Everyone knows the answer to 2+2 is 5. Duh.

Do you want to argue this to... it's just math.

#Clown.

Haha this is the highlight of all these exchanges. Was supposed to be new regulations in the UK that would stop bookies acting like ***** on withdrawals but its probably worse then ever.

Not intellectually dishonest, just overlooked in the context of the discussion. You're right of course.

LOL what? How is half Kelly unacceptable for a <10k bankroll? Other than some very contrived scenarios, Kelly is pretty much always the correct framework for betting. And it's definitely the correct approach for OP. The advice should be to bet full Kelly, which for him currently is zero.

You bagged 2.15 about a 2.00 perceived chance, half-kelly would reccomend a bet of about 3.25% of Roll..... too much.

Do that too many times and as Thorp says, betting too much even though each bet is +EV can be ruinous.

He doesn't know what his edge is at this point.

Let's discuss. This requires certainty of inputs, does it not?

We live in a world of probablities, not certainties, do we not?

How do you propose pulling this off and living to tell about it?

Also, I'm a dyed-in-the-wool Kelly guy, I have no issue w/ the use of Kelly, it's the chosen multiple in this case (.50).

Man who could ever live to tell about betting full Kelly on a $200 bankroll, which isn't even a Kelly bankroll unless OP lives in a favela or something.

Also and I'm not trying to patronize you but you do understand that there is a 50% chance of having a 50% Drawdown w/ FK.... In the real world, the Withdrawal Risk using this approach is massive.

You truly expect a Newbie to assess his edge w/ the granular accuracy necessary to use FK? Your edge isn't some point-estimate calculation.

Ok. So that's how it is. The fact is, I expected as much.

Full Kelly for him is zero. HTH. What is "withdrawal risk" (another weird non-standard term)?

F**k this, I'm out. Basically, System abandonement. The emotional hazing of a large Drawdown is just too much to take.

So what bankroll size are we talking about in your version of a "real world"? If you can't handle large swings in your bankroll, this isn't the field for you or OP.

My version of real world wasn't referring to BR size, it was referring to human nature.

I never used FK and would never reccomend it, I prefered some assumed humility w/ Fractional Kelly.

hedgie,

Fair. You're one of the few guys I will reread if I get a response that doesn't confirm.

All,

Parroting Ganch (RIP, maybe should message him to see if he is dead or not), but Kelly BR is "beg borrow steal" BR. Child sex slaves, dick sucking behind dumpsters (S/o poogs), etc.

Half Kelly for your actual deposited amount is generally super conservative in this context. Or you're an idiot and super rich already. k.

Damn. How am I contributing in a real way? I just came for the lolz.

4 page discussion about bet sizing lol who really cares at the end of the day

how many people really get the chance to bet full kelly anyways?

**** like this and modelling are all just face value ****

Some of the GameBred posts are really dumb.

Fractional Kelly for "human nature" and not the skew of risk from overbetting vs underbetting? Variance moons and growth stalls out pretty fast overbetting! Then again almost all of us base our BR on money we put in a separate arbitrary box called "bankroll" which represents a tiny fraction of our Kelly BR. So even "full" kelly is super super tiny fraction when assessed in this light.

You're so Deeply Confused it's comic relief.

2+2=5

I still havent gotten over "trading your equity curve".

You need to get over yourself first; good luck w/ that.

lol good one

youre a finance guy, arent you? The wanna be verbiage gives it away. Gotta say... perfectly on-brand. Over confident without knowing the very basics, lots of ridiculous words that legit NO one in the industry ever uses, super defensive, loves to 'give advice' and call anyone who disagrees with whatever it is youre saying in here (still no idea) an idiot.

LMAO. I give my word as a Gentleman, I don't know the first thing about Financial Trading (Cue joke about the same applies to Sports Investing).

I may not know the basics but I do know that 2 + 2 = 5...

#WashedKing

Hi! I have a question about Kelly. My betting models based on retrospective data: I collect different football statistics and also collect opening and closing bookmakers odds. Then I set up filters in the way they show positive trend in the past. Of course, it’s obvious that in the future the size of ROI will be less than in the past. So how should I manage my bankroll due to Kelly if everything that I know is ROI only for the retrospective models in the past?

You could start forward-testing your model and use Bayesian updating to estimate your ROI. Start with a prior distribution that assumes you're losing to the vig and is weighted/centered toward -EV's. As your sample grows, your posterior distribution (which gets continually fed back in as your new prior distribution) will become more accurate. If/when your posterior distribution says you're +EV, you can start kelly-betting according to your Bayesian ROI, but keep updating your ROI and adjusting your bet size accordingly.

At least, that's how I'd do it. An alternative would be to gather a big enough sample for a desired confidence interval, ie the Frequentist approach.

IMO the key is not to assume a positive ROI based on results from a period whose data your model was trained on. Any recent data your model wasn't trained on is fair game for estimating your ROI.