple years.

>>413527
Mainly by looking for other autists via Google-fu. (name of sport) +csv or xml are a good place to start, sometimes throwing in the name of a stat you want to include. "Historic" is a good search term to include as well.
http://www.repole.com/sun4cast/data.html is where I get my basic handegg csv's, for example.
You can also scrape data from major sports websites, obviously that's a bit more work.

How do I get into sports math?

Blogs? Websites? Books?

>>413563
Take a class or read books on probability and statistics. Look around for blogs of people who are into it - there are plenty out there who explain their models pretty thoroughly. See if there are aspects to their model or process that you don't understand, and research them. Also, seeing if you can find flaws in their logic can be a good exercise.

>>413574
And by find, I mean find and actually quantify.

>>413574
> there are plenty out there who explain their models pretty thoroughly

Like? Not him, but how do I know if one math autist is actually right and the other is a retard?

>>413574
>>413576

got any links to books/blogs that would be good for a beginner?

i am a huge sports fan and like to bet on it sometimes but never looked into it like this.

>>413646
>For general betting related stuff
http://forumserver.twoplustwo.com/40/sports-betting/propplayers-awesome-sports-betting-faq-1204065/
>intro regarding modeling
http://www.bettingexpert.com/blog/how-to-build-a-sports-betting-predictive-model-getting-started
>example of a decent blog
http://andrewgelman.com/category/sports/

Seriously, google is your friend as far as the blogs go.

>>413461
Excellent point. If you look at my original post
>>403468
The thing to note is that the graph is actually a combination of a training data set and testing data set. On the x-axis is the game # before present (1230 games in a season)
Games before 930 were used for training the model and the last 300 of the season were for testing purposes. That's why the red (anticipated average net profit) and blue lines intersect at x=300

I was worried about this strategy because it introduces a degree of randomness because it roughly coincides with the trade deadline, but it seems that my concerns were unfounded. Of course the success may not carry over to next season but it's a good sign at least

>>413527
I scrape mine directly from the NHL site and a fan stat site called behindthenet, but it's a fair bit if work if you're just getting into it

>>413563
key here is to choose (at least to start) the sport you're most familiar with. if that's cricket, study cricket. if that's curling, study curling.

spend a few hours listing the contributing factors that lead to success in that sport. try to rank them. see if those factors overlap or are correlated (or inversely). when you feel like you've done everything you possibly can, then head to google and see how close you've gotten to others' theories.

that's how I started at least

>>413859
apologies, wrote that from my phone

>correction
the training dataset is everything to the left of 300 on the x-axis (not 930), and the testing dataset is everything to the right

>>414711
Hey statsbro if you do decide to bet this season and intend to do it online, consider 5dimes... they have a giant exploitable hole IMHO (giggidy) in that they allow same game parlays on NHL games if they are total+moneyline (not total + spread though, unfortunately).

As for me, I decided to just give in and go to Vegas this year. Zeus help me.

I started with making £10 bets and a bankroll of £50 in January and I've made about £10k betting on football (soccer) this year.

You DO NOT need an in depth knowledge of statistics or mathematics, you need to understand "value", have some discipline, and to be resourceful. Bookies offer so many games in-play that they're never going to be able to price them all correctly. It's incredible how badly they can fuck up. There is a fortune to be made out there.

>2013+1
>not betting with bitcoins
>not using betvip.com

>>415201
looks legit

>>415178

Jesus Christ.

Post some proof thats fucking phenomenal.

You must have got lucky though, betting 20% of your bank roll is not good strategy regardless of how much you made.

Really interesting thread.

My strategy is completely different. I use an Elo rating system based on the opening odds of one of the biggest bookie.

EA = 1 / ( 1+10^ (Rb-Ra)/400 )
EA winning probability
RA Rating Home
RB Rating Away

I won't tell you the whole calculation behind it because I'm working right now on a paid service. Almost all bookies limited me.

These guys right here use a similar approach: http://awaypayday.blogabet.com/
But they are using a diffierent staking plan and some of their picks makes no sense to me. But thats another story.

