The Results Are In!!!! Time of Day---> ROI/ITM Correlation Test... [Archive]

Messy_Jesse

08-11-2005, 04:00 AM

Ok. I compiled the list of stats from this thread. (http://forumserver.twoplustwo.com/showthreaded.php?Cat=&Board=singletable&Number=308 2351&Forum=&Words=&Searchpage=0&Limit=25&Main=3082 351&Search=true&where=bodysub&Name=22981&daterange =1&newerval=1&newertype=w&olderval=&oldertype=&bod yprev=#Post3082351) Basically, I had people submit their results based on the time of day, and compiled and weighted the results to find how people's ROI and ITM vary according to the time of the day. After a sh!tload of painstaking data entry, I came up with some general stats. They encompass approximately 10,500 tournaments, so should be pretty accurate. Here are the results:

ITM ROI #Tourneys

Morning: 36.8% 11.9% 654
Forenoon: 37.2% 7.9% 830
Afternoon: 37.3% 8.8% 2088
Dinner: 38.6% 16.4% 2720
Evening: 38.9% 18.1% 2994
Night: 37.9% 14.2% 2077

A quick look at those stats, and you can see the bell-curve nature of it. The dinner and evening stats (as I expected) were significantly better than those of other times during the day.

Comments?

P.S. I realize that there are a few spots where the accuracy of these results could be called into question- however, I believe that these are trivial in nature and that for the most part, these results are statistically significant.

-Jess

Shilly

08-11-2005, 04:02 AM

Thanks for compiling all of that data. I don't think the results are all that interesting, however--it basically just confirms what most of us seemed to think.

Myst

08-11-2005, 04:06 AM

[ QUOTE ]
Thanks for compiling all of that data. I don't think the results are all that interesting, however--it basically just confirms what most of us seemed to think.

[/ QUOTE ]

You could be a a little bit more grateful for someone who takes the time to actually confirm, through evidence, what seems to be common sense.

Thanks Jess.

raptor517

08-11-2005, 04:09 AM

nice post. i wanted to do that but im too lazy. anyways.. i play maybe .5% of my games in the forenoon period. im just NEVER awake at that point. 8-12.. thats just out of my range for awakeness. nice luckbox btw /images/graemlins/wink.gif holla

Shilly

08-11-2005, 04:10 AM

[ QUOTE ]
[ QUOTE ]
Thanks for compiling all of that data. I don't think the results are all that interesting, however--it basically just confirms what most of us seemed to think.

[/ QUOTE ]

You could be a a little bit more grateful for someone who takes the time to actually confirm, through evidence, what seems to be common sense.

Thanks Jess.

[/ QUOTE ]

I did thank him.

TT_fold

08-11-2005, 04:11 AM

[ QUOTE ]
Ok. I compiled the list of stats from this thread. (http://forumserver.twoplustwo.com/showthreaded.php?Cat=&Board=singletable&Number=308 2351&Forum=&Words=&Searchpage=0&Limit=25&Main=3082 351&Search=true&where=bodysub&Name=22981&daterange =1&newerval=1&newertype=w&olderval=&oldertype=&bod yprev=#Post3082351) Basically, I had people submit their results based on the time of day, and compiled and weighted the results to find how people's ROI and ITM vary according to the time of the day. After a sh!tload of painstaking data entry, I came up with some general stats. They encompass approximately 10,500 tournaments, so should be pretty accurate. Here are the results:

ITM ROI #Tourneys

Morning: 36.8% 11.9% 654
Forenoon: 37.2% 7.9% 830
Afternoon: 37.3% 8.8% 2088
Dinner: 38.6% 16.4% 2720
Evening: 38.9% 18.1% 2994
Night: 37.9% 14.2% 2077

A quick look at those stats, and you can see the bell-curve nature of it. The dinner and evening stats (as I expected) were significantly better than those of other times during the day.

Comments?

P.S. I realize that there are a few spots where the accuracy of these results could be called into question- however, I believe that these are trivial in nature and that for the most part, these results are statistically significant.

