MEMORANDUM OPINION

Memorandum Opinion by Justice CARTER.

For his part in stealing beer from a garage, and after enhancement, a jury assessed Allan Cunningham twenty-five years' incarceration for burglary of a habitation. On appeal, Cunningham argues that we should reverse his conviction because: (1) the trial court failed to charge the jury that Bobby Lemon, the codefendant who actually entered the building and took the beer, was an accomplice witness as a matter of law, (2) the court erred by denying his motion for an instructed verdict based on the failure to connect him with the crime by nonaccomplice evidence, and (3) the evidence was legally and factually insufficient to support the verdict. He also contends he received ineffective assistance of counsel because counsel did not request an accomplice-witness instruction, failed to object to the admission of evidence about juvenile adjudications, and failed to adequately impeach a witness. We affirm the judgment of the trial court.

Background Facts

Cunningham was riding around with Lemon and Christopher Ray (in Ray's sister's car) when they backed into the driveway of John Lightfoot's house. As Lemon testified, they were all pretty well "lit." Lemon testified that he went inside the garage to steal something, and ended up taking beer out of a cooler. When Lemon came back out with stolen beer in hand, Cunningham was arguing with one of Lightfoot's neighbors (Larry Buster) who came to make inquiry. A fight ensued, and both Cunningham and Lemon struck Buster before they drove away, leaving him unconscious in the driveway. Ray testified that he stayed in the car, that Lemon said it was a friend's house, and that he had no idea a crime would occur.

Buster testified that, when he saw the car backed into Lightfoot's driveway, he went over to see what was going on, and he saw Cunningham in the car. Buster testified that Cunningham yelled at him and got out of the car, and about then he saw someone else come out of the garage with beer in his hand, and that both Cunningham and Lemon hit and kicked him until he blacked out.

Lightfoot testified he was in the house and did not hear the fight. Buster, bloody from the battle, banged on his door and told him what had happened. Lightfoot checked his garage and realized some beer was missing from an ice chest.

Lemon testified that, after this occurred, while they were driving around, he saw a pickup truck with tools in the back. He got out and stole the truck and drove away — with Cunningham following in Ray's sister's car. Lemon stopped near a business, unloaded the items from the truck into the car, left the truck there, Lemon got in the car, and Cunningham drove the automobile away.

Officer Steven Hill testified that he worked the assault and burglary. Officers stopped the car and arrested Lemon and Cunningham. When officers inventoried the car, they found three cans of beer of the type in Lightfoot's cooler, and also found tools, a C.D. case, a rifle, a cell phone, a video camera, and other items that were identified as property stolen from the pickup truck.

Failure to Instruct the Jury on the Accomplice-Witness Rule

A. Lemon was an accomplice as a matter of law — egregious harm standard.

Lemon was an accomplice as a matter of law in this case. He was indicted and convicted for the same offenses and was in prison at the time of this trial. "A conviction cannot be had upon the testimony of an accomplice unless corroborated by other evidence tending to connect the defendant with the offense committed; and the corroboration is not sufficient if it merely shows the commission of the offense." Tex. Code Crim. Proc. Ann. art. 38.14 (Vernon 2005). Cunningham did not request an instruction to the jury on the requirement of corroborating an accomplice witness' testimony, and no such instruction was given. In such event, when error occurs in failing to properly instruct the jury on the accomplice-witness rule, our review of the charge is under the Almanza standard. Almanza v. State, 686 S.W.2d 157 (Tex.Crim.App. 1984). The appropriate harm analysis depends on whether the defendant preserved error by bringing the improper omission to the trial court's attention. When the error is preserved, we must reverse if "some harm" is shown. But when, as in this case, the defendant has not preserved error, he must show egregious harm. Under either instance, we must review the entirety of the record in reaching our determination. Id. at 171. The difference in harm standards impacts how strong the nonaccomplice evidence must be for the error in omitting an accomplice-witness instruction to be considered harmless. Herron v. State, 86 S.W.3d 621, 632 (Tex.Crim.App. 2002).

The Texas Court of Criminal Appeals has held that, in applying the egregious harm standard, the omission of an accomplice-witness instruction is generally harmless unless the corroborating (nonaccomplice) evidence is "so unconvincing in fact as to render the State's overall case for conviction clearly and significantly less persuasive." Id.; Saunders v. State, 817 S.W.2d 688, 692 (Tex.Crim.App. 1991).