The benefits of this approach:
- you do not need to analyse your games, your program calculates the picks
- you do not need to specialise, you can bet on every league worldwide
- on the long-run I have a average payout of 117.28%
- you can use this calculation in every team sport or every sport where two players compete


Oh by the way I'm talking about soccer/football.
The world cup was a loss for me, it is really hard to compare nations.

>>415213
I was curious so I ran some numbers on this.
Assuming even bets minus a 10% vig, 20% of your bankroll is the optimal size single bet if you can maintain a 61.9% win rate on those bets. With a starting roll of 50x, you'd be expected to cross 10,000x after 149 bets on average.

>>415251

I don't think he had a strategy. May he played some combinations and got lucky.

>>415251

It's not sound long term strategy though. Everyone has bad days and you could lose it all or most of it.

5-10% is a much safer limit.

>>415255
I can't really speak for him, I was just curious what sort of edge was needed to make that betting size a good strategy
>>415257
>It's not sound long term strategy though
It's the optimal strategy over any term, assuming that winrate and payout ratio, and no betting minimum of course. You can't lose your whole bankroll betting 1/5 of your bankroll.
Assuming an 5x minimum size bet he'd have had a 50/50 shot of "busting" rather than raking it in. Of course, if he can afford to put in another 50, he'd have had an additional shot at the whole 10,000 in this hypothetical scenario.
If the minimum bet was 2x, he would have a 1/5 chance of busting over the 149 bets.
And so on.

>>415257
>>415263
Also, for giggles, assuming that same win rate, if he bet 10% of his bankroll, his expected total after 149 bets would be around 760. If he limited himself to 5%, his expected total would be 150.
Not trying to be a dick, btw, just giving examples of why it's good to be able to actually quantify this stuff. For fairness I'll work out the cost of overestimating winrate next.

>>415269
So let's say you think you're a 62% winner and are betting as per circumstances above, but you were overestimating yourself and your true win rate is more like 55%. Now your average bankroll after 149 bets is 224 - but that doesn't really tell the whole story, since 93.75% of the time, you go bust before you make it there.

>>415213
>>415251

>>415213
>>415251

20% is dumb yeah, but I found at the start that i didn't have the patience to spend all day looking for £2-5 bets, so you could say I got lucky.

I can't provide conclusive proof because I've used loads of accounts, and I currently only have access to an ipad (lol), but this is my first bet365.

>>415290

This is my sportingbet

This is another 365, lasted about a week before being limited to £5 lays

>>415290
>>415298


thats awesome, you got any tips?

do you post in /bet/ on /sp/?

>>415302

I post in /bet/ a bit yeah, there are a few who make this kind of, or greater, profit.

Like I said I only have an ipad at the moment, makes researching/betting pretty tricky, nothing of my own at the moment. /bet/ seems to be on flora tallinn, but I couldn't comment on it. There are load of friendlies at the moment, which can be a goldmine if you have the patience to research. If you look into and find team news for most, you're bound to find a mispricing.

>>415310

no i dont mean match tips haha i mean overall tips to do what you did.

>>415290
>>415301
How do you re-open another account after being limited?

Do you then re-use the sign up bonus?

>>415338

Mainly look into obscure games and find things that bookies haven't taken into account she pricing the games. One of my favourites was a hong kong reserve league game where a team had had to release most of their players because they'd filed for bankruptcy, i think it was on a -1 handicap and was something like 8-0 at half time. Last week an Estonian team could only field nine players (found this out on twitter), and were still priced as favourites. Yesterday the Netherlands women's under 19 team had a senior side player playing up front. Stuff like that, you just need to dig.

>>415386
Open them in friends name. You don't need to go through their bank, all you need is their proof of id and address. You can use the sign up bonus again yeah.

>>415404
> Britfag here
> Over 40 bookmaker accounts.
> Arbitrage across bookies/exchanges, started in uni firstly arbing bonuses then running 2-8% arbs found using software.

Fair play, have done this using friends names playing them a small commission after getting limited/banned at most major bookies.

Not as great at straight betting as you, prefer to take a less risk approach very interested by your style.

How do you price your games? What sources do you use to get this info? Is it just lots of research and looking through odd games that bookies do little research into?

