Hours played

[Bobby Digital]

Assume I play online at 3 tables for 4 hours. Should I treat this as 4 hours of play or 12 hours of play?

[sj_poker]

It seems to me that, at least from a statistical standpoint, it would depend on the question(s) you want to answer.

If, for example, you want to see if your overall win rate is better playing three tables at once rather than one or two, then you would want to treat it as four hours. And use the data to do a difference of means test of this compared with your hourly rates playing one or two tables at once.

If, however, you want to know how playing an extra table or two affects your play at each table (i.e. whether you play better when you can concentrate on only one game at a time), you would treat it as 12 hours and you could compare this to your hourly rate(s) while playing one or two tables at a time by treating an hour at one table as an hour of play regardless of how many other tables you might be playing at the same time.

These questions are all sort of related, but the various distinctions between them are important -- for example while you may do wose at EACH TABLE by playing three rather than two tables, you may still do better overall. This could happen, for example, if you averaged 1BB/hr at each table while playing two at once, but only .7BB/hr at each table playing three at once. Then, you'd make 3*.7=2.1BB/hr playing three games compared with 2BB/hr playing two games even though at each table, your play suffers a bit from lack of attention.

My guess is that you want to answer a question like "Could I make more money by playing three rather than two tables at once?" In this case, you're interested in whether your hourly rate (where hourly here means actual hours of time spent) is better or worse, so you should analyze your results of how much total you win (or lose) each hour over all the tables you're playing. But it's not a bad idea to keep both sorts of records so that if you ever want to ask a different question, you will have the data.

Also, it's important to keep records of your actual status at every hour mark (not just to record that you were up 12 BB at the end of a 12 hour session). If you don't do this, the sort of inferences that you will be able to make from your data will be severely limited. So after every hour you play, record the level your chip stack at each of the tables you're playing. Then, at the end, you can have the option of doing both sorts of analyses.

Hope I didn't go overboard with the response, but I am a B&M player, so I hadn't really thought about the interesting ways internet players' record-keeping could differ from mine.

[Nottom]

[ QUOTE ]
Also, it's important to keep records of your actual status at every hour mark (not just to record that you were up 12 BB at the end of a 12 hour session).

[/ QUOTE ]

Maybe its just me, but I really don't think this is particularly important. Hourly measurements will just record variance throughout the session which isn't really all that important in most cases.

The only real benefit I can imagine is knowing if you perform worse either near the beginning of a session or after a certain amount of time. If the data shows for example that you tend to perform worse in the 4th hour maybe you should shorten your sessions to reflect that and take a break instead of continuing to play.

[sj_poker]

I'm just a beginning player, but I feel fairly confident in my grasp of the statistics behind most poker record-keeping, so take my ideas for whatever you want. I don't claim to be any sort of expert on this topic.

To me, the problem with not recording hourly stats is that, if you assume each "session" is the same as all the others, any inferences you attempt to draw from these data will be severely polluted.

At best, you will lose a ton of precision around your win rate. Think of measuring your car's fuel efficiency by recording how many days you go before you have to refill your tank. This will give you some idea of how it's doing, but if you want to get a meaningfull measure, you'd use miles per gallon since you don't drive the same distance every day. At worst, your results will be severely biased due to correlation between your winning/losing status and the amount of time you play in a given "session" as well as many other factors.

Even if you record the total number of hours played and total win/loss amount, but not the hourly stats, you will not be getting all of the information you could and, therefore, will lose a ton of power for your statistical analyses. If you're serious about keeping records and learning from them, it seems to me that hourly stats are by far the best way to do it. They're also much easier to interpret and analyze than info on sessions that vary in length.

All of the poker authors I've read (which, admittedly is just a few) have endorsed keeping hourly records. Once you have this data, you can begin to make inferences about your performance as well as how it changes with time, type of games played, and other factors.

[huzitup2]

1. The answer is 12 hours - case closed.

2. IF (and it's a big if) you find that your concentration level does not suffer - or suffers only a very small amount - there are TWO major benefits from playing multiple tables.

The first is you can play TWO tables of 2-4 with a much smaller bankroll than you need for one table of 4-8.

(Expert blackjack players can verify this; you need less money to spread 2 hands of 25-100 than you need to play 1 hand spreading 50-200).

The second advantage is that you will get into the "long run" much faster - AND have smaller "swings" - playing two games at half the stakes as opposed to 1 game at [full] stakes.

That said, I will state one more time; 3 tables for 4 hours is 12 hours of play. There is no disputing this.

I hope this was of some help.

- H

[nummerfire]

In the essay section on this site Mason describes a way of computing the standard deviation without needing to take notes every hour.

Kim

[emanon]

sj_poker makes some important points that other's seem to miss.

The important issue is "Why are you keeping statistics?"
Only when that question has been answered can you determine what you need to track.

Are you trying to make an estimation of how much you can earn over the course of a year?
What your overall standard deviation is?
Or are you primarily interested in how your 3 table performance compares to 2 table performance?

Once you can answer this, it will help you establish how much information you need to track, and what calculations you want to make.

[mrbaseball]

"1. The answer is 12 hours - case closed"

I would totally disagree with this. I play multiple tables online as a norm. If I play for 4 hours I have played for 4 hours. The 4 hours are relevent and actual. The number of tables I play is moot. I may play 1, 2, 3 or 4. How much I win (or lose) in those 4 hours is my actual performance for 4 actual hours of my life.

I am only concerned with actual time spent and actual money earned. But since I am primarily an online player I consider all of my earnings calculations from number of hands (ie bb per 100) rather than time. But I realize that the more hands (tables) I can play effectively the more money I can earn per hour.

[David Steele]

Mason's calculation is an estimate, but a good one.
I don't recall ever hearing 2+2 authors ( or any one else) suggesting you keep individual hour results.

D.

[TobDog]

[ QUOTE ]
I don't recall ever hearing 2+2 authors ( or any one else) suggesting you keep individual hour results.

[/ QUOTE ]
Lou Krieger mentioned it in his book on how to keep track of your standard deviation.