[boxcar]

I've had an odd result using the calculator and maybe someone can help explain. If I enter 4 simultaneous bets, and for ease of discussion say they are all -150 and have a 80% chance of winning. If I ask for the full Kelly (I.e., Kelly multiplier = 1), the result is something like 3.25% per bet. If I set the multiplier to .5, the stake returned INCREASES to the neighborhood of 10%. If you go by .1 steps on the multiplier starting at .1, it starts at around 4%, peaks at around 10% at .5 multiplier, and goes down again until it's at the 3.25% for full Kelly (my numbers aren't exactly right but you get the gist -- don't want to go back and reenter everything in the claculator). Can anyone explain this?

[MonkeyF0cker]

It's not a malfunction. Single bet staking is higher, while 2, 3, and 4 team parlay staking is significantly lowered in this instance.

[MonkeyF0cker]

Notice that you have 93.75% of your bankroll at risk with full Kelly versus 41.9132% at half Kelly. Click on the numbered tabs (or All) to see your staking for the various parlays.

[boxcar]

Is this true for individual bets? I'm talking individual straight bets and not parlays?

[Pancho sanza]

I've had an odd result using the calculator and maybe someone can help explain. If I enter 4 simultaneous bets, and for ease of discussion say they are all -150 and have a 80% chance of winning. If I ask for the full Kelly (I.e., Kelly multiplier = 1), the result is something like 3.25% per bet. If I set the multiplier to .5, the stake returned INCREASES to the neighborhood of 10%. If you go by .1 steps on the multiplier starting at .1, it starts at around 4%, peaks at around 10% at .5 multiplier, and goes down again until it's at the 3.25% for full Kelly (my numbers aren't exactly right but you get the gist -- don't want to go back and reenter everything in the claculator). Can anyone explain this?

-150 @ 80 % win probability for full Kelly = bet half your bankroll, so for 4 identical bets, that would mean 2 X your bankroll is at stake.

It looks like the calculator has the required constraints built in for simultaneous bets.

[boxcar]

But why would the amount to bet be higher for half Kelly than it is for full Kelly? Doesn't make sense.

[RickySteve]

I'm not sure why you aren't understanding but your questions have been answered in the thread.

[boxcar]

Well I don't have a stats background, but I am an attorney. I don't say this to be arrogant but just to say that I'm not an idiot, and I don't understand how the answers address what appears to me to be a divergence from common sense that a half Kelly bet would be larger than a full Kelly bet for four independent and simultaneous wagers. I hope I don't sound disrespectful, don't mean to, but I don't get it. Why is the bet size smaller at .1 Kelly, larger at .5 Kelly, and smaller again at full Kelly?

[Pancho sanza]

Well I don't have a stats background, but I am an attorney. I don't say this to be arrogant but just to say that I'm not an idiot, and I don't understand how the answers address what appears to me to be a divergence from common sense that a half Kelly bet would be larger than a full Kelly bet for four independent and simultaneous wagers. I hope I don't sound disrespectful, don't mean to, but I don't get it. Why is the bet size smaller at .1 Kelly, larger at .5 Kelly, and smaller again at full Kelly?

Are you looking at the single bets tab, ie tab labelled 1? Look at the tab labelled "All", total bets are smaller under 1/2 Kelly.

The software is simply optimizing bet sizes across single bets, 2, 3 and 4 team parlays.

When you go from full to half kelly, the constraint becomes less of an issue and a higher % of the overall bets are thus allocated to the single bets, like any good optimization software.