Can anyone make sense of this?


I have been discussing this with Hanna but wanted to share it here too.

Not sure how to upload my graph but this is the situation:

Around 4000 bets. After 2400 bets I sat at 3300 EUR profit now after 4000 I have lost 90 % of that and sit at 330 EUR.
I use Kelly 30, value 1-20%, max stake 1% and generally bet last 3 hours before matches start but not exclusively.
First 2500 bets were mainly with Bet365 but my account “died” (got severely restricted) on September 28th where I lost almost 1000 EUR in a day. I then regrouped and have mainly used sharp bookies (Sportmarket) since.
Yesterday I lost 1000 EUR (must be a -25% ROI on the day) again which takes me down to the figure above.

I have checked some of my bets on oddsportal to see if I got value and it seems I beat the closing odds most of the time. Yet I have now lost for almost 6 weeks in a row.

Is this really just variance?

Any help would be appreciated.

Oh, forgot to say that my expected profit is almost 5000 EUR but seems impossible ever to catch that blue line again.

Also Sportmarket tells me that I actually have a ROI of 4,5 % placing 1x2 and AH bets whereas over/under bets have a NEGATIVE ROI of 6,7 %. This is after 1500 bets so don’t know if it has any statistical value.


I am in almost the same situation. My expected profit is about 2100 euro and my results are -1600 so a 3700 difference. As you say it feels impossible to catch the blue line ever again.

My settings are Kelly 30, 1% max stake and minimum value 2%. I make 50-125 bets per day.

I want to hear from the staff if this is normal or not.

Hi Marie,

When you check your reports under menu do you see a lower ROI for Over/under than for your other bet types? I think I heard from someone else here who also had much worse results from those bets than from 1x2 and AH.

I hope one of the founders will respond to this. I read an article on a competitor website (their product is practically the same) saying that all their customers who had made more than 10.000 bets were in profit whereas there were still 10-20% of customers with between 5.000-10.000 total bets who were losing money. If that is really true we still might be underestimating how volatile this is.

As stated above I really have my doubts though. My average bet is 1.87 and I hardly ever go above 2.25 so cannot quite fathom a 6 week losing run like I am having.

I can make a web tool for simulations visualization, but I am also very lazy :).

So, by now just several cases screens (n = 4000, coeff = 1.87, value = 5%, stake = 25)
I didnt choose, just 10 random consecutive simulations in a row:


ok not 10, but 8. I am really lazy.

And look if n = 100 000:

So, two summary points:

  1. do not underestimate sample size, say it as a data scientist. it always needs more than we think.
  2. ofc its true only if rebel value 5% is really true 5% (as coeff 2.1 in coin flip). but who actually knows? :slight_smile:
1 Like


Thanks Ryan,

So my graph looks a bit like your second one but most of your graphs move in the right direction. My EV is not 5 though rather around 3.4.

However, sometimes I am not very lazy:

1 Like

Thanks Ryan,

If I can pick your brain:

My stats are:

Asian handicap bets: 1926 bets, 11354 units profit, 2.2 % yield
Over/under bets: 1625 bets, 6057 units loss, -1.5 % yield

Would you stop doing the over/unders?

It is not an easy question, as a minimum, you should calculate p-value and estimate it. If your two-side p-value less than 0.01 its could say about a high probability of some problems in O/U value estimation.

I am sorry about the late reply.
My ROI for OverUnder is the worst of all, -4,4% after 733 bets.
I still think the volume is too low to claim that there is something wrong with the value estimation regarding OU. A few thousand bets and it should even out, I just hope my money lasts that long =)
But with that said; if I wouldnt have played on OU I would be +10400 instead of -8800. Also basketball is hurting me pretty bad

I also have my worst results with Over/Under bets, at -1.7% from over 1500 bets. Seems like there could be a problem with the system.

I just checked some community stats and with more than 90,000 O/U bets on football, the yield is positive but a bit lower then on other markets such as 3-way and AH. So there is still value to be made on O/U :ok_hand:t2:

1 Like

Thank you for that great info!
I have actually cut my OU bets by 50% , slowly recovering.
Is there any similar data on sports? Basketball is -8,6% for me. As with OU I am still betting on it but carefully.

I’m not sure if this is good or bad but I have almost the exact opposite results.
After +2500 bets, O/U yields all of my my entire gains!?! (total yield about 2,3%)
3way has now improved to -7,4% from a long period at -11%-ish.
Win and AH are about even Steven.
I do realize that there’s pretty much a consensus on the forum that O/U is the weakest yielding market, but at the same time we can’t really deny my stats.