In Saunders, the court found harm under that standard because the corroborating nonaccomplice evidence was weak and was contradicted by other evidence. The alleged crime was arson and the corroborating evidence involved financial circumstances that seemed somewhat suspicious at first glance but were given persuasive innocent explanations. As a result, the corroborating evidence, even if believed, did not have a very strong tendency to connect the defendant to the crime. Further, much of the evidence in the defendant's favor was uncontradicted, and the court observed that "[r]ational jurors may not utterly disregard undisputed evidence without a sensible basis for thinking it unreliable any more than they may simply assume a critical part of the proof without evidence having an inclination to confirm it." Saunders, 817 S.W.2d at 693.

In determining the strength of a particular item of nonaccomplice evidence, we examine (1) its reliability or believability, and (2) the strength of its tendency to connect the defendant to the crime. Herron, 86 S.W.3d at 632. The test for weighing the sufficiency of corroborating evidence is to eliminate from consideration the accomplice's testimony and then examine the remaining evidence to determine if there is evidence that tends to connect the defendant with the commission of the offense. Munoz v. State, 853 S.W.2d 558, 559 (Tex.Crim.App. 1993); Reed v. State, 744 S.W.2d 112, 125 (Tex.Crim.App. 1988); Hall v. State, 161 S.W.3d 142, 149 (Tex.App.-Texarkana 2005, pet. ref'd). The nonaccomplice testimony does not have to directly link the accused to the crime, it alone need not establish guilt beyond a reasonable doubt, nor must the corroborating evidence prove all the elements of the alleged offense. Gill v. State, 873 S.W.2d 45, 48 (Tex.Crim.App. 1994). The accused's presence at the scene of the crime is, by itself, insufficient to corroborate an accomplice's testimony. However, "evidence that an accused was in the company of the accomplice close to the time of the offense, coupled with other suspicious circumstances, may tend to connect the accused to the offense." Id. at 49.

B. The nonaccomplice testimony connects Cunningham to the burglary.

In this case, there was nonaccomplice evidence from the neighbor, Buster, that Cunningham was in the car and that Cunningham attacked him for no reason other than his inquisitiveness. This response could reasonably have been interpreted by the jury as an effort to aid and assist in the burglary and to prevent being seen with items stolen from the house. The jury could reasonably have concluded that this violent reaction evidenced Cunningham's intent to help Lemon complete the theft and escape.

Ray, the other person in the car, testified that Cunningham drove the vehicle to the home that was robbed; Ray saw Cunningham kicking and stomping Buster, and Ray told Cunningham and Lemon that they had done a bad thing by attacking and beating the man — after which they beat Ray up and left him (although still driving Ray's sister's car).

There is no evidence in this case that Cunningham personally stole the beer. The jury necessarily convicted him (as provided by the charge) as a party criminally responsible for the offense committed by Lemon. That type of responsibility exists if "with intent to promote or assist the commission of the offense, he solicits, encourages, directs, aids, or attempts to aid the other person to commit the offense." Tex. Pen. Code Ann. § 7.02(a)(2) (Vernon 2003).

In this case, there is nonaccomplice testimony that Cunningham was driving the vehicle in which Lemon and Ray were riding when they went to Lightfoot's house. Lemon entered the garage and came out with beer. As Lemon was engaged in the burglary, a neighbor came to investigate. Cunningham got out of the car and attacked, without provocation, the neighbor who came to inquire about their actions. Cunningham assisted in attacking Buster, and Cunningham then drove them away, leaving him on the ground. The evidence shows that Lemon stole the beer from the house, that Lemon and Cunningham were arrested together, and that the fruits of the theft were found in the car. Intent may be directly proven, or it may be inferred from circumstantial evidence such as acts, words, and the conduct of the appellant. Guevara v. State, 152 S.W.3d 45, 50 (Tex.Crim.App. 2004).

Although it is clear that an accomplice-witness instruction should have been provided to the jury, we do not believe that egregious harm resulted from its absence. The nonaccomplice evidence was of a nature that a jury could reasonably believe was reliable, persuasive, and connected Cunningham to the burglary. The nonaccomplice evidence, as summarized above, is not "so unconvincing in fact as to render the State's overall case for conviction clearly and significantly less persuasive." Herron, 86 S.W.3d at 632. Points of error one and two are overruled.

