def _calculate_normalized_play_pct(positon_players, normalized_data):
for player in positon_players.values():
rbiz_successes = player["routinebizfield"] / player["routinebiz"] * normalized_data["routinebiz"] if player["routinebiz"] > 0 else 0
lbiz_successes = player["likelybizfield"] / player["likelybiz"] * normalized_data["likelybiz"] if player["likelybiz"] > 0 else 0
ebiz_successes = player["evenbizfield"] / player["evenbiz"] * normalized_data["evenbiz"] if player["evenbiz"] > 0 else 0
ulbiz_successes = player["unlikelybizfield"] / player["unlikelybiz"] * normalized_data["unlikelybiz"] if player["unlikelybiz"] > 0 else 0
vulbiz_successes = player["veryunlikelybizfield"] / player["veryunlikelybiz"] * normalized_data["veryunlikelybiz"] if player["veryunlikelybiz"] > 0 else 0
player["adj_play_pct"] = (rbiz_successes + lbiz_successes + ebiz_successes + ulbiz_successes + vulbiz_successes) / normalized_data["totbiz"] if normalized_data["totbiz"] > 0 else 0
player["adj_fieldedballs"] = (player["routinebizfield"] + player["likelybizfield"] + player["evenbizfield"] + player["unlikelybizfield"]
+ player["veryunlikelybizfield"]) / ip_to_num(player["fieldingip"]) * 9 * 162 if player["fieldingip"] > 0 else 0
player["adj_totbiz"] = player["totalbiz"] / ip_to_num(player["fieldingip"]) * 9 * 162 if player["fieldingip"] > 0 else 0
player["adj_zr"] = player["zr"] / ip_to_num(player["fieldingip"]) * 9 * 162 if player["fieldingip"] > 0 else 0
def calculate_league_stats(ovr_data, vl_data, vr_data, splits):
tot_batr = 0.0
tot_bsr = 0.0
tot_zr = 0.0
tot_pas = 0
for player_data in ovr_data.values():
tot_batr += player_data["batr"]
tot_bsr += player_data["baserunningruns"]
tot_zr += player_data["zr"]
tot_pas += player_data["pa"]
league_adj = -1 * (tot_batr + tot_bsr + tot_zr) / tot_pas
replacement_player_adj = 228000 / 185553
batter_stats = {
"avg":
lambda pd: pd["hits"] / pd["ab"] if pd["ab"] > 0 else 0,
"obp":
lambda pd:
(pd["hits"] + pd["walks"] + pd["timeshitbypitch"]) / pd["pa"]
if pd["pa"] > 0 else 0,
"ops":
lambda pd: (pd["hits"] + pd["doubles"] + pd["triples"] * 2 + pd[
"homeruns"] * 3) / pd["ab"] +
(pd["hits"] + pd["walks"] + pd["timeshitbypitch"]) / pd["pa"]
if pd["ab"] > 0 else 0,
"bsr_600_pa":
lambda pd: ovr_data[str(pd["t_CID"])][
"baserunningruns"] * 600 / ovr_data[str(pd["t_CID"])]["pa"]
if pd["pa"] > 0 else 0,
"zr_600_pa":
lambda pd: pd["zr"] * 600 / pd["pa"] if pd["pa"] > 0 else 0,
"batr_600_pa":
lambda pd: pd["batr"] * 600 / pd["pa"] if pd["pa"] > 0 else 0,
# last number is replacement adjustment, so it's not based on average
"war_600_pa":
lambda pd: (pd["batr"] + pd["zr"] + pd["baserunningruns"] + league_adj
+ replacement_player_adj) * 600 / (pd["pa"] * 10)
if pd["pa"] > 0 else 0
}
pitcher_stats = {
"sp_k_per_9":
lambda pd: pd["sp_k"] * 9 / ip_to_num(pd["sp_ip"])
if pd["sp_ip"] > 0 else 0,
"sp_bb_with_hbp_per_9":
lambda pd:
(pd["sp_bb"] + pd["sp_playershitbypitch"]) * 9 / ip_to_num(pd["sp_ip"])
if pd["sp_ip"] > 0 else 0,
"sp_hr_per_9":
lambda pd: pd["sp_hra"] * 9 / ip_to_num(pd["sp_ip"])
if pd["sp_ip"] > 0 else 0,
"rp_k_per_9":
lambda pd: pd["rp_k"] * 9 / ip_to_num(pd["rp_ip"])
if pd["rp_ip"] > 0 else 0,
"rp_bb_with_hbp_per_9":
lambda pd:
(pd["rp_bb"] + pd["rp_playershitbypitch"]) * 9 / ip_to_num(pd["rp_ip"])
if pd["rp_ip"] > 0 else 0,
"rp_hr_per_9":
lambda pd: pd["rp_hra"] * 9 / ip_to_num(pd["rp_ip"])
if pd["rp_ip"] > 0 else 0,
"ip_per_gamesstarted":
lambda pd: pd["sp_ip"] / pd["gamesstarted"]
if pd["gamesstarted"] > 0 and pd["gamesstarted"] == pd["games"] else 0,
"ip_per_gamesrelieved":
lambda pd: pd["rp_ip"] / pd["games"]
if pd["games"] > 0 and pd["gamesstarted"] == 0 else 0,
"sp_war_per_220_ip":
lambda pd: pd["sp_war"] * 220 / ip_to_num(pd["sp_ip"])
if pd["sp_ip"] > 0 else -50,
"rp_war_per_100_ip":