-Jess

[/ QUOTE ]

Wow, these numbers are far from trivial. The jump in ROI from afternoon to dinner is astounding! This makes me want to reconsider playing a bunch of SNGs during the day tomorrow. Why do all the cool social events have to occur during Party's fishiest hours? /images/graemlins/frown.gif

Sabrazack

08-11-2005, 04:27 AM

I am assuming all these times are after the partypoker time. But my question is, between what hours do you count the different "times"? Like, when is dinner?

raptor517

08-11-2005, 04:34 AM

[ QUOTE ]
I am assuming all these times are after the partypoker time. But my question is, between what hours do you count the different "times"? Like, when is dinner?

[/ QUOTE ]

if the memory serves right.. im too lazy to look it up.. dinner is 4-8, evening 8-12, night 12-4, morning or whatever 4-8, forenoon 8-12, and whatever else 12-4. im tired, cant think. holla

Newt_Buggs

08-11-2005, 04:42 AM

Awesome post, thanks Jesse

The difference may not be nearly as dramatic as it seems though. 10,000 is a lot of tournaments, but spread across the time intervals it isn't too many. I'm sure that a confidence interval on the 654 for morning will say that it the numbers don't show you too much for that time period.

ilya

08-11-2005, 04:47 AM

Thanks a lot Messy /images/graemlins/laugh.gif

One thing that's interesting about these results is that while the ROI differences are dramatic, the ITM differences are quite small. What's going on? Are daytime opponents better HU? Or are they better at accumulating chips early on while 2+2ers are at their weak-tightest?

sobroke

08-11-2005, 05:37 AM

great info.....this confirms my thoughts......it be even better if you had a breakdown of the different levels of buyins too.....great post tho

bennies

08-11-2005, 05:44 AM

Quote:
--------------------------------------------------------------------------------

I am assuming all these times are after the partypoker time. But my question is, between what hours do you count the different "times"? Like, when is dinner?

--------------------------------------------------------------------------------

if the memory serves right.. im too lazy to look it up.. dinner is 4-8, evening 8-12, night 12-4, morning or whatever 4-8, forenoon 8-12, and whatever else 12-4. im tired, cant think. holla

--------------
-------------

Could anyone confirm this? And are we talking American East Coast time?

Thanks for posting by the way! I'm a little surprised the difference is not bigger but it's definatly big enough!

The once and future king

08-11-2005, 05:45 AM

Messy Jesse. Great work, thanks.

Could you please confirm the time ranges for each period. If you could also add an asjustment for GMT that would be fantastic.

Myst

08-11-2005, 06:26 AM

The times are from Aleo's spreadsheet, so thats why some people understand them, and others dont.

12:00 AM-4:00 AM (Night)
4:00 AM-8:00 AM (Morning)
8:00 AM-12:00 PM (Forenoon)
12:00 PM-4:00 PM (Afternoon)
4:00 PM-8:00 PM (Dinner)
8:00 PM-12:00 AM (Evening)

Something Jesse didnt consider is that the times are totally client side, not server side based... In other words, what my time is on my computer is what's recorded on the spreadsheet. Hopefully though, most of the times are given in EST :P

bennies

08-11-2005, 06:33 AM

lol

"P.S. I realize that there are a few spots where the accuracy of these results could be called into question- however, I believe that these are trivial in nature and that for the most part, these results are statistically significant."

AA suited

08-11-2005, 11:05 AM

hm.. i must be doing something wrong at dinner time (all 50+5)

Time #SnG Top 3 ROI
Night 150 40.0% 32.7%
Morning 4 25.0% -54.5%
Forenoon 126 42.9% 29.1%
Afternoon301 40.2% 28.7%
Dinner 156 32.7% -3.3%
Evening 257 37.4% 19.9%

# SnGs = 1000
overall roi = 21.6%

btw- can you break up the #'s into each level. ie: individual stats for $11/22/33/55/109/215?

edit:
Eastern time

Myst

08-11-2005, 11:07 AM

[ QUOTE ]
hm.. i must be doing something wrong at dinner time (all 50+5)

Time #SnG Top 3 ROI
Night 150 40.0% 32.7%
Morning 4 25.0% -54.5%
Forenoon 126 42.9% 29.1%
Afternoon301 40.2% 28.7%
Dinner 156 32.7% -3.3%
Evening 257 37.4% 19.9%

# SnGs = 1000
overall roi = 21.6%

btw- can you break up the #'s into each level. ie: individual stats for $11/22/33/55/109/215?