>>415176
>same game parlays on NHL games if they are total+moneyline @ 5Dimes
wow that's unbelievably stupid on their part
a great scenario for this would be betting the Under+Dog(moneyline). all you need is for the underdog's goalie to hold the favourite to 1 or 2 goals and bam there's a payout in the neighbourhood of 5:1

>As for me, I decided to just give in and go to Vegas this year. Zeus help me.
congrats on the big move Justice. how do the Vegas lines compare to 5Dimes, Pinnacle etc?

>>415220
interesting strategy. will have to look into that further. thanks for the link anon

>>415257
People love to espouse the whole "bet 2-5% of your bankroll" thing but what they seem to overlook is the ability to simply deposit more. Say I've set aside $500 from my personal budget for betting. that's my bankroll -- NOT however much of that $500 I decided to deposit. why would you tie up so much cash in an online account where (unless you've hit a string of bad luck) you're only going to be dipping into at most $100 or $200 on a given day?

my personal strategy is to use an average of:
>a dollar figure - ie. $20
>a percentage of my bankroll - ie. 8%

as my balance is growing, the bet size is driven up by the $ (because there is less risk) and when it has grown a bit, it's held in check by the % (because a losing run would be much more costly once my bankroll has grown substantially)

I'm sure your strategy works just fine, that's just my 2c

plebs get maine tipster for reduced price rasic95luka@gmail.com

tipster for like 3 € haha not even joking

>>415575
>>415581
shillfag pls go.

>>415290
what was your total amount deposited during this run? curious to hear an ROI type figure

>>415290
and also, could you guess at approximately how much time per week you would put into researching and placing these bets?

Is anyone familiar with the tax regulations surrounding gambling winnings in Canada and the US? I believe that in Canada winnings are non-taxable (in theory) but how would one go about consistently depositing sizeable sums from foreign bookies? Wouldn't it raise eyebrows from the tax man?

I was wondering when /sp/ would get into sports betting, albeit I don't visit it as much as I should.

I've made close to $2,500 the past two seasons in college football. IMO, these games have the loosest lines of any sport, especially the first 6-7 weeks of the season.

>>416091

How come? What are your tips for college football?

You're welcome to come to /bet/ on /sp/ and post tips and pick up others.

Where can an American get an online betting account with a semi-respectable outfit?

>>416415
SBR has a good outline of which books are good for different locations if I recall

>>416907
happened to be talking with a friend about this today so I'll just answer

As a Canadian, my best option would be to line shop at 5Dimes, Pinnacle (known to be among the top 2-3 books for payout %) and maybe bet365 if there's a good bonus/promotion going on.

sadly, of those, only 5Dimes serves the US, so I'd go with that and maybe William Hill or Ladbrokes

>>417440
dun goof'd ladbrokes and WH both hate Liberty apparently

sportsbookreview .com/sportsbooks/

Bovada is used by a lot of americans I believe.

>>415562
>congrats on the big move Justice
Thanks, kind of nerveracking for me though. I was looking to move somewhere anyway though, so I figured why not? Being in The Land Of The Free (tm), there's hardly anywhere that sports betting is legal. Most of my friends think I've lost it.

>how do the Vegas lines compare to 5Dimes, Pinnacle etc?
Spread lines appear to be pretty similar. I guess they can't afford to deviate too crazily from online bookmakers anymore now that the process is so fluid. Payouts on certain types of bets, like parlays, don't seem to be as high on average.

>all you need is for the underdog's goalie to hold the favourite to 1 or 2 goals and bam there's a payout in the neighbourhood of 5:1
Seems like a good plan, I'm not going to get into hockey any time soon but if I did I'd look for goalies with an unusually high volatility in their performance, particularly in middle-of-the-pack teams.

Found a pretty good set of articles on handicapping and modeling. The guy's a pro tout, so the site as a whole is a little heavy on trying to convince you to subscribe to his picks, but his articles are pretty solid, and he seems to be well regarded.
https://www.drbobsports.com/essays.cfm?cat=0

I'm thinking about putting together a pastebin of resources, if anyone's interested.

>>417624
I think everyone would appreciate it, might even end up in the /bet/ pastebin.

