Hi JPG 1. You wrote, "My perception of the NC is : connecting a high performance state to the context of the “problem” state" There is no point to limiting the application to a problem state - any context in which you desire new options. There is much to be said for selecting contexts in which you are already successful to radically increase the set of choices available for you there, for example. Problem states (or contexts, since this is the focus of the work) are the traditional domain of psychotherapy. NLP casts its nets much wider. 2. You proposed, "So, my guess is that the MAIN motivation behind step 2(“localize physically this hallucinated context …This is also the opportunity for the coach to calibrate your present state response to the context in question”) is to get an anchor of the context RATHER than of the problem state. I mean some triggers that the client see, hear, smell when he enters the context problem." Correct on all points - I would add the comment that the intention behind physically locating the context selected in a specific physical location (spatial anchor) is that most frequently, the player, having entered a high performance state, is startled by the news (step 4) that there could be such a thing as a problem state - since the key point in step 4 is to connect the stimuli (auditory and visual) that define the context selected with the high performance state, it is at least prudent to locate the context selected physically to ensure the connection gets made. 3. You wrote, "...much emphasises on the problem state (sometimes an intense negative feeling) the two anchors might collapse (like in the collapsing anchors pattern but that NC is more in the spirit of a Swish)." In my opinion, it is closer formally to the collasping of anchors except there is no collapsing at step 4 only the connection to context to ensure an automatic unconscious re-activation when the player next enters that context (or anyone of the contexts related by FA similarities). There are few aspects that overlap with SWISH. I have never witnessed a "problem" state that was intense enough to challenge the high performance state although logically it is possible (see Mitch's postings and my responses for a possible exception). 4. You wrote, "- One to investigate (not too much) about the problem state (find the triggers, possibly calibrate the response of the client) - another where  to anchor the context problem" I appreciate your intention but personally my calibration that the player has selected the context is more than adequate as an investigation of the problem state - by capturing the context (all the auditory and visual stimuli that define the context in which the problematic response occurs, you are guarenteed to pick up the triggers (including internal triggers that are elicited by those stimuli). Perhaps you intend something I missed. I don't get the point of determining as you said, "... where  to anchor the context problem"? You will have to elaborate for me to offer a response. John