[/ QUOTE ]

Ok, now give the one in which you had your bad streak over 640 sngs.

nWirb

08-11-2005, 11:10 AM

For nordic, italian and german players (GMT+1):

 ....................................ITM .. ROI ... #T 
Morning ... 10:00-14:00: 36.8% 11.9% . 654
Forenoon . 14:00-18:00: 37.2% . 7.9% . 830
Afternoon .18:00-22:00: 37.3% ..8.8% 2088
Dinner ..... 22:00-02:00: 38.6% 16.4% 2720
Evening ... 02:00-06:00: 38.9% 18.1% 2994
Night ....... 06:00-10:00: 37.9% 14.2% 2077

Sabrazack

08-11-2005, 11:13 AM

This is exactly the same as OP isn't it?

Edit: Ignore me please.

nWirb

08-11-2005, 11:24 AM

No, you're way to cute to ignore (I assume your avatar is a photo of you).

Anyways, if Swedes like us would like to turn pro it would be best to adjust our "working hours" to somewhere betwen 22-10.

Myst

08-11-2005, 12:00 PM

[ QUOTE ]
No, you're way to cute to ignore (I assume your avatar is a photo of you).

Anyways, if Swedes like us would like to turn pro it would be best to adjust our "working hours" to somewhere betwen 22-10.

[/ QUOTE ]

Omg your avatar is the most disgusting thing I have ever seen in my life. Lol.

DesertCat

08-11-2005, 12:17 PM

Thanks, Jesse. But besides adjusting all times to EST, I think there is another statistical issue to consider.

Is it possible that the results are skewed by the 2+2'rs who supplied you with data? I.e. if all the best players you polled prefer to play at night, wouldn't that improve the results for that time period?

I suspect it's both, that evenings are both fishier, and are when the best 2+2'rs make a living. Also my experience multitabling limit tables directly supports your conclusion, as I was always able to find higher VPIP tables in the late afternoons/early evenings (PST).

Myst

08-11-2005, 12:27 PM

[ QUOTE ]
Thanks, Jesse. But besides adjusting all times to EST, I think there is another statistical issue to consider.

Is it possible that the results are skewed by the 2+2'rs who supplied you with data? I.e. if all the best players you polled prefer to play at night, wouldn't that improve the results for that time period?

I suspect it's both, that evenings are both fishier, and are when the best 2+2'rs make a living. Also my experience multitabling limit tables directly supports your conclusion, as I was always able to find higher VPIP tables in the late afternoons/early evenings (PST).

[/ QUOTE ]

Well isnt late afternoon for you early evening for those on the east coast? And early evening for you late evening for thoese on the east coast? Lol.

Freudian

08-11-2005, 12:30 PM

[ QUOTE ]

Is it possible that the results are skewed by the 2+2'rs who supplied you with data? I.e. if all the best players you polled prefer to play at night, wouldn't that improve the results for that time period?

[/ QUOTE ]

Of course it is. If most of these results are from 10+1 it may not be all that valuable to a 100+9 player.

But basically all it does is that it somewhat confirms what we already suspected. The more players on the Party Poker network, the more beatable the games are.

schwza

08-11-2005, 01:05 PM

[ QUOTE ]
Thanks, Jesse. But besides adjusting all times to EST, I think there is another statistical issue to consider.

Is it possible that the results are skewed by the 2+2'rs who supplied you with data? I.e. if all the best players you polled prefer to play at night, wouldn't that improve the results for that time period?

I suspect it's both, that evenings are both fishier, and are when the best 2+2'rs make a living. Also my experience multitabling limit tables directly supports your conclusion, as I was always able to find higher VPIP tables in the late afternoons/early evenings (PST).

[/ QUOTE ]

i would guess the 2+2'ers playing during the day are better for the same reason that the competition is better. schlubs like me are at work while pros are home playing.

does anyone play higher at night than during the day?

