Nba playoffs thread

[QUOTE=TheDude;41936]Diff between 11.5 and 12 is worth a double unit? I am not the math guy obviously but could you explain this on such a non key #? Thx[/QUOTE] Maybe a little more explanation might be useful. Even though the spread +12 isn't a key number, the NBA push rate is roughly 3.6% so it is fairly significant. So this is how Fezzik correctly gets the 1.8% difference in win rate between +11.5 and 12. However if we assume the +12 two-weight bet (assuming 1 wgt = 1% of bankroll) is a 55% winner, the fractional Kelly is 0.3628. Certainlly this ia a reasonable Kelly fraction to bet given the uncertainites involved with a NBA spread of 12. At the same Kelly, the 11.5 should be bet at 0.624 units. So with these assumptions, probably a half weight bet at +11.5 is more appropriate. As Fezzik points out, with the 55% win assumption, the +12 bet is about 3 times better EV than the +11.5. However, making the +12 bet a three-weight increases the Kelly fraction to around 0.58 which doesn't seem justified and is probably much too high. Hope this helps.

[QUOTE=TheDude;41947]Next question: On 4-17-(sunday) when you recommended OKC UNDER 204.5 as a 2weight, and then the total went to 206. [B]Q?[/B] is how big of a weight does it become for each 0.5 point increment there? Thanks again for your time.[/QUOTE] An NBA total of around 206 has a push rate of about 2.3%. Using the same logic as above, i.e. if the 2 wgt bet UNDER 204.5 is a 55% bet, then the Kelly fraction is 0.3628 %. Each half point roughly gives an extra 1.15% edge in win rate and at the same Kelly fraction that gives an extra 0.88 units per 0.5 point.

You're saying, mathematically, that two units at +12 means .624 units at +11.5. Is this because 1.376 is the square root of 1.8? Knowing this would make configuring bet size much easier for me. [QUOTE=ComptrBob;41948]Maybe a little more explanation might be useful. Even though the spread +12 isn't a key number, the NBA push rate is roughly 3.6% so it is fairly significant. So this is how Fezzik correctly gets the 1.8% difference in win rate between +11.5 and 12. However if we assume the +12 two-weight bet (assuming 1 wgt = 1% of bankroll) is a 55% winner, the fractional Kelly is 0.3628. Certainlly this ia a reasonable Kelly fraction to bet given the uncertainites involved with a NBA spread of 12. At the same Kelly, the 11.5 should be bet at 0.624 units. So with these assumptions, probably a half weight bet at +11.5 is more appropriate. As Fezzik points out, with the 55% win assumption, the +12 bet is about 3 times better EV than the +11.5. However, making the +12 bet a three-weight increases the Kelly fraction to around 0.58which doesn't seem justified and is probably much too high. Hope this helps.[/QUOTE]

Good explanation Bob. Also, it's important to point out the circumstances of why you are getting the +12 or under the 206 instead of the 204.5. The reason we can't bet full kelly is because we don't have 100% certainty of the edge on our bets. When the market gives us a better price, it also should add some amount of doubt to your original bet. Best case scenario, you are getting a new +12 at a rogue square books and the market is staying around 11.5. No reason to add doubt to your original bet, and you can grab some +12 to add to your position. If the whole market goes to +12, there's a decent chance your +11.5 wasn't as good to begin with and you should sit tight. Now, you could chalk it up to public money betting the big favorite Bulls and still justify your original play here and add some. The case of a total with the whole market moving against the play should be the most concerning. You are more likely looking at a situation where you were around 48%-50% on the original bet.

Thanks Bob and Twoniner Good explanation. Now the question for each individual is to use Kelly (or 1/4 or 1/2 Kelly) etc...or not. Good Luck

NYKNICKS +8-135 greek 1 weight I would buy up the +7-115 to +8-135. Worth a 1 unit bet IMO......... at +7-115 OR +8-135 (SLIGHT preference to +8,since +7 is SUCH A HUGE MONSTER)

724 ORL -9 1 weight zig ZAG missed the -8.5, but no way I see atl winning this game, given this -9 worth a bet.

[QUOTE=joelshitshow;41950]You're saying, mathematically, that two units at +12 means .624 units at +11.5. Is this because 1.376 is the square root of 1.8? Knowing this would make configuring bet size much easier for me.[/QUOTE] Nope, the square root of 1.8 is 1.342, but you have to use the actual Kelly formula to calculate the "equivalent" unit bet for a different spread.

[QUOTE=TheDude;41968]Good explanation. Now the question for each individual is to use Kelly (or 1/4 or 1/2 Kelly) etc...or not. Good Luck[/QUOTE] Yes, good points by TwoNiner. IMHO, the Kelly criterion should never be applied blindly to sports betting. Uncertainty in never knowing your exact edge and working with small and dynamic sample datasets vitiates much of its usefulness. However, it is useful in proportionally comparing different bets at different odds and spreads as we have done above. Certainly, full Kelly should only be rarely, if ever, bet in sports. I would recommend [B]never betting over 1/2 Kelly.[/B]

[QUOTE=ComptrBob;42014]Certainly, full Kelly should only be rarely, if ever, bet in sports. I would recommend [B]never betting over 1/2 Kelly.[/B][/QUOTE] And for the markets where hedge funds and prop firms bet Kelly x10? Good posts as always, CB.