dissenting.
I agree with my colleagues that the district court properly dismissed Count One of the complaint, in which Santiago alleged that Correctional Officers Rednour and Butler failed to protect him from Harris, a fellow inmate at the Menard Correctional Center. I also agree that the district court properly dismissed Count Two, which alleged that Sgt. Suemnicht and “C.O. John Doe” used excessive force in breaking up the fight between Harris and Santiago in the prison cafeteria. I disagree, however, with the decision to reinstate Count Four. That claim — alleging that Warden Walls failed to protect Santiago from an attack by his cellmate Castro— was also properly dismissed for reasons I will explain in a moment.
More significantly, I cannot agree that the district court abused its discretion by declining to recruit pro bono counsel for Santiago during discovery. In Pruitt v. Mote, 503 F.3d 647 (7th Cir.2007) (en banc), we held that 28 U.S.C. § 1915(e)(1) places no thumb on the scale either for or against recruiting volunteer counsel for an indigent litigant, either in general or in any particular category or type of case. Id. at 654. We also held that a district court’s decision not to recruit counsel is entitled to substantial deference on appeal. Id. Unlike my colleagues, I see no abuse of discretion in the way in which the magistrate judge handled Santiago’s various requests for recruited pro bono counsel during the course of this case. Accordingly, I respectfully dissent.
First, I disagree with the majority’s conclusion to reinstate Count Four, which alleged that Warden Walls failed to protect Santiago from assault by Castro. As applicable here, this claim has the following two elements: (1) that Santiago was incarcerated under conditions posing a substantial risk of serious harm; and (2) that Warden Walls was deliberately indifferent to that risk, that is, that he personally knew of the risk and deliberately disregarded it. Farmer v. Brennan, 511 U.S. 825, 834, 114 S.Ct. 1970, 128 L.Ed.2d 811 (1994); Pinkston v. Madry, 440 F.3d 879, 892 (7th Cir.2006). Santiago attached a copy of the grievance relevant to this claim as an exhibit to his complaint. Based on the contents of that grievance, he pleaded himself out of court.
The grievance is dated June 2, 2002— four days before Santiago’s fight with Castro — and it was filed directly with the warden’s office. In this grievance Santiago notably did not complain that he was at risk of serious harm from Castro, nor did he complain that Castro has a history of assaulting cellmates. Instead, he lodged a more general objection to his current cellmate — whom he did not name — and said that this inmate should have been placed on his “enemies list.” He also claimed that unnamed prison employees were retaliating against him by “placing me in cells with inmates they knew I wouldn’t be able to live in the same cell with with [sic] the same intentions of provoking me into a physical confrontation and keeping me in the segregation unit.” He contended further that these prison employees “continue to either provoke me into altercations of confrontation or place me into physical confrontations with other inmates.” Finally, he said: “I am currently being housed with another inmate, dispite [sic] my numerous request [sic] for single cell.”
Read generously, this grievance at most establishes that Warden Walls was aware that Santiago wanted a single cell to avoid being provoked into confrontations with cellmates he couldn’t get along with and that his current cellmate should have been on his “enemies list” for unstated reasons. It also establishes that the warden was aware that Santiago accused unnamed prison employees of placing him in sitúa*768tions that would provoke him into altercations with other inmates and claimed these employees were motivated by a desire to keep him confined in the segregation unit. Importantly, the grievance does not say Santiago feared being attacked by his cellmate. Instead, Santiago complained that he might be “provoked” into assaulting his cellmate or another inmate. Based on the contents of this grievance, Santiago cannot prevail. As a matter of law, Warden Walls cannot have been deliberately indifferent to a risk that Santiago would be attacked by his cellmate when the grievance in question alerted the warden to an entirely different sort of risk — a risk that Santiago himself might attack another inmate.