Legal and Factual Sufficiency

Cunningham next contends that the evidence is legally and factually insufficient to support the verdict. In our review of the legal sufficiency of the evidence, we employ the standards set forth in Jackson v. Virginia, 443 U.S. 307, 319 (1979). This calls on the reviewing court to view the relevant evidence in the light most favorable to the verdict and determine whether any rational trier of fact could have found the essential elements of the crime beyond a reasonable doubt. Johnson v. State, 23 S.W.3d 1, 7 (Tex.Crim.App. 2000). In our review, we must evaluate all the evidence in the record, both direct and circumstantial, whether admissible or inadmissible. Dewberry v. State, 4 S.W.3d 735, 740 (Tex.Crim.App. 1999).

In a factual sufficiency review, the appellate court views all the evidence in a neutral light and determines whether the evidence supporting the verdict is too weak to support the finding of guilt beyond a reasonable doubt or if evidence contrary to the verdict is strong enough that the beyond-a-reasonable-doubt standard could not have been met. Threadgill v. State, 146 S.W.3d 654, 664 (Tex.Crim.App. 2004) (citing Zuniga v. State, 144 S.W.3d 477, 486 (Tex.Crim.App. 2004)).

We have summarized the evidence above. There is evidence that the crime occurred and that Cunningham took violent action against a witness to the crime. The accomplice witness testified that he did not remember telling Cunningham before the burglary that he intended to steal from the house. ("I'm not saying they didn't know but I'm saying that I was drinking and I'm saying I don't remember . . . discussing this with somebody.") However, the district attorney's investigator testified that Lemon had stated to him that Lemon and Cunningham had discussed "stealing something" at the Lightfoot house before the burglary. The fruits of the crime were in the automobile driven by Cunningham at the time of his arrest. The evidence is both legally and factually sufficient to support the verdict. The contentions of error are overruled.

Ineffective Assistance of Counsel

Finally, Cunningham contends that he received constitutionally ineffective assistance of counsel at trial because of the failure to obtain an instruction on accomplice-witness testimony, because counsel did not object when his juvenile adjudications were admitted into evidence, and because of the failure to impeach a witness with a prior statement that the codefendant was the aggressor.

When ineffective assistance is raised on direct appeal, appellate counsel and the court must proceed on a trial record not developed for the object of litigating or preserving the claim, and thus, it is often incomplete or inadequate for this purpose. Freeman v. State, 125 S.W.3d 505, 506 (Tex.Crim.App. 2003).

The standard for testing claims of ineffective assistance of counsel is set out in Strickland v. Washington, 466 U.S. 668 (1984). To prevail on a claim of ineffective assistance, an appellant must, by a preponderance of the evidence, prove: (1) trial counsel's performance fell below an objective standard of reasonableness, and (2) counsel's deficient representation prejudiced the appellant's defense. Id.; Bone v. State, 77 S.W.3d 828, 833 (Tex.Crim.App. 2002). To meet this burden, an appellant must show that the attorney's representation fell below the standard of prevailing professional norms and that there is a reasonable probability that, but for the attorney's deficiency, the result of the trial would have been different. Tong v. State, 25 S.W.3d 707, 712 (Tex.Crim.App. 2000). In other words, the appellant must prove counsel's representation so undermined the proper functioning of the adversarial process that the trial cannot be relied on as having produced a just result. Strickland, 466 U.S. at 686. If, however, "there is at least the possibility that the conduct could have been legitimate trial strategy," then we must "defer to counsel's decisions and deny relief on an ineffective assistance claim on direct appeal." Murphy v. State, 112 S.W.3d 592, 601 (Tex.Crim.App. 2003).

Our review of counsel's representation is highly deferential, and we indulge a strong presumption that counsel's conduct falls within a wide range of reasonable representation. Strickland, 466 U.S. at 689; Tong, 25 S.W.3d at 712. An appellate court should not second-guess through hindsight the strategy of counsel at trial, nor will the fact that another attorney might have pursued a different course support a finding of ineffectiveness. Blott v. State, 588 S.W.2d 588, 592 (Tex.Crim.App. 1979). That another attorney, including appellant's counsel on appeal, might have pursued a different course of action does not necessarily indicate ineffective assistance. Harner v. State, 997 S.W.2d 695, 704 (Tex.App.-Texarkana 1999, no pet.). Any allegation of ineffectiveness must be firmly founded in the record, and the record must affirmatively demonstrate the alleged ineffectiveness. Thompson v. State, 9 S.W.3d 808, 813 (Tex.Crim.App. 1999).

