Record Keeping
As any moderately serious poker player knows record keeping is important to a player's development and long term success. I would be interested in finding out how devoted and indepth people are in relation to this. Thus I pose the following:
Do you record results every single session? B&M, Internet and Home games.
How do you differentiate in your recording of ring games and tournament play?
When recording results, what do you record?
Do you regularly record/write down information on other players? B&M, Internet and Home games.
How often do you revisit/review the notes you have made?
How do you record the information? Pen and Paper? Computer spreadsheet? Computer program (if so which). Combination of them all?
Have you ever conveniently forgotten to record a bad session? Or do you catch yourself searching for reasons not to?
Any other comments are appreciated?
Do you record results every single session? B&M, Internet and Home games.
How do you differentiate in your recording of ring games and tournament play?
When recording results, what do you record?
Do you regularly record/write down information on other players? B&M, Internet and Home games.
How often do you revisit/review the notes you have made?
How do you record the information? Pen and Paper? Computer spreadsheet? Computer program (if so which). Combination of them all?
Have you ever conveniently forgotten to record a bad session? Or do you catch yourself searching for reasons not to?
Any other comments are appreciated?
Comments
For me, it is a real hobby only, but I really enjoy it, and would play much more on-line if I were single (catch my drift? lol). Having said all of that, I record every single session I play, in an excel spreadsheet that I made. I seperate every game type, so I track each buy in level for a S&G seperately, so I can zero in on the most profitable level for myself to focus on. I track all live games and on-line games sperately as well, so a $20 buy in one table tournament live would be seperate from an on-line one.
I record, date, time, buy in, result and general notes for each session. Overall, I calculate my hourly rates, rates per tournament, and EV for each level. I also track winnings per tournament by day of the week, just for interest sake. I also plan on creating a chart which breaks things down by time of day, such as afternoon / lunch time versus evenings, just to see if anything interesting results. I don't track the day of the week and time of day in anticipation of my game being stronger or weaker in one or the other, but I suspect a different calibre of competition shows up on-line Monday at 2 PM, versus Saturday night at 9 PM. I don't have nearly enough data yet to conclude anything, however early on the results are in line with what I would suspect (Sat games easier than Monday).
I'm a huge baseball stats kind of guy anyways, so the tracking part isn't a chore at all for me. I find it very interesting, and of course, my wife thought the baseball Strat league I do is nerdly enough, let alone poker stats...lol, they just don't understand.
I have never "omitted a bad session". No point, that's just lying to yourself, and you might as well not track a thing if you are doing that. The information you have is invalid at that point. If you are doing that, it would seem to me you aren't tracking results for yourself, rather as something you use for bragging rights.
Yes I record every single session. In statking you can manage a list of locations so I have different locations setup that are B&M, Internet and Home games. It allows me to produce summaries on any (or all) of those.
As far as differentiating between ring & tournament play, this I don't like that much in StatKing. But basically you manage a list of games and I have games setup such as $5-10 hold'em, $1-2 NL hold'em, Tournaments, etc. Basically I record them in the proper category and again I can produce summaries on whatever I want. When I set it up I setup Internet and Live action separately (i.e. I have a $10-20 Hold'em and a Internet $10-20 Hold'em). No good reason I did this because I can use my location to separate my results ... but I did. :-P
I record the basics in the program: Date, Location, Game, Time Played, Result. There is no spot for "other notes" which I would really like. When I recorded using an excel spreadsheet I would always have notes -- not sure how useful they were actually. :-P
I do not write down stuff on other players. I use it a bit online (player's notes) but probably not as much as I should.
Usually I look at my monthly results each time I play. It provides summaries based on time periods and I usually have mine seutp on the current month. About once a month I will look at some longer range stuff (how I am doing for the year etc) as well as how I am doing in individual games. If I see a short term trend (say I lost 3-4 sessions in a row playing $5-10 shorthanded) I will look at my historical results in that game as compared to other games (i..e is this shortterm fluctuation, have I played enough hours to get a reasonable gauge on how I am doing, should I consider playing in a different game?) Often it is more of a calming effect and I realize it is short term fluc and I should just work my way through it.
I can honestly say I have recorded every single session (good or bad) into the program. There are a few things I do, however, that it would probably be best not to do. Here are a few of the mistakes I make:
In a house game if I play a variety of limits/games I often record it all in the game we played the most. Say we play $1-2 NL for most of the night but at the end we switch to some mini-tournaments. Often I just put my total result in $1-2 NL.
Sometimes I do not record the time I spent playing as properly as I should. Often I know my result and record is 2-3 days later, but I can't recall the actual time played. Sometimes 8hr sessions were actually 9.5hr sessions ... this effects my earning / hour, but is still reasonable accurate.
Often I had trouble recording the tournaments. I went through a pattern where I would play a number of $7+1 tournaments and after I lose say $24 I would go to a live game and make that money back. Often I didn't record it at all (it was a wash). But again this affects my game results (tournaments lower, live game higher) as well as my earn / hour. I have since stopped doing that and have gone back to recording them individually.
The biggest problem I have is how I record freeroll tournaments (such as party poker's $7+1 I was just talking about). For example, say I played in 3 of them and I win the 3rd. Now all I get is a freeroll into a $64+6 tournament and say I lose that one. Have I lost $24 (3*8) or have I lost $86 ($2*8 + $70). I think you would say it is likely -$24.
Now take it one step further where I play in the 4th one and I win, then I win the $64+6. So now I am in a $600+40 tournament and I don't make the money. I guess this is another $8 loss .... bottomline is you are going to show a LOT of small losses and then potentially one huge win. Since I really want this to be a tracking of my long term results, butting in a big fluctuation like that isn't fair. Fact is I probably would not record a huge win either ... so putting in all those $8 losses isn't fair either.