There is an additional reason not to reinstate this claim. Although Santiago alleged in his complaint that Castro attacked him, at trial he told quite a different story — one more consistent with the contents of the grievance. He testified that he and Castro were having a heated argument about flushing the toilet in their cell, that the argument escalated, and that he — Santiago—threw the first punch. A full-blown fight ensued and he got the worst of it, sustaining various injuries. In my judgment, on remand the warden would be entitled to summary judgment based on Santiago’s trial testimony. Santiago cannot recover on his claim that the warden was deliberately indifferent to a risk of harm from Castro when Santiago himself started the fight. I would affirm the dismissal of Count Four.
This brings me to the majority’s reversal of the district court’s decision not to recruit pro bono counsel until the time of trial. In Pruitt this court sat en banc to clarify the legal standards that guide the district court’s exercise of discretion when confronted with a request for recruitment of pro bono counsel under 28 U.S.C. § 1915(e)(1). We held that when an indigent plaintiff requests counsel under § 1915(e)(1), the district court must make the following inquiries: “(1) has the indigent plaintiff made a reasonable attempt to obtain counsel or been effectively precluded from doing so; and if so, (2) given the difficulty of the case, does the plaintiff appear competent to litigate it himself?” Pruitt, 503 F.3d at 654. We further held that “[although [§ 1915(e)(1) ] ‘legitimizes’ the court’s request for a pro bono lawyer, its language suggests no congressional preference for recruitment of counsel in any particular circumstance or category of case.” Id. (quoting Johnson v. Doughty, 433 F.3d 1001, 1006 (7th Cir.2006)).
We also addressed the separate question of the appellate standard of review, emphasizing that appellate review — for abuse of discretion only — is limited and highly deferential. “As with any discretionary determination, the question on appellate review is not whether we would have recruited a volunteer lawyer in the circumstances, but whether the district court applied the correct legal standard and reached a reasonable decision on facts supported by the record.” Id. at 658.
The majority notes the district-court and appellate standards from Pruitt but essentially undertakes a wholly independent analysis of Santiago’s request for pro bono counsel, requiring recruitment of counsel during the discovery phase of this case for basically two reasons: (1) Santiago’s claims require proof that prison officials were deliberately indifferent to a risk to his safety; and (2) Santiago had been transferred to another prison by the time he filed his lawsuit. The first of these factors is present in many prisoner cases; as such, the majority’s holding could be understood to suggest a general rule requiring pro bono counsel in deliberate-indifference cases. The second factor is also not uncommon and is not significant *769enough to warrant reversal of the magistrate judge’s discretionary decision.
Our decision in Pruitt specifically addressed the limited nature of the appellate role in reviewing a district court’s decision not to recruit pro bono counsel: “[W]e do not undertake our own analysis of the degree of case difficulty as against the plaintiffs competence to litigate it himself; that is the district court’s inquiry, not ours.” Id. We emphasized the district court’s superior position to evaluate requests for pro bono counsel and reiterated that the reviewing court’s inquiry asks only whether the court’s decision was reasonable: “ We ask not whether [the judge] was right, but whether he was reasonable.’ ” Id. at 659 (quoting Farmer v. Haas, 990 F.2d 319, 322 (7th Cir.1993) (alteration in original)). Properly focused, then, appellate review of decisions declining to recruit counsel under § 1915(e)(1) will result in reversal only in the rare case where the district court’s decision is based on an error of law or clearly erroneous factfinding, or is arbitrary or irrational. Id. at 658.
The magistrate judge’s decision here suffers from none of these infirmities. Though he decided the matter before Pruitt was issued, the judge touched all the bases. First, he specifically addressed the nature of Santiago’s claims — excessive force, failure to protect, insufficient medical care, and retaliation — and concluded they were not overly complex in the circumstances of this case. The judge then evaluated Santiago’s capabilities and the resources available to him in prison, as Pruitt requires. The judge noted that Santiago is literate, understood the nature of his claims, and had demonstrated the ability to “comprehend and follow rules of procedure.” The judge further noted that Santiago had “personal knowledge of the relevant facts” and reasonable access to documents, postal services, a library and legal materials, and photocopying services. The judge reserved judgment about whether Santiago would be competent to try his own case but concluded that he was competent to handle the “pretrial phase of this litigation.” However strongly my colleagues may disagree with this decision, nothing in the judge’s reasoning reflects an abuse of discretion.
