mathew : Len ta opravim v tom, ze ak si hovoril ze malokto rozumie variancii tak si tym myslel ze malokto ma skusenosti s varianciou. Lebo rozumiet je podla mna rozumiet matematike, ktora za nou stoji a mat urcitu predstavu (vyjadrenu aspon priblizne CISLOM) ako je dana vec pravdepodobna.
BTW to co sa stalo tebe ked si siel na vyssie limity je smola, ale nemyslim ze je to menej pravdepodobne ako keby Slovensko prehralo v Armensku, doma s Irskom a este domaci zapas s Rusmi. (pravdepodobnosti tychto udalosti zoberme podla kurzov v stavkovych a Rusko berme ze ma tretinovu sancu u nas vyhrat).
Chcel som tym povedat, ze je to smola, ale nic extra nepravdepodobne.
Este chcem podotknut, ze som sa vyjadroval iba k mathewovej smole, nie ku 7 castlovej, lebo na to nemam podklady.
Ale ma to zaujalo natolko, ze som si urobil urcity vypocet ktory zverejnim

Na vypocet pravdepodobnosti som pozil nasledovny model :
Namiesto standardneho s&g som pocital turnaj s 3 ludmi, kde vitaz berie vsetko. Nakolko pri s&g berie tiez tretina ludi, je to podobne. Tento model ma NIZSIU varianciu ako klasicky s&g, ale nemalo by to byt velmi vyrazne a staci aby sme ziskali urcitu predstavu.
Pri modeli som bral 8% fee pre hernu a 1,29% ROI. Co povazujem za brutalny skill, lebo to znamena, ze ak by neexistovalo fee bolo by ROI 9,5%.
Tabulka udava pravdepodobnost, ze po 1000 odohratych turnajoch bude ROI nizsie, alebo rovnake ako uvedena hodnota :
ROI -5,14%. -2,57%. 0,00%. 2,57%. 5,14%
pravd. 5,32% 17,05% 38,48% 64,20% 84,57%

Tabulka pre 5000 turnajov :
ROI -5,14%. -2,57%. 0,00%. 2,57%. 5,14%
pravd. 0,01% 1,41% 23,61% 77,33% 98,65%

Ako je vidiet na 1000 ofohratych turnajoch stale existuje slusna sanca byt v minuse aj po RB, pri 5000 sa to uz priemeruje pomerne dobre.
Podotykam este raz ze uvedeny model ma o nieco nizsiu varianciu ako realne s&g

Neda sa inak...

Azntracker:

"For this example, let's assume the player plays 1 marathon session of 1,000 single table $210+15 turbo SNGs on P****Stars. 1,000 games is a standard benchmark number that apparently can determine if a player is a consistent winner. Let's also assume that the player is a decent regular and places 1st 13% of the time, and gets 2nd and 3rd 12% of the time, for a 37% ITM rate. The math works out to a 5% ROI which is pretty decent for those SNGs. After a simulation of this session 1,000 times, the result was a simulated ROI (ROI for each 1,000 game set) that ranged from as low as -8% to a whopping 25%. In only 23% of the 1,000 SNG sets was the simulated ROI within one percent of the actual 5%. This goes to show that a sample size of 1,000 games is hardly enough to predict your true ROI by any means.

Here are some more interesting ROI statistics in non-pargraph format, for the sake of easier of reading:

In 12% of those 1,000 SNG sessions, the player experienced a negative ROI
In another 10% of those sessions, the player had a 0% ROI.
What about upswings! In 14% of those 1,000 sessions, the player had more than double the “true” 5% ROI. He even had a few sessions of 16-17% ROI.

The program also outputs buy-in (BI) downswing statistics as well as ITM/OOTM streaks. Here were some interesting stats:

In 1% of sessions, the player experienced an 86+ BI downswing.
10% of the sessions had a 62 or greater BI downswing, with the greatest downswing at 123 buy-ins.
The most common downswing was 34 BI's.
In only 10% of sessions did the player experience a 24 BI or less downswing.
The player once had 13 ITM finishes in a row.
The most common OOTM streak was 12 games, with the max at an astonishing 32 in a row.
In 10% of sessions, the player experienced a 17 game or more OOTM streak.
In 58% of sessions, there was a 12 game or more OOTM streak.

Just for fun, a few simulations of a single session of 1,000,000 games was run. Some nifty stats – there was a 306 BI downswing, the longest ITM streak was 17 games, and the longest OOTM streak was 37 games!

One final practical calculation that yielded a telling point was the simulation of 365 sessions of 50 games each. Every calculation represented one “year” of playing. In doing so, the range of simulated ROIs obtained was 2.43% to 7.54%. In other words, if you were to play a 50 game session every day of the year, at the end of the year, your ROI could be as much as 50% off your actual ROI, given the assumptions made above. Even after that much grinding (18,000 games), after which incidentally you would be Supernova Elite on PS, your true ROI is still some fickle thing that simply depends on...variance.

Chalani, len aby bolo jasne (hoci som to pisal uz asi 100x ale pre istotu):

Dufam ze je jasne ze mne je vsetko toto absolutne jasne v teoretickej rovine, dokonca si myslim ze na rozdiel od mnohych je mi to jasne aj v praktickej rovine.

To nic nemeni na tom, ze je v celku pochopitelne ze sa nad tym pozastavim vo svojom blogu, a trochu ludsky sa aj postazujem, snad sa chapeme

Btw pre tych ktorych by to zaujimalo, odporucam kuknut SS, to co pred 2 dnami vyzeralo ako dno som som uz davno prekonal

Dneska si dam volny den, zajtra idem do Montesina na ten sialeny deepstack.

Nam je to jasne ze tieto veci su ti jasne Este nam to tu aj arakas podrobnejsie rozobral.