A. Failure to request an instruction on the accomplice-witness rule.

Counsel did not request a jury instruction that the accomplice-witness testimony must be corroborated. We have recently addressed this issue in our opinion in Hall v. State, 161 S.W.3d 142 (Tex.App.-Texarkana 2005, pet. ref'd). As we pointed out there, the failure of counsel to request an accomplice-witness instruction when facts warrant such an instruction may constitute deficient performance. Id. at 152; Henson v. State, 915 S.W.2d 186, 197 (Tex.App.-Corpus Christi 1996, no pet.) (citing Ex parte Zepeda, 819 S.W.2d 874, 877 (Tex.Crim.App. 1991)). It is uncontested that Lemon was an accomplice as a matter of law, and thus, the instruction, which could only favor the defense, was mandatory on request.

Unlike the situation in Hall, however, the evidence corroborating the accomplice's testimony in this case is not weak. Even without the accomplice-witness testimony, there is evidence that strongly supports a conclusion that Cunningham was involved in the burglary. Buster walked up on this burglary as it was occurring. After routine inquiries, Cunningham attacked him in what the jury could reasonably believe was an effort to conceal the burglary. Without repeating the evidence previously elicited, we find that the corroborating evidence was both compelling and persuasive. Even though counsel's failure to request the instruction was deficient, based on the analysis previously conducted on the evidence elicited at the trial, as well as all of the surrounding proceedings, we are not convinced from this record that the failure created a reasonable probability that the result of the trial would have been different in the absence of the deficient conduct. Strickland, 466 U.S. at 687.

B. Failure to object to juvenile adjudications.

Cunningham also argues that his counsel was deficient because he did not object to the admission of evidence about his juvenile adjudications at the punishment stage of the trial. The complained-of evidence consists of a list of juvenile adjudications, many from another state, and some of which were felony level, while others were misdemeanor. We also recognize the State mentioned Cunningham's juvenile history several times during argument as a basis for a sentence of lengthy incarceration.

Article 37.07, Section 3(a)(1), states that "evidence may be offered by the state and the defendant of an adjudication of delinquency based on a violation by the defendant of a penal law of the grade of: (A) a felony; or (B) a misdemeanor punishable by confinement in jail." Tex. Code Crim. Proc. Ann. art. 37.07, § 3(a)(1) (Vernon Supp. 2006). However, such evidence is limited. Article 37.07, Section 3(i), states as follows:

Evidence of an adjudication for conduct that is a violation of a penal law of the grade of misdemeanor punishable by confinement in jail is admissible only if the conduct upon which the adjudication is based occurred on or after January 1, 1996.

Tex. Code Crim. Proc. Ann. art. 37.07, § 3(i) (Vernon Supp. 2006). See generally Wallace v. State, 135 S.W.3d 114, 120 (Tex.App.-Tyler 2004, no pet.).

The net effect of these provisions is that juvenile adjudications of delinquency which occurred before January 1, 1996, are not admissible, as prior adjudications of delinquency, unless the adjudication was for a felony-grade offense. An objection to juvenile adjudications of the grade of misdemeanor, all of which occurred before January 1, 1996, should have prevented their admission as adjudications. However, evidence of the conduct underlying such an adjudication may be admissible as extraneous crimes or prior bad acts if properly proven. Tex. Code Crim. Proc. Ann. art. 37.07, § 3(a)(1); Rodriguez v. State, 975 S.W.2d 667, 687 (Tex.App.-Texarkana 1998, pet. ref'd). In this case, no evidence of such bad acts was presented during the State's case, other than the attempt to introduce these offenses as adjudications of delinquency. The document introduced, Exhibit 29, was only a summary in a presentence report, not certified copies of judgments. No other evidence was presented during the State's case concerning the conduct of Cunningham that led to the adjudications. Exhibit 29 shows Cunningham's convictions as follows:

November 28, 1993 motor-vehicle larceny — (unauthorized use of a motor vehicle — or burglary of a motor vehicle) (F)

November 28, 1993 Injured property (M) November 28, 1993 Larceny (M) February 1, 1994 Motor vehicle larceny (F) February 1, 1994 Breaking and entering a motor vehicle — (burglary of a motor vehicle) (three counts) (F) February 24, 1994 Motor-vehicle larceny (F) December 7, 1994 Assault (M) October 26, 1995 Community threat November 29, 1995 Disorderly conduct November 29, 1995 Credit card abuse November 29, 1995 Assault of a police officer (M) November 28, 1995 Injured property (M)

The State acknowledged at trial that "some of these are misdemeanors and some are felonies."

