Break Stats -- Turning Stone XX 9-Ball, Jan. 2013

[AtLarge]

Aggregating the numbers for pushes in the 3 most recent Turning Stone events (XX, XIX, and XVIII), we have the following.

Out of 622 games in 43 streamed matches, 78 games (12 1/2%) involved a push out, and the results were as follows:

• Breaker pushed and won the game -- 26
• Breaker pushed and lost the game -- 22
• Non-breaker pushed and won the game -- 20
• Non-breaker pushed and lost the game -- 10

So, overall, the person who pushed won 46 of the 78 games (59%) and lost 32 (41%). Both breakers and non-breakers who pushed won the majority of their pushes. This contradicts conventional wisdom.

Of the 78 pushes, 32 were returned (passed back to the pusher to shoot), and the pusher won 21 of those 32 games (66%).

 

[CreeDo]

The push stats are a nice unexpected bonus, thanks very much

I was always curious whether pro-level push plays were truly 'traps' as some like to claim, or just "I'll give the other guy some rope and hope he hangs himself".

I wonder if the pro player's ego is part of the reason the "pushee" seems to have less success.

Basically, the guy who steps to the table after the push thinks "I'm a pro, I should be able to handle anything situation this guy throws at me. Whatever he thinks he sees, I also see. If he thinks that shot's too hard for me, I'm about to show him how wrong he is." Their ego won't let them pass back as often as they should.

 

[jalapus logan]

The push stats are a nice unexpected bonus, thanks very much

I was always curious whether pro-level push plays were truly 'traps' as some like to claim, or just "I'll give the other guy some rope and hope he hangs himself".

I wonder if the pro player's ego is part of the reason the "pushee" seems to have less success.

Basically, the guy who steps to the table after the push thinks "I'm a pro, I should be able to handle anything situation this guy throws at me. Whatever he thinks he sees, I also see. If he thinks that shot's too hard for me, I'm about to show him how wrong he is." Their ego won't let them pass back as often as they should.

Yes, thanks again! Looks like the pusher has a 9% edge on the pushee. Beware the pusherman!

 

[Bob Jewett]

Aggregating the numbers for pushes in the 3 most recent Turning Stone events (XX, XIX, and XVIII), we have the following.

Out of 622 games in 43 streamed matches, 78 games (12 1/2%) involved a push out, and the results were as follows:

• Breaker pushed and won the game -- 26
• Breaker pushed and lost the game -- 22
• Non-breaker pushed and won the game -- 20
• Non-breaker pushed and lost the game -- 10

So, overall, the person who pushed won 46 of the 78 games (59%) and lost 32 (41%). Both breakers and non-breakers who pushed won the majority of their pushes. This contradicts conventional wisdom.

Of the 78 pushes, 32 were returned (passed back to the pusher to shoot), and the pusher won 21 of those 32 games (66%).

Thanks for the info.

When the stats are based on a fairly small number of events, you have to consider the significance of the results. In general, if an event (like "push and win") happens N times, then the normal variation in the number is expected to be something around the square root of N. So, if you see 4 of a certain kind of outcome in match, the "fuzziness" of the result is something like +-2 on the count.

You most often see this rule applied to polls where the often say something like "a thousand people were polled so the statistical uncertainty is +- 3%." Here 3% is 1/sqrt(1000) or about 1/30.

The fuzziness decreases if you include more events and so get larger numbers overall -- you get a better average.

 

[Jude Rosenstock]

Thanks for the info.

When the stats are based on a fairly small number of events, you have to consider the significance of the results. In general, if an event (like "push and win") happens N times, then the normal variation in the number is expected to be something around the square root of N. So, if you see 4 of a certain kind of outcome in match, the "fuzziness" of the result is something like +-2 on the count.

You most often see this rule applied to polls where the often say something like "a thousand people were polled so the statistical uncertainty is +- 3%." Here 3% is 1/sqrt(1000) or about 1/30.

The fuzziness decreases if you include more events and so get larger numbers overall -- you get a better average.

This is, quite literally, statistics 101. As in, I actually learned all of this in my statistics class in college a couple decades ago. Yeah, the standard deviation from the mean is related to the distribution of events and the number of events. I did great in the class but it's been too long to be able to regurgitate any relevant formulas. I will say this, all of the formulas were ridiculously long. When you're looking at a sample size of only a few dozen win/loss outcomes, the distribution can be all over the place. What's more, the statistics in something like this are inherently tainted since it's winner-breaks and oftentimes the winner is also a better player. So, it's not a perfect evaluation of the importance of the push necessarily. Of course, more data would have more conclusive results.

So, in order to get a more accurate understanding of these statistics, you might actually need to plot out a few other charts. Like, for example, what was the winning percentage of the breaker when he did not have to push versus pushing? What was the winning percentage of the non-breaker? I'm not saying this necessarily paints a complete picture but it might give us greater insight to its statistical significance.

 

[lewis_rva]

Event, B & R's , Total Games Played, B & R %
Tunica 10-Foot 10-Ball July 2012, 13, 150, 8.67%
Derby City Classic 10-Ball Jan. 2012 22, 115, 19.13%
Turning Stone XVII 10-Ball 70, 375, 18.67%
10-Ball Masters event in Virginia 49, 256, 19.14%
SBE 10-Ball Pro Players Championship in Pennsylvania 21, 119, 17.65%
Johnny Archer Classic 10-Ball, Oct. 2012 19, 86, 22.09%

Total 194, 1101, 17.62%

Including the Tunica 10 foot 10 ball stats skews the 10 ball difficulty. If those games are not included the 10 ball break and run percentage is 19.0%.

 

[AtLarge]

Thanks for the info.

When the stats are based on a fairly small number of events, you have to consider the significance of the results. In general, if an event (like "push and win") happens N times, then the normal variation in the number is expected to be something around the square root of N. So, if you see 4 of a certain kind of outcome in match, the "fuzziness" of the result is something like +-2 on the count.

You most often see this rule applied to polls where the often say something like "a thousand people were polled so the statistical uncertainty is +- 3%." Here 3% is 1/sqrt(1000) or about 1/30.

The fuzziness decreases if you include more events and so get larger numbers overall -- you get a better average.

I certainly agree, Bob, and I often mention the "small number" caveat. But I still enjoy looking at the results.

 

[pt109]

Aggregating the numbers for pushes in the 3 most recent Turning Stone events (XX, XIX, and XVIII), we have the following.

Out of 622 games in 43 streamed matches, 78 games (12 1/2%) involved a push out, and the results were as follows:

• Breaker pushed and won the game -- 26
• Breaker pushed and lost the game -- 22
• Non-breaker pushed and won the game -- 20
• Non-breaker pushed and lost the game -- 10

So, overall, the person who pushed won 46 of the 78 games (59%) and lost 32 (41%). Both breakers and non-breakers who pushed won the majority of their pushes. This contradicts conventional wisdom.

Of the 78 pushes, 32 were returned (passed back to the pusher to shoot), and the pusher won 21 of those 32 games (66%).

Thanx for this info....but the results don't really surprise me.
The pusher has more control over the situation...he CHOOSES the
battle ground....he can push to his strength and his opponent's weakness.

In a competitive situation, I prefer to bet on the guy asking the questions
rather than the guy who has to come up with the answers.

pt..<...played a lot of 'roll out'

 

[Bob Jewett]

I certainly agree, Bob, and I often mention the "small number" caveat. But I still enjoy looking at the results.

I like to see the numbers too.

If you get a dozen tournaments in which the "breaker wins the rack" percentages are all around 50%, it really starts to tell you that the break is not a plum advantage. This seems especially the case because in lop-sided matches the breaker wins percentages are way up, assuming winner breaks (as was the case at TS).