Tak som si bol vo Viedni vyvetrat hlavu od online, v Montesine na 550e Deepstack turnaji. 50k chips, prvy level 500bb deep, hodinove levely. Vedeli sme ze to bude dlhe, ale aj tak ma zaskocilo ked vysvitlo ze hrame prvy den 12 (!) levelov, na konci to bolo uz hodne huste

Ale so far so good, po Day 1 som s 234k solidne nad avg, komplet chipcout tu: http://www.montesino.at/index.php?optio ... Itemid=121, v sobotu pokracujeme od 1500/3000/400, takze velmi dobrych 80bb.

Zacnem 2 rukami, ktore som odohral fakt zle, jedna ma ale paradoxne nastartovala.

Este na prvom stole som mal asi 60k, ked pri 200/400 niekto limpol EP, lahko citatelny Rakusan isoraisol z CO na 1500, za nim na BTN Madar co hral dost necitatelne ale celkom solid (a mal asi 140k uz) call, tak som sa zamiloval na SB do A 5 , squeezol som to na nejakych 6100, Rakusan dost rychly call Madar fold. Nebola to uplne ideal situacia, ale u Rakusana som si bol isty ze ked netrafi cbet to vybavi, ze zlozi najhorsiu cast range, a ze bude pasivny postflop. Tiez rolu zohralo to, ze som predtym asi hodinu nehral, lebo sa original nedalo, co ma jednak uz trochu narusilo (bohuzial), ale tiez to sposobilo ze som predpokladal ze moj RR dostane nejaky kredit.

Flop AQ4 s jednym , nevidel som dovod cbetnut, a nasledne mi total zlyhal live tell Chlapik sa na 1sek tak zarazil, a p mojom checku dost rychlo hodil do potu 10k zeton, co mi prislo ako klasicke "mam KK, KQ, TT whatever a skusim to bluffnut", tak som si povedal ze skusim jednu street a uvidim ci to nebude stacit, pripadne ci mi nedojde druhy par alebo sa mi notvori FD. Turn 6, chlapik sa pomylil a myslel si ze ide prvy, insta hodil dalsich 10k, to uz som vedel ze je to presne naopak, ze ma AQ, tak som c/f, tvaril sa ze mam KK a samozrejme hrdo ukazal AQ. Citil som sa ako uplny debil, ale zostalo mi este 40k takze nebol dovod na paniku.

Druha haluz, ktora bola ale paradoxne klucovo pozitivna, sa mi udiala na novom stole. Mal som mozno 65k, ked pri 500/1000/100 raisol rakusky Turek UTG nejakych 3200, za nim znamy bratislavsky regular Oleg L call, dalsi zle hrajuci francuz call, tak ja s mojimi 99 si vravim ok, dead money by boli, UTG ma tak 30k, takze s nim si to zatocim, ak ma JJ+ so be it, ale nesparovanych ruk je predsa viac, este aj JJ by mohol foldnut, tak som tam poslal nejakych 14k, vsetci fold, Turek ship, ostatni fold.

Turek otoci QQ, ked zbada moje 99 tak teatralne zakrici "Yessss", to uz si vravim sam pre seba ze tesit sa je dobre az na riveri, samozrejme flop s prehladom A93, do riveru uz sa nic nezmenilo a ja som zhrabol pot. Vrchol ale prisiel potom, ked som si zratal zetony tak som zistil ze mam asi 120k, cize Turek nemal 30k, ale takmer 50k Doteraz to nechapem, ci som mal nejake zatmenie, keby som vedel ze ma 50bb tak ma hento ani nenapadne... ale bolo by to dopadlo velmi podobne, Oleg mal A9o, takze by som zobral miesto Turka jeho (kedze by som to bol na 95% overlimpol)

No a "kratko" k zvysku. Prvy stol som mal dobry, v zasade ziadni dobri hraci, mnozstvo limpovania, handy napisane na cele, ziadna kreativita (teda az na Madara, ktory isoraisol T2s, dostal 3 cally, na flope 962 alebo tak nejak vyhral cbetom 2 superov, double barreloval turn T a na riveri T uz si vychutnal Slovaka, ktory bol v tilte so 77 po predtym prehratej ruke), proste radost hrat, akurat som mal iba 3 value ruky, vzdy QQ, 1x mi zhodili PF, a dva razy som mal 2 streets of value, akurat nie vo velkych potoch.

Epic ruka bola ta co roztiltovala toho Slovaka, chalan limpol KQ UTG, za nim limp, za nim chlapik so stats 50/0 (ktory predtym limp/reraisoval takze evidentne limpoval vsetko) call, BB check. Flop KKQ, Slovak check, ultra limper overbetol pot, Slovak call, ostatni fold. Turn blank, zas to iste, check, potbet, call. River A, Slovak donkol nejak malo (bolo jasne ze ma JT na rivered str8 alebo KQ), typek vyzeralo ze foldne, ale zrazu tam poslal masivny overbet, Slovak samozrejme call a limper ukazal AA . Inak haluz je v tom, ze ja by som to KQ zlozil, lebo by som hentoho dal na AK, ale AA ma zabilo (overlimpoval ako 3.)

to be continued...

http://www.pokergaming.cz/Forum/ThreadDetail/1746/PokerGaming-Hero-profile-9-7castle

lol si trafil jediny out a este to aj napises tak ze sa nechumeli?
najs one

tucniak napísal:lol si trafil jediny out a este to aj napises tak ze sa nechumeli?
najs one

Hehe esteze to nevedel ten Turek Tak ale co mam napisat, musim mat aj ja niekedy stastie, a hlavne ma to vobec netrapilo, proste to je vrchol amaterizmu ked sa niekto teatralne tesi pred flopom...

Pekny rozhovor, mas u mna bod za ten rum A gl v montesine