The State is allowed in the punishment hearing to prove evidence of extraneous offenses and even bad acts that are deemed relevant and shown beyond a reasonable doubt by the evidence to have been committed by the defendant regardless of whether he or she was charged or convicted of such crimes or acts. Tex. Code Crim. Proc. Ann. art. 37.07, § 3(a)(1). One method of providing such proof beyond a reasonable doubt is to question the defendant when he or she voluntarily submits to cross-examination. Cunningham testified and acknowledged his juvenile record. On direct examination, Cunningham was asked by his counsel, "So, you do not deny the juvenile record that has been presented to the jury today? A. No, sir." He was then asked if he denied the misdemeanor record and answered that some were dismissed. On cross-examination, the State questioned Cunningham about the specific events shown on his juvenile record.

It is conceivable that counsel's actions were strategically planned. Since Cunningham was requesting the jury grant him community supervision in lieu of confinement, counsel may have thought it necessary to present Cunningham's testimony, knowing such action subjected him to cross-examination, which would allow the State to prove his juvenile record as constituting admissible bad acts. See Rodriguez, 975 S.W.2d at 687; Strasser v. State, 81 S.W.3d 468, 470 (Tex.App.-Eastland 2002, no pet.). Counsel could have believed that the only effective method of convincing a jury to grant community supervision to Cunningham was to allow him to testify; let him explain that his prior juvenile record occurred while he was young and without guidance; and urge that he had reformed. To accomplish this goal, counsel may have calculated that it was more persuasive to admit that Cunningham had a poor record as a young person, but explain he had never been convicted as an adult of a felony offense and was eligible for community supervision. Such an attempt could have been to attempt to personalize Cunningham to the jury as a young person who had no parental guidance as a child, but who now admitted his prior mistakes and accepted societal rules. The evidence presented by Cunningham appears to substantiate this as the defense strategy. Cunningham testified that he had never been convicted as an adult of a felony; he had no parental assistance as a child; he took full responsibility for his bad choices; before the charges were filed he had enlisted in the United States Army; he was attempting to change his life by enrolling in a computer school; and that, if he received community supervision, he would still have his job. In closing argument, counsel argued that Cunningham had turned his life around and asked the jury to grant community supervision. We have no record from which to evaluate the rationale of counsel in choosing the course of action that he did, and it is beyond our purview to second guess counsel's strategy even if the choice of strategy was not successful. As the Texas Court of Criminal Appeals has held many times, "In most instances, the record on direct appeal is inadequate to develop an ineffective assistance claim." Ex parte Torres, 943 S.W.2d 469, 475 (Tex.Crim.App. 1997). If there is at least the possibility that the conduct could have been legitimate trial strategy," then we must "defer to counsel's decisions and deny relief on an ineffective assistance claim on direct appeal." Murphy v. State, 112 S.W.3d 592, 601 (Tex.Crim.App. 2003).

From this record, Cunningham has not shown this Court that trial counsel's actions were below an objective standard of reasonableness.

C. Failure to impeach witness.

Cunningham also contends that counsel was ineffective because he did not impeach Buster with his prior written statement to police. Buster testified that both Lemon and Cunningham were hitting and kicking him and that he could not tell who threw the first blow. On cross-examination, he testified he was not sure who had attacked him first. Appellate counsel argues that Buster's statement had indicated that Lemon had initially attacked him and that trial counsel was inadequate because, although referring to the statement, he did not seek to introduce the statement by the use of extrinsic evidence. See Tex. R. Evid. 613(a).

This constitutes a question of trial tactics. We do not have a copy of the statement, thus, we do not know its contents, but from counsel's description it appears to show nothing more than which of the two hit Buster first. This is not a particularly important issue, and counsel may well have felt it was of insufficient importance to spend additional time and energy — in effect — showing the jury, again, that his client hit and kicked a man in his mid-70's who was trying to protect an even older man living next door. The contention of error is overruled.

We affirm the judgment.