ChuckNorris

08-11-2005, 01:21 PM

Very nice post. I'm just gonna have to keep living like some freaking vampire. No such luxury as daylight for us europeans /images/graemlins/frown.gif

Thanks to the guy who converted the times to German time, I think it's like an hour off the time in Finland.

J-Lo

08-11-2005, 01:25 PM

i play higher during the nights-- my regular game is 4 tables of $55's after 5pm and 9 tables of 22's before 5pm. i hate the $33's

Slim Pickens

08-11-2005, 01:56 PM

[ QUOTE ]
Something Jesse didnt consider is that the times are totally client side, not server side based... In other words, what my time is on my computer is what's recorded on the spreadsheet. Hopefully though, most of the times are given in EST :P

[/ QUOTE ]

In the original thread, I gave directions for how to adjust your spreadsheet to any time zone. Depending on how people import their data into the sheet, it might be different. PT records all times Eastern, and there's a check box to correct that to your local time zone in the autoimport macro, but some people in other time zones might not use it.

Thanks a lot Jess. Now we've got something to point to when people ask "r the 22s getting hsrder in the day????"

SlimP

morello

08-11-2005, 02:16 PM

One question: Is the proportion of high buyin tournaments (109/215) higher during the "off-peak" hours?

It's not unreasonable to assume that the people playing the highest buyin tournaments are pros, and therefore play on a different schedule than the people with "real" jobs who only play during the night.

(meaning the effect of time may be exaggerated due to not also reflecting the lower average ROI)

Just a thought, it might not be true at all.

Slim Pickens

08-11-2005, 03:13 PM

[ QUOTE ]
Now we've got something to point to when people ask "r the 22s getting hsrder in the day????"

[/ QUOTE ]

I like how this happened about 1/2 hour after I posted.

ilya

08-22-2005, 08:13 PM

[ QUOTE ]
Quote:
--------------------------------------------------------------------------------

I am assuming all these times are after the partypoker time. But my question is, between what hours do you count the different "times"? Like, when is dinner?

--------------------------------------------------------------------------------

if the memory serves right.. im too lazy to look it up.. dinner is 4-8, evening 8-12, night 12-4, morning or whatever 4-8, forenoon 8-12, and whatever else 12-4. im tired, cant think. holla

--------------
-------------

Could anyone confirm this? And are we talking American East Coast time?

Thanks for posting by the way! I'm a little surprised the difference is not bigger but it's definatly big enough!

[/ QUOTE ]

I think most of the people reporting for this survey are from the EST zone, and the reason I think that is that the "Night" stats are better than I would expect them to be for late-night-west-coast/late-night-early-morning-east coast. They seem more consistent with late-evening-west-coast/night-east-coast.

Messy_Jesse

08-22-2005, 09:05 PM

After looking at the responses, it seems that the biggest concerns revolve around the sample size and quality of players that submitted data and their corresponding influence on the results.

Although I agree that these are valid concerns, I think that one observation clears this up. If you go back to the original thread and look at each individual player's data, in basically EVERY single case, each player's personal results corresponded to the ultimate conclusion that peak hours are more profitable. i.e. for each person's data, their dinner, evening, and nightime ROIs were higher than their ROIs during other periods of the day. Thus, the cumulative results reflected the individual results of all the players contributing the data, which, in my mind, goes a long way towards confirming the ultimate conclusion.

-Jess

Myst

08-23-2005, 12:13 AM

[ QUOTE ]
After looking at the responses, it seems that the biggest concerns revolve around the sample size and quality of players that submitted data and their corresponding influence on the results.

Although I agree that these are valid concerns, I think that one observation clears this up. If you go back to the original thread and look at each individual player's data, in basically EVERY single case, each player's personal results corresponded to the ultimate conclusion that peak hours are more profitable. i.e. for each person's data, their dinner, evening, and nightime ROIs were higher than their ROIs during other periods of the day. Thus, the cumulative results reflected the individual results of all the players contributing the data, which, in my mind, goes a long way towards confirming the ultimate conclusion.

-Jess

[/ QUOTE ]

English please, for those less educated.

You geek.