lambda pd: pd["rp_war"] * 100 / ip_to_num(pd["rp_ip"])
if pd["rp_ip"] > 0 else 0,
}
_calculate_stats_per_data(vl_data, batter_stats, pitcher_stats, "vL")
_calculate_stats_per_data(vr_data, batter_stats, pitcher_stats, "vR")
_calculate_stats_per_data(ovr_data, batter_stats, pitcher_stats, "ovr")
progress_bar = ProgressBar(len(ovr_data.keys()), "Adding extra ovr data")
for card in ovr_data.values():
cid = str(card["t_CID"])
lefty_stats = vl_data[cid] if cid in vl_data else None
righty_stats = vr_data[cid] if cid in vr_data else None
if lefty_stats == None or righty_stats == None or lefty_stats[
"pa"] < 10 or righty_stats["pa"] < 10:
# Set to some default low values
card["batr_600_pa_ft"] = -10
card["batr_600_pa_vr"] = -10
card["batr_600_pa_vl"] = -10
card["war_600_pa_ft"] = -10
card["war_600_pa_vr"] = -10
card["war_600_pa_vl"] = -10
progress_bar.increment()
continue
position = "catcher" if card["position"] == "C" else "fielder"
ft_vr_split = splits["FT"]["vR%"][position]
vr_vr_split = splits["vR"]["vR%"][position]
vl_vr_split = splits["vL"]["vR%"][position]
card["batr_600_pa_ft"] = ft_vr_split * righty_stats["batr_600_pa"] + (
1 - ft_vr_split * lefty_stats["batr_600_pa"])
card["batr_600_pa_vr"] = vr_vr_split * righty_stats["batr_600_pa"] + (
1 - vr_vr_split * lefty_stats["batr_600_pa"])
card["batr_600_pa_vl"] = vl_vr_split * righty_stats["batr_600_pa"] + (
1 - vl_vr_split * lefty_stats["batr_600_pa"])
card["war_600_pa_ft"] = ft_vr_split * righty_stats["war_600_pa"] + (
1 - ft_vr_split * lefty_stats["war_600_pa"])
card["war_600_pa_vr"] = vr_vr_split * righty_stats["war_600_pa"] + (
1 - vr_vr_split * lefty_stats["war_600_pa"])
card["war_600_pa_vl"] = vl_vr_split * righty_stats["war_600_pa"] + (
1 - vl_vr_split * lefty_stats["war_600_pa"])
progress_bar.increment()
progress_bar.finish()
print()
def merge_player_data(old_info, new_info):
if ip_to_num(add_ip(old_info["fieldingip"], new_info["fieldingip"])) > 0:
old_info["cera"] = (
old_info["cera"] * ip_to_num(old_info["fieldingip"]) +
new_info["cera"] * ip_to_num(new_info["fieldingip"])) / ip_to_num(
add_ip(old_info["fieldingip"], new_info["fieldingip"]))
if (new_info["sp_bf"] + old_info["sp_bf"] > 0):
old_info["sp_fip"] = (new_info["sp_bf"] * new_info["sp_fip"] +
old_info["sp_bf"] * old_info["sp_fip"]) / (
new_info["sp_bf"] + old_info["sp_bf"])
if (new_info["rp_bf"] + old_info["rp_bf"] > 0):
old_info["rp_fip"] = (new_info["rp_bf"] * new_info["rp_fip"] +
old_info["rp_bf"] * old_info["rp_fip"]) / (
new_info["rp_bf"] + old_info["rp_bf"])
old_info["leverage"] = (new_info["rp_bf"] * new_info["leverage"] +
old_info["rp_bf"] * old_info["leverage"]) / (
new_info["rp_bf"] + old_info["rp_bf"])
if (new_info["pa"] + old_info["pa"] > 0):
old_info["woba"] = (new_info["pa"] * new_info["woba"] +
old_info["pa"] * old_info["woba"]) / (
new_info["pa"] + old_info["pa"])
if ip_to_num(add_ip(old_info["sp_ip"], new_info["sp_ip"])) > 0:
old_info["sp_era"] = (
old_info["sp_era"] * ip_to_num(old_info["sp_ip"]) +
new_info["sp_era"] * ip_to_num(new_info["sp_ip"])) / ip_to_num(
add_ip(old_info["sp_ip"], new_info["sp_ip"]))
old_info["sp_whip"] = (
old_info["sp_whip"] * ip_to_num(old_info["sp_ip"]) +
new_info["sp_whip"] * ip_to_num(new_info["sp_ip"])) / ip_to_num(
add_ip(old_info["sp_ip"], new_info["sp_ip"]))
if ip_to_num(add_ip(old_info["rp_ip"], new_info["rp_ip"])) > 0:
old_info["rp_era"] = (
old_info["rp_era"] * ip_to_num(old_info["rp_ip"]) +
new_info["rp_era"] * ip_to_num(new_info["rp_ip"])) / ip_to_num(
add_ip(old_info["rp_ip"], new_info["rp_ip"]))
old_info["rp_whip"] = (
old_info["rp_whip"] * ip_to_num(old_info["rp_ip"]) +
new_info["rp_whip"] * ip_to_num(new_info["rp_ip"])) / ip_to_num(
add_ip(old_info["rp_ip"], new_info["rp_ip"]))
old_info["sp_hra"] += new_info["sp_hra"]
old_info["sp_hitsallowed"] += new_info["sp_hitsