Uechi-Ryu.com

The Gracies would further claim that their's was the most important and practical system to learn, because based on their experience fighting on the streets of Rio de Janeiro as well as in numerous challenge matches, "ninety percent of all fights go to the ground.

The American Society of Law Enforcement Training provides statistical data along with a 1991 study of LAPD incident reports involving a physical confrontations[...]
The LAPD study does not show that “90% of fights go to the ground.” Instead, the LAPD study shows that 95% of altercations took on one of five familiar patterns (with which any street cop will be intimately familiar). It also shows that of that 95%, 62% ended up with both the officer and the suspect grappling on the ground.

Journal of Non-lethal Combatives, Jan 2007
Going to the Ground: Lessons from Law Enforcement
By Chris Leblanc
Copyright © Chris Leblanc 2007. All rights reserved.


Law Enforcement Officers (LEOs) go “hands on” in both armed and unarmed physical confrontations more often than perhaps any other armed professionals. Within the self-defense and martial arts communities, this naturally has led to a great deal of interest in the experiences of officers in physical encounters. And no other information coming from the law enforcement community has received as much attention as an elusive set of statistics that purportedly show that 90% (or more) of physical altercations “go to the ground.”
The responsibility for the popularizing of this statistic is most often laid at the feet of the famous Gracie family, proponents of the art of Brazilian jujitsu, and dismissed as a shameless attempt at marketing themselves and their family fighting system which, not coincidentally, emphasizes fighting on the ground.

Unfortunately, I have yet to see a single source within the martial arts community -- affiliated with the Gracies or otherwise -- that accurately cites the actual study, or that does not either accept the statistics (or repudiate them) almost wholesale.

The statistics provided here are quoted directly from the ASLET (American Society of Law Enforcement Training) pamphlet for their July 1997 Use of Force Training Seminar. [...]In 1991, Sergeant Dossey, an exercise physiologist with the LAPD, completed a comparative study of use of force incidents reported by LAPD for the year 1988. Sergeant Dossey looked at all 5,617 use of force incident narratives written by officers in 1988, and devised a method for codifying the information contained and analyzing it for what they identified as dominant altercation patterns. The study was replicated in 1992 by LAPD’s Training Review committee.

During 1988, there were 316,525 arrests made by LAPD.

5,617 (1.7%) of these arrests required the completion of a use of force report.

2,031 (0.6%) altercations developed from these arrests. “Of the 5,617 reports examined, only 2,031 incidents contained a sufficient level of aggressive resistance by the suspect toward the officer to qualify as an altercation.”

Five scenario patterns accounted for 95% of the altercations: “[...]As for the five patterns, they were:

Subject pulls away from an officer’s attempt to control the subject’s arm. “33.7% Officer grabbed the subject by the arm and the subject pulled his arm away; the most frequent second act was the officer applying a joint lock (32%) and the most frequent final subduing act was the officer taking the subject down to the ground (46%)”

Subject attempts to punch or kick the officer. “25.4% Subject ran at the officer and swung punches and kicks; the most frequent second act was the officer evading the subject and striking him with the baton (26%; a close second was taking the subject to the ground 22%) and the most frequent final subduing act was taking the subject to the ground (35%).”

Subject refuses to assume a searching position. “19.3% Subject refused to assume a searching position as verbally ordered by the officer; the most frequent second act was the officer applying a joint lock (35.5%) and the most frequent final subduing act was taking the subject to the ground (36.5%).”

Subject flees and officer pursues. “10.5% Subject ran from the officer, officer chased the subject; the most frequent second act was the officer taking the subject to the ground (40%) and the most frequent final subduing act was also taking the subject to the ground (39.5%).”

Subject takes a combative posture, but does not attempt to strike the officer. “6.8% Subject assumed a fighting, martial arts, or boxing stance but did not attack the officer; the most frequent second act was the officer striking the subject with the baton (38%) and this was also the most frequent final act (41%).”

Once the officer was down:

The subject continued to assault the officer once the officer was down (64%)

The subject fled (31%)

The subject waited for the officer to get back up to continue the fight (5%)

Of the ground fights, suspects generally continued with grappling and pinning techniques (77%), or used punches, kicks, and strikes (66%). However, in 21% of the cases, the subjects attempted to disarm the officer, with 5% being successful. As a side note, the FBI states that of 594 law enforcement officers killed between 1992 and 2001, 46 were killed with their own weapon.

On the ground, the officers tended to use weapons other than firearms:

Pepper spray (OC) was used 29% of the time

Impact weapons (sticks, batons, flashlights, handcuffs, etc.) were used 26% of the time

Hands, feet, holds, etc., were used 24% of the time

How many fights REALLY go to ground?

More relevant question might be: How often does it go to the ground when one combatant wants it to?

Answer:

Most of the time..
[...]Bottom line: The ground exists as an environmental variable.. If you find yourself on the ground what are your options, how experienced are you? If the answer is not experienced at all then it makes sense to get some... Otherwise one is vulnerable and severely limited when faced with a common real-world scenario--earth in the face..

That's supported by the LAPD report as well...most of the time the police end up on the ground because the assailant pushes or pulls them down. I found the stats on what the subject does once the policeman is down and what techniques are used interesting.