>>415997
I am very much not a lawyer and have no experience with Canadian tax law but I thought you might find this interesting:

"In the 2006 case of Leblanc v. The Queen, the Leblanc brothers won an average of $650,000 per year between 1996 and 1999 by gambling on sports lotteries in Ontario and Quebec. The brothers wagered $10 to $13 million per year, developed a computer program to analyse bets, and negotiated a 2-3% discount from ticket retailers. The Canada Revenue Agency treated this gambling operation as a business, “because it was managed and organized with the object of realizing a profit (para 7).” However, the Tax Court, acknowledging “that certain types of gambling can in some circumstances be a business but it is a rare circumstance (para 36),” characterized these gambling activities as a personal endeavour (para 48):

The appellants [the Leblanc brothers] are not professional gamblers who assess their risks, minimize them and rely on inside information and knowledge and skill. They are not like the racehorse-owner, who has access to the trainers, the horses, the track conditions and other such insider information on which to base his wagers. Nor are they seasoned card players or pool players who prey on unsuspecting, inexperienced opponents. Rather, they are more accurately described as compulsive gamblers, who are continually trying their luck at a game of chance."

tl,dr:
Analytical handicapping does not involve minimizing risk (kek), not based on inside information, and not a contest of skill (double kek) and therefore can't be reasonably expected to make a profit and isn't taxable.

From http://www.lawnow.org/are-gambling-winnings-a-prize-under-the-income-tax-act/

>>417643
fuckin hell lad, is that part of a comedy routine? screw Vegas, you should move up north and become a syrup salesman like me.

as a side point, I wonder what made them stop. quit while the going's good?

>>417615
>I'd look for goalies with an unusually high volatility in their performance
that's an analysis I've been meaning to get to. wouldn't be hard by any means either. the square bettor, if they look at statistics at all, almost exclusively looks at "averages" which really never tell the whole story. good data on volatility would be a nice weapon to have in the bag. interestingly though, when I built the NHL model, I fed trailing save% data for the starting goaltender of each team and my model trained itself to completely ignore both home and away goalkeepers. could very well be due to just how volatile the majority of keepers are...

bump

Do you guys use the Kelly criterion?

>>419092
I'd only use Kelly if I knew to within 1-2% what % of bets I'd win. anyone who claims that they can predict their own accuracy with such precision is either deluding themselves or ought to be a professional handicapper.

>flat % betting for me

>ellid goal

w-what

>>419150
I'm testing it right know with my bet history. I'll show a comparison tomorrow.

>>419411
would be interesting to see the side by side. eager to be proven wrong

>>419411
any luck with that?

>>405672
>studying
>know the game
>solid theories

In other words two sentences of nonsense to convince yourself a bet is good.

>>401753

This exactly.

It's exactly why Manchester United to lose when they play at home is a long term winning bet.

Rabid fans that will sink all their money into the team even when they're playing a weaker team.

If you're Euro or have access to a European/Canadian bank account then Eurovision betting is a fucking gold mine. If you don't have absolute shit taste in music you'll be able to pick out at least a couple that are surefire top 10 picks, and those usually pay 35-50% of the bet.

Been following it for 7 years now and the early sportsbook favorite has always made top 10. Usually only pays about 12%, but I've yet to see it go wrong and I've yet to see them pick an absolutely terrible song as the favorite.

Other important things:

1. The winner from last year doesn't win the next year, in fact they rarely even make top 10. Sweden's entry last year was very popular in the juries but got slammed by the public who didn't want Sweden to win two years in a row.
2. The following countries will not make top 10 either: UK, Spain, Italy, France, Germany. These people get an automatic entry into the finals since they fund much of the event. People find their favorites during the semi-finals so they miss out on a valuable opportunity to win people over before the actual contest. Same thing goes for the winner the previous year, the automatic entry into the finals screws them.
3. The song that wins doesn't have to be the best but it has to have a wide appeal. The song should fit well into a genre and be a rather mainstream genre, or have classical roots.
4. Gimmicks almost never do well, but it did this year so that's up in the air.
5. Sweden does well ever year except 2013. They have their own song contest that's big by its own right, and is well liked by the other countries geopolitically. They're practically a guarantee to win.

>>419950

>It's exactly why Manchester United to lose when they play at home is a long term winning bet.


lol maybe last season but you realise they hardly EVER like once a season or less lost at home before >Moyes?

>>419952

Even then they are still overfavored, the handicap is never less than 1 goal.

