[Poisec]

I got this in my emails today. I do trust the results because it is certified by the Smart Betting Club.

My concern is: the average odds are 80/1 (+80000).

Question: is the sample of 1600 bets big enough considering these unusual average odds?

[Alfa1234]

Your real concern should be being able to get on at the quoted odds.

Ask for a free trial and see if you can match their odds.

[Poisec]

Your real concern should be being able to get on at the quoted odds.

Ask for a free trial and see if you can match their odds.

They do offer a 60 day trial, I am confident I can get the odds though as most bets are placed on the betting exchange. Thanks for your advice.

[u21c3f6]

The answer to your question is that 1600 wagers at 80-1 odds is not even close to how many wagers you would need for any degree of confidence. Based on those odds just by luck you would have about 19/20 winners. 13.4% return is only about 22 winners. A difference of 2 or 3 winners is well within what one could expect by chance alone. This does not mean that they may not be on to something but these results are not unusual.

Joe.

[Poisec]

Okay that's a constructive reply, thanks, and the reason why I was asking, I had the same doubts. Betting on this kind of odds must be very frustrating... until you hit a winner or two.

[danshan11]

think of it as the lottery, the lottery is bad ROI in most cases but for the guy that won, it was a good ROI for sure. You need to really look at bought and closed lines to get any sort of idea if his tips are worth it.

if you randomly select some 100-1 shots and if you hit one before the bell rings 100, you are a winner, in general does that sound like a long term winner to you?

[HeeeHAWWWW]

Around a 1 in 6 chance it's just a coinflipper. You'd want 4300ish bets for that to be 1 in 20, and double that for 1 in 100.

This is a massive problem for anyone betting at such long odds. At evens, with such high yield, you'd only need 120ish bets for that 1 in 100 level.

[danshan11]

Around a 1 in 6 chance it's just a coinflipper. You'd want 4300ish bets for that to be 1 in 20, and double that for 1 in 100.

This is a massive problem for anyone betting at such long odds. At evens, with such high yield, you'd only need 120ish bets for that 1 in 100 level.

yeah that is what I said LOL

[trytrytry]

email scam misleading results times 10 ignore it

[semibluff]

Do NOT bet on outright tournament wins with golf. Just take the odds for the top 100 listed players and work out the total book price. You're probably looking at 150%+ and you still don't have every outcome covered.

[Barrakuda]

At evens, with such high yield, you'd only need 120ish bets for that 1 in 100 level.

So you are saying a 71-49 record (+14% ROI at -110 juice) has a 99% chance of being non-random? That would be close if the record was randomly sampled. But since it's not (the record was discovered bc it was so unusual), it's extremely likely to be pure chance.

Short version: never trust a small sample Z score unless the sample was truly randomly selected.

[HeeeHAWWWW]

So you are saying a 71-49 record (+14% ROI at -110 juice) has a 99% chance of being non-random? That would be close if the record was randomly sampled. But since it's not (the record was discovered bc it was so unusual), it's extremely likely to be pure chance.

Well, I did specify a coinflipper :-) Point taken though, not a realistic scenario - especially at one of these tipping sites, where likely thousands of others have given up, and the lucky survive.

[danshan11]

Well, I did specify a coinflipper :-) Point taken though, not a realistic scenario - especially at one of these tipping sites, where likely thousands of others have given up, and the lucky survive.

Survivorship Bias pure and simple, if you follow the plays starting now you will see reality set in very quickly

[Barrakuda]

Survivorship Bias pure and simple, if you follow the plays starting now you will see reality set in very quickly

Even if you force a group of 1,000 talentless people to track all of their plays for a couple years (eliminate survivorship bias), you'd expect to see ~50 people with records 2 stdevs above coinflipping. So beyond the fact that unsuccessful bettors stop tracking their plays, there is the issue of hunting for outperformance making luck look like skill.

That said, handicapping skill does exist. But you have a know a hell of a lot more than a W/L record to identify it.

https://bettingiscool.com/2017/11/13...less-tipsters/

[danshan11]

very valid point, that is what drives me insane, people screaming Joeblow is 97-3 in his last 100 plays tail him, in most scenarios assuming he is throwing darts his chances of hitting the next ATS is roughly 50% like everyone else. it drives me insane and that is why I keep harping over and over. look at me I am at 48% NBA over a couple hundred plays and I suck right but my sides are at 56% so am I genuis, of course neither is true

[zorba74]

Around a 1 in 6 chance it's just a coinflipper. You'd want 4300ish bets for that to be 1 in 20, and double that for 1 in 100.

This is a massive problem for anyone betting at such long odds. At evens, with such high yield, you'd only need 120ish bets for that 1 in 100 level.

How did you come up w/ that?

I'm coming up w/ about a 1 in 2 chance of achieving a 13.4% yield after 4,300 bets @ 81.00 Avg Odds.

I'm also coming up w/ about a 1 in 7 chance of achieving a 13.4% yield after 120 bets @ 2.00 Avg Odds.

Thanks.

[HeeeHAWWWW]

How did you come up w/ that?

It was one of 12xpert's helpful sheets. Away from home atm so don't have the link.

I'm coming up w/ about a 1 in 2 chance of achieving a 13.4% yield after 4,300 bets @ 81.00 Avg Odds.

Hrrrm, that can't be right if the base chance is 50% - the median should be losing the juice.

[zorba74]

Hrrrm, that can't be right if the base chance is 50% - the median should be losing the juice.

I just double-checked the figures and getting the same answer (1 in 2 chance) using a one-sample t-test for a population mean.

[zorba74]

I just double-checked the figures and getting the same answer (1 in 2 chance) using a one-sample t-test for a population mean.

Actually, about a 1 in 3 (2.8 chance). Sorry.

[HeeeHAWWWW]

It was one of 12xpert's helpful sheets. Away from home atm so don't have the link.

https://twitter.com/12Xpert/status/1083699386440474624


Can't get it working on this laptop for some reason, but it was def that one.

Major factor might be fls? At 80/1 the juice is massive.

[zorba74]

https://twitter.com/12Xpert/status/1083699386440474624


Can't get it working on this laptop for some reason, but it was def that one.

Major factor might be fls? At 80/1 the juice is massive.

Thank you and LMAO. Just clicked the link and can't believe I never saw this. I'm GameBred and I used Joe's P-value Calc to come up w/ the figures I used in this thread. I'm going to check out the new Calc now and try and see why we are apart on this issue.

Joseph is a great guy.

[Poisec]

Thanks guys by the way, replies were very constructive and helpful.

[Poisec]

https://twitter.com/12Xpert/status/1083699386440474624


Can't get it working on this laptop for some reason, but it was def that one.

Major factor might be fls? At 80/1 the juice is massive.

80/1 odds have big juice but not on a betting exchange when someone can Back @ 80 and another user can Lay @ 85. I would use only a betting exchange (Betfair) if I decided to tail such tipster.