Somewhere in Bob Martin's book HOW WE DO IT there is bound to be an answer & a full reference.

https://www.researchgate.net/profile/Robert_Martin5

https://www.researchgate.net/profile/Robert_Martin5

The cycle day of interest in relation to ovulation is calculated by assuming that ovulation occurs 14 (Beckmann et al., 1998; Knaus, 1929; Ogino, 1930) or 15 (Dixon, Schlesselman, Ory, & Blye, 1980) days prior to the first day of the next menstruation. Trussell et al. (1998) suggest that subtracting 13 days from a woman’s usual cycle length (i.e., 14 days before the next menses begins) provides a less biased estimate of the timing of ovulation than Dixon et al.’s (1980) method of assuming ovulation occurs 15 days prior to next menstruation, so it depends on who you side with.  

the best reference days for luteal phase inn a womam will dependent on the regulatory of her cycle,if cycles are every 30days then mostly it is 14 days prior to the next menstrual period but in case of irregular cycles it will have to be calculated with combining ulltrasonography to find when a dominant follicle does form in an ovulatory cycle and accordingly after ovulation roughly it willbe 14 days from that day but in anovulatory cycle one cannot fix the length but will have to give withdrawl bleed and stimulate thae lady by ovulation inducion and the day dominant follicle forms to calculate from the day of follicle rupture approximately 14 days