As I have noted, the majority has focused on essentially two factors: (1) the supposed complexity of proving deliberate indifference; and (2) the fact that Santiago had been moved to a different prison by the time he filed his case. Taken individually, neither of these factors justifies reversal under our deferential standard of review; nor are they enough considered together to call into question the magistrate judge’s exercise of discretion.
Santiago’s case included claims of excessive force, failure to protect, delayed or insufficient medical care, and retaliation, all stemming from two fights in the prison. I do not agree that the state-of-mind elements in these claims are either inherently complex or uniquely difficult in the circumstances of this case.1 Each defendant’s state of mind is inferred primarily from the circumstances surrounding the assaults in question and the grievances Santiago filed alerting prison officials to his complaints about Harris and Castro. As the magistrate judge properly noted, Santiago had personal knowledge of these facts and circumstances and did not re*770quire assistance of counsel to discover them.
The majority concludes that Santiago was incapable of conducting discovery on his own — in particular, that he was incapable of preparing interrogatories and taking depositions “so that he could determine whether any of the defendants had the requisite knowledge to give rise to an Eighth Amendment violation and to discover the identity of Dr. John Doe.” Maj. op. at 764. The identity of Dr. John Doe was obtainable by simple interrogatory or document request; nothing in the record suggests Santiago was incapable of submitting such a straightforward question to the defendants himself. And I am hesitant to conclude from our appellate-court vantage point that this case required depositions to probe state-of-mind issues. It bears repeating that it is the district court’s prerogative — not ours — to determine whether a pro se litigant is generally capable of navigating pretrial discovery under the particular circumstances of the case. Although my colleagues disagree with the magistrate judge’s assessment of Santiago’s competence to handle the pretrial preparation of his case, the judge’s decision was hardly unreasonable — and that is the appropriate question on appellate review. See Jackson v. Kotter, 541 F.3d 688, 700 (7th Cir.2008) (applying Pruitt and deferring to the district court’s decision declining to recruit counsel in a case involving similar allegations of excessive force, deliberate indifference, and inadequate medical care).
It is true the magistrate judge did not consider the fact that during the pendency of his case, Santiago was incarcerated at a prison other than Menard, where the events at issue took place. It is not clear why this should make any difference, let alone a difference significant enough to displace the magistrate judge’s decision. The majority does not explain why a prisoner’s incarceration at a different prison makes discovery inherently more difficult. With the exception of depositions, discovery is conducted largely through written requests; depositions are initiated upon written notice and, to the extent they are necessary, obviously need not be conducted at the place where the events at issue in the litigation occurred.' The magistrate judge specifically considered Santiago’s capabilities and the resources available to him, noting he was literate and had access to documents, postal services, a law library, and photocopying services to assist him in preparing his discovery requests. In my view, the judge’s failure to separately address Santiago’s incarceration at a different prison is not significant enough to warrant reversal under our deferential standard of review.
The majority also takes issue with the magistrate judge’s language in his order denying Santiago’s fourth motion for counsel, suggesting it “impermissibly prevented Mr. Santiago from making later requests that would have been reviewable in this court.” Maj. op. at 764. With respect, I find this criticism of the judge particularly unwarranted. We held in Pruitt that the district court has no duty to monitor whether an indigent litigant is competently litigating his claims throughout the litigation and therefore no obligation to revisit an earlier denial of pro bono counsel. Pruitt, 503 F.3d at 658. We acknowledged, however, that the court has the discretion to proceed incrementally on this question if it chooses, id., and that’s exactly what the judge did here. The judge held that Santiago was “competent to represent himself throughout the pretrial phase of this litigation” but that his skills “may not suffice at trial,” and on this basis denied Santiago’s request for pro bono counsel “without prejudice to his right to seek representation for purposes of trial.” *771The judge later recruited pro bono counsel to represent Santiago at trial.