When they pay +1000 on a loss it doesn't matter if if the rarely lose at home. This season's been even more profitable than usual.

>>419950
>long-term winning bet
>citation needed

last 10 seasons, betting $1 a game nets you -$58.05 for a great return of -30.53%

>>419960
>>419950
seriously though statements like "this is always a good bet" are just nonsense because:

1. They're usually said in hindsight and the books will adjust accordingly.
2. Plucked out of someone's ass.

>>419962

How are you finding the prices of old games?

>>419959

I still dont believe this is a good bet. I am just going off the fact that United basically never lose at home so if you have stats to prove what you are saying then that's fine.

Last season was an anomaly and this season United will be far better again.

>>419968
football-data.co.uk

only has the prices from uk books but you get the picture

>>419949
chuckled

>>419950
>Rabid fans that will sink all their money into the team even when they're playing a weaker team.
It's just important to remember to use a book that allows bets from such fans. I once made the mistake of thinking that the New York Rangers would be heavily shaded on bet365 due to their massive fanbase. pic related


>Eurovision betting is a fucking gold mine
I don't doubt it anon. My only gripe would be how qualitative the betting decisions would have to be. for me, if you don't have the quantitative analysis going for you, that's how emotion gets involved and that's a dangerous path

>>419960
R*KT

>>419968
bet explorer is good too

>>419970
Even at the beginning/first quarter of last season, public sentiment was pretty positive on Moyes so you would've had to know something everyone else didn't to realize that losses were coming

>>402707
How do you get involved in F1 betting? I got into the sport last year and some bets would make the races more fun to watch.

>>420942
Wouldn't know myself. Don't follow the drivers or teams but it would be quite interesting. My only issue with it would the relatively small sample size of data.. what are there, 20 races a year?

>>419909
I'm busy right now. Sorry.

It's not that easy as I thought because I would like to involve my payouts and my changing bankroll. I'm working on it.

This shit is really interesting.

the general key to winning long term is finding mispriced odds and fixes.

>>412394
i bet you're the kind of guy who bangs solid 7's and feels good about himself.

do you ever bet on Barcelona?

Hello, I'm here for the fixes.

you want make real money bro?

interest bro

Tears tbh

>>421994
Fixes? What do you know?!

guys pls gib winning algorithm tips

The trick is to bet on the team that will win

Can I borrow 4k pls

Adelaide to win tomorrow is amazing value, they are in their own country, playing a team that came 6th.
would put my house on it

>>422004
What algorithm model did u use? Post proof or not betting

>>422005
I know the manager and he wants to beat a spanish team and malaga aren't trying because half of their star players are still playing in the world cup

I used maths and several calculators to come to this conclusion.

Adelaide will win tomorrow if they play soccer(also known as football) in Australia

If you bet $1000 you will win your initial $1000 + profit (see attachment for proof)

Gonna be making mad money on the fixes tomorrow when these clowns paper bet. Absolute tears tbh

>>422009
you should start a facebook group or a website like mrsoccer(footy)tips.com

>>422009
>>422007

Well that's all the proof I need. The graph must have taken ages to figure out, solid work

>>422009
>>422015
This point right here is where the real profit starts. Class half time lump.

post yfw these /biz/ nerds don't know about the fixes

>>422018

>>422018

post ending in 69 decides what I do with this folder

>>422025
open it?

>>422025
Gibe tips

>>422018
I know right... like the rigged sports betting has been well documented countless times

If it isn't organized crime it's just some other greedy fucks!

>>422035
>Not wanting free money
Are you even Jewish?

>>422037
I mean if I had a connection to the rigged results I would do that, but I don't so I won't feed the house my money

>>422038
W-what

>>422039
he scared

>>422042
N-no! I j-just have better ways to invest my m-money!

that sure was hilarious!

>>407237
josh reynolds detected

jesus christ no wonder /biz/ is filled with autists
>muh paper trading
>muh statistics
>muh 5% bankroll growth per bet

>>422175
I like /bet/ but jesus christ there's more than one way of doing things.

Tears at people who missed the real fix last night.

What programming language should I learn to automate data collection and stuff?

>>422336
In general, R is pretty well regarded for all things data analysis.