I have considered doing a "tournament" database separately so I could track all those little tournaments and any big wins. I think it would be a big fat negative number (in fact I know it would) and it would not show me a good representation of my year to date.
Bottomline is I have problems recording these freerolls in a way I am happy with. Anyone with any ideas it would be very helpful. I guess I can also summarize by ring games to see my long term historical results and then show big picture by entering all the tournament results in.
Now when I play in single table sats (say $30+3) it is just money so I record that straight up (there is no freeroll aspect to it all). Those have contributed positively to my yearly result and I consider them as part of my normal play. Hmmmm I dunno.
Those are your answers in a BIG FAT message. Yes I rock.
The most accurate way to record these, I think is to simply record the satellite buy-in that you win as the same as cash. Consider it as exactly the same as a situation where you where you entered a $7+$1 tournament and won $70. Your profit is $62 for that tournament. Then you choose to enter another tournament for $64+$6. Say you don't win a prize in that tournament. That's a $70 loss. Your net P/L is an $8 loss, which is how you have previously (correctly) been recording the P/L.
What you have *not* recorded correctly by netting out the two tournaments is the your *variance*. Using your method will grossly underestimate your variance. (Technically, you will not even be able to get a proper estimate of your per tournament variance from this data at all.) This only matters if you care about using these records to estimate variance of course.
BTW, whether you made the choice to enter the $64+$6 tournament after winning $70 ($62 profit) in a $7+$1 cash tournament, or made this choice before entering a $7+$1 satellite tournament should be immaterial to the record keeping.
ScottyZ
-8
-8
+62 ($70 win -8 entry fee)
-70 (lose the $64+6)
-8
+62 (win another $8)
+570 (win $640 seat - $70 it cost to enter)
-640 (don't win in the big tournament)
I guess what I am saying is you are showing HUGE variance even though in reality you are only putting $8 up in each of them. I guess what I am saying is a lot of people, given the choice, would take the $640 cash rather then freerolling into the main event.
Now your records show the proper loss (it all nets to $8) but really unrealistic swings. I think if anything I would rather record the -$8 with the extra time it took to eventually lose that $8.
So the records would be more like ...
-8 -> 30mins
-8 -> 45mins
-8 -> 1.5hrs (got to play in $64+6)
-8 -> 15mins
-8 -> 4hrs (got to play in $64+6 & 600+40)
atleast this is tracking the proper net win/loss, and a true loss rate over the time you got to play without all the wild swings of freeroll money (which you never really had).
But then again who knows ... if you win a second one you would end up seeing a big +632 since they give you the $640 cash.
.kw
But, if you are interested in variance, the fact that your bankroll changes over time (even if your bankroll includes satellite buy-ins, or something like T$ or W$ at PokerStars) needs to be preserved in your data set.
Yes, the variance in that satellite example is high, but it is correctly so if you are in fact counting *per tournament* variance.
Note that if you are more interested in your variance per week (say) then you don't need to know your tournament resuts for each tournament. You just neet to net out your win/loss at the end of each week. (Don't forget to include T$ and W$ equivalents.)
The main thing I am worried about is that people (dangerously) miscalculating their variance from the data set 0, -8, -16, -24, -32, ... and concluding that playing satellite tournaments *really* cuts down their variance compared to cash tournaments. It does no such thing. All you're really doing is "hiding" the payouts by netting your results. This won't affect linear calculations (like win/loss) but will affect non-linear calculations (like variance).
ScottyZ
The varince calculation would be the same for satellites and cash tournaments since losses *and wins* count towards variance. The formula is:
sum (x_i - mean(x))^2
So, a large *positive* x_i (a win) will also contribute a lot towards the variance. Notice that variance is not in itself a bad thing necessarily--- it's a measure of deviations, but not necessarily in one specific direction.
Let's say that it goes
+62 (win)
+570 (win)
-640 (loss)
The per tournament mean P/L is -$8/3 = -$2.666...
The per tournament variance is about $370,000 (or most people prefer to think of standard deviation which is about $600).
The calculation is the same whether you were paid at each stage in cash, or if the prizes are some kind of buy-ins.
One important thing here is I keep saying "per tournament". This is vital. You *can* measure variance in pre-determined time intervals, such as per tournament, per day, per week, per year. However, you *cannot* properly measure variance by dividing your measurement intervals stocastically. That is, you can't stop your intervals at unknown random events, such as when you lose a tournament. That's why you can't calculate the variance properly with the data set 0, -8, -16, -24,...
The *illusion* (and possibly a large part of the appeal) of low variance for multi-stage satellite tournaments comes from the fact that people can easily convince themselves that individual wins never took place because they were never paid out in real money. Then this can easily be taken one step further so that people can explain away the *losses* in the target tournament(s) as never having taken place.
In fact, it can be argued that variance is *higher* in multi-satellite structures for highly skilled players. Why? Because winning entry via satellite into $640+$60 tournament might even be worth around $1000 in EV to a very good player. The win is worth more than a cash win in the sense of EV, and a loss will be more devistating.
Of course, I hope you noticed the flaw in the last argument. If a $640+$60 tournament buy-in is worth $1000 in EV to a good player, surely $700 in cash must be worth at least that much in EV too.
Actually, there is even a small flaw in the last counter-argument, since some skilled players might have better results in satellite tournaments than cash tournaments due to the flat payout structure.
I might as well stop arguing with myself and let everyone else jump in.
ScottyZ