The majority faults the judge for using the language I have quoted, saying it was too “definitive” and may have inhibited further requests for counsel during discovery. Maj. op. at 764. This is highly doubtful. Prison inmates generally are not shy about filing motions and are not often deterred by concerns about trying the court’s patience. More importantly, the majority has effectively suggested that the district court must proceed incrementally on the question of recruitment of counsel, always keeping the matter open for further review at any point along the way. We specifically held in Pruitt that there is no such rule. 503 F.3d at 656-58.
Finally, a few words about the majority’s prejudice analysis. The majority notes that the claim against Dr. John Doe was dismissed prior to trial because Santiago could not identify him and suggests this is evidence of prejudice. As I have already explained, there was nothing that prevented Santiago from discovering the doctor’s identity. All it took was a simple interrogatory question or document request; assistance of counsel was not required for that. The majority also concludes that Santiago’s inability to locate other witnesses to the assaults and “gather pertinent evidence such as the surveillance tapes” hampered his ability to present his case. Maj. op. at 766. But it’s important to focus on the crux of the claims that actually went to trial. Four counts survived screening and proceeded to trial: Count Three, against Correctional Officer Keys for allegedly yanking on Santiago’s handcuffs while escorting him from the infirmary to segregation after the fight with Harris; Count Five, against Warden Walls for allegedly failing to protect Santiago by housing him in the same cell block as Harris; Count Six, against Correctional Officers Jines and Cox for allegedly failing to provide medical treatment after the Castro assault; and Count Seven, against Walls and Keys for retaliation. Notably, none of these counts involved a dispute between the parties about the basic facts of either the fight between Santiago and Harris or the fight between Santiago and Castro. Rather, the dispute centered on what happened before and after the fights, and the witnesses to these events were Santiago and the defendant prison officials. Accordingly, identifying additional witnesses would not have made a difference.
As for the “surveillance tapes,” there is nothing in the record to suggest that anything relevant to the surviving counts was captured on surveillance videotape. Santiago believes there was a surveillance camera on the route from the prison cafeteria (where the fight with Harris occurred) to the infirmary and also across from a holding cell outside the segregation unit. The record does not substantiate this. It is not, in any event, enough to establish prejudice, which requires “a reasonable likelihood that the presence of counsel would have made a difference in the outcome of the litigation.” Pruitt, 503 F.3d at 659. Speculation that surveillance videotape might exist (seven years after the fact) and might contain relevant evidence is simply too slim a reed to support a finding of prejudice.
In the end, I return to a concluding point we made in Pruitt:
The principles reiterated here are intended to ensure that requests for pro bono counsel are resolved according to a consistent framework calibrated to the nature of the discretionary judgment called for by § 1915(e)(1). They are not meant to move the exercise of discretion toward recruitment of counsel more often than not, or more often than is now the case; we repeat that the inquiry is *772individualized to the plaintiff and the case before the court.
Id. at 661 (emphasis added). For the reasons I have explained, the majority’s decision in this case may have the effect of suggesting to our district judges that they had better “move the exercise of discretion toward recruitment of counsel more often than not” and “more often than is now the ease.” Id. This will come at a cost to the bench and bar alike and was manifestly not what we intended in Pruitt. For all the foregoing reasons, I would affirm the judgment of the district court.
. The cases the majority cites — Swofford v. Mandrell, 969 F.2d 547 (7th Cir. 1992); Merritt v. Faulkner, 697 F.2d 761 (7th Cir. 1983)— predate Pruitt and must be read in light of the principles explained in our en banc decision. To the extent these cases suggest deliberate-indifference claims are inherently too complex for an indigent pro se litigant, they have been superseded by Pruitt.