>>417624
>I'm thinking about putting together a pastebin of resources, if anyone's interested.
pls do so if you're still watching this thread

>>422286
>>422175
>>422038
>>422033
>>422025
>>422023
>>422020
/bet/ posting itt

>>422017
9/10 did laugh

>>422806

bump for this

bump

>>424738
>>425506
>>422806

I'm still around I'm just swamped right now.
At the moment I don't have enough interesting stuff to fill a pastebin so for now I'll just post useful items I find here, and if this becomes a recurring thread I'll collect them into a pastebin.

For now here's an Excel workbook from which you can import odds history from covers.com for the NBA, NFL, NHL, and MLB.
https://docs.google.com/file/d/0B5dydT9Oe6JbU1BVMlJqVXA0VkU/edit?usp=sharing
This was not made by me.

and here's a link to some articles by Kenneth Massey (creator of one of the BCS computer polls)

>>425918
>says here's a link
>doesn't post link
http://www.masseyratings.com/theory/

>>425918
>>425919

thats awesome.

>>425918
>>425919

Thanks Anon.
I'm curious how this approach would work with soccer. In a few weeks I have some time and I'm gonna work on it.

tears

guys whats the best website for betting on elk?

Hi guys, I'm pretty certain that I can predict every match-outcome by running through an origami paper-model and doing some high-level algebra.

Will post model afterwards.

>>412137
>haven't bet a dollar so far, only back-tested or "paper traded" as mentioned earlier.

>he does it for free

tears at you, mug. lrn2shrewd

>>427405
>>427405

UAC tbh

>>427405
>>427413
what kind of paper do you use to run the model?

Does anyone here know what DNB means?

>>427405
>>427413

This model can predict FT ( Stands for full-time) scores

>>427419


>>427415
I use standard photographic paper, because the paper need to glide without too much friction during the origami-making.

who /dac/ here

So which match is scientifically fixed?

>>427419
I used to use it but currently I've run out of oil

>>427423
>>427419
>>427417
>>427414
>>427413
>>427406
>>427405
>>427416

Yawn, /bet/ is here.

>implying I'm not profiting more than you long-term and consistently.

>>427419
>>427423

It can also predict when a team will turn a match around and make a comeback.

>>427428
so what are you doing on 4chan? bragging about it to your online friends? go out and spend all that money bro

>>427431

>bragging

well you guys have come here thinking you're somehow elitist in your style of gambling so I'm just telling you that you're not

>>427428
How much has your paper-model made you then? Show us.

>>427434
>arguing with SP

Oil pls

>>427423
thanks for the heads up on that. i've only started doing rudimentary origami models using legal pad paper with 1.25' margins. i find its a good system so far but needs to be refined.

>>427436
shhh no tears, only SP now.

>>427437
You need photographic paper for dat non-friction. It's very important or else the orignami-model won't be calibrated correctly and the paper-model will be flawed.

>>427434

I don't use a paper model, I bet with my real money, and a heck of a load more than anybody in /bet/ wagers with

>>427444
Nice photoshop. Every self-respecting bettor knows that paddypower doesn't have 4.2k to pay out.

>>427444
SP what's your bankroll?

>>427449

It's a champions league qualifying game on an unlimited paddypower account, ofcourse you can get that much on

>implying you've ever tried

I have a monopoly money + calculator model

I use monopoly money (I got it from my parents board game cupboard, we have also have classics like Jenga and Hungry Hungry Hippos but they don't have any money to use) and a scientific calculator. If my sister or brother is using the calculator, I have a Casio calculator watch which can help me with the more simple equations.

First I look at a sports game I like, usually something with good odds (just ask me if you don't understand what that means) then I ask my father to be the 'bookie' but he prefers the term Banker because he works as a bank teller and he doesn't really understand what I'm doing. I use the money and pretend it's real, I deposit lump sums on particular markets with those nice odds I was telling you about earlier, then I use the calculator to work out how much I win. If I lose the bet I can simply 'rewind' the bet and get my money back. It's so fun =)

I've made over 20 hotels worth of monopoly money this week alone

>>427453
Nice one mate, you made money from a fix. How can I be as shrewd as you?

>>427453
>You in charge of reading-comprehension.

Paddy doesn't have 3.3k in their bank...

>>427444
>stealing that from me

>>427460

what the fuck? how?

>>427455
Impressive system! What's your ROI? 2%? Any paper-system that can be applied to this? Maybe a swan-origami would fit nicely with this.

>>427462
Was the max stake m8, you should know that

>>427465

no but it's placed at exactly the same time as mine

>>427468
That's when the lumps came in

>>427469

You've become too shrewd for your own good

>>427443
do you prefer wet-folding or pureland styles of origami?

>>427479
I use a mixture of modular and wet-folding. Certain pieces needs to be fit together with the wet-folding method to yield accurate predictions.

tears

>greyhound racing
>theres no logic to it
>bet on 4 out of 6 of the doghorses if they all have odds greater than 4/1
>made stupid returns doing this but unfortunately peasant tier money to start

>>427405
Get the fuck outta here!
>low odds without value
>friendly matches

I'm 100% sure you are losing money on the long run.

>>427932
DAC m8 DAC

bump 4 more funny /bet/ systems I want to make free money

bamp

don't let this shekel-trove of intellect die!

curious: is it better to optimize to maximize profit or best approximate win % with your predictions

>>431031

Return on investment is key.

btw why is /bet/ so anti friendly?

>>432611

they arent there are just a vocal minority who are. its really shitty.

i dont know where you live but /bet/ gets cool around sort of 10-11pm onwards (UK time) and a lot more chilled out.

in the day theres always people trying to inflate their ego by showing off how much betting knowledge they have.

Steve Stevenson here

always bet the underdogs so that way you are winning from the start

ayy lmao

.

last page bump

Hi, what do if you're mathematically ignorant but would like to develop a method to quantify your knowledge of a particular sport and bet more efficiently?

Like, if I consider a particular piece of information to impact a team's chances of obtaining a result, how do I combine all such data points and calculate my opinion of what the "true" price of an event would be?

>>432967
>>432967
>>432967
>>432967
Bet the under, that way you're winning for at least the majority of the game

>>436026
A lot of the shit these guys are talking about can be done in microsoft excel. It's really all about using something to keep track of your data and knowing how to properly analyze it. I've never given much thought to sports betting, but after seeing this thread I think I'm going to toy around with it come this college football season and see what happens. This season I probably won't bet any money and just see if I can come up with or find out from someone else a decent method for predicting outcomes.

>>419950
conchita wurst...

>>437470
I know I'm not expressing myself properly, but I guess what I'm asking is how do I quantify my gut feeling?

Basically, how do i take numerous disparate pieces of information into account and calculate my own odds? In a way where I can then look at lines offered by bookmakers and say "gee whiz, event X is 5/1 but my gut algorithm says they should be a lot shorter. this is worth a punt".

If anyone can help set me on the right track I would be extremely grateful

>>437806

Collect data. Are your sure your gut feeling is profitable?
Place your bets in an excel sheet and write down why you are betting on this outcome. After hundreds of bets you'll see if you are profitable and if there is some kind of correlation.

>>437863
key word: hundreds
none of this "hurrrr I was profitable last week ima move to Vegas baby"

>>437863
>Are your sure your gut feeling is profitable?
Not extremely. But I believe it would be if I had a framework in which I could play with the weights of different variables and see how that improves the predictions.

>>438527
what are the variables and sport you're referring to?

I'll dump this here

>>439909
What is this?

>>439933
On /sp/, they maintain sports betting threads and some autistic dutchman logged every single tip that anons posted in these threads and updated a spreadsheet to compare how good some nations tipped in the respective sports.
After a while he lost motivation to do it because it took him literally a few hours every day.

As you can see, horses and greyhounds made good money. Football as well if you're able to detect the shitty tipsters.

Too bad there's not spreadsheet nowadays and we don't know how that performance has changed today. Maybe some of you /biz/ people know how to write a script that logs all tips automatically?

>>440597
There are really some autistic faggots with too much freetime. That's crazy.
I'm really suprised that overall it was profitable.

A return of 105% in football is okay. Nothing special.
I'm looking forward to the new football season. Some of the smaller leagues already started.

>>439587
It doesn't matter what sport.
The variables can be any piece of data I consider to have an effect on the outcome

>>440650
Also who states returns as 105% instead of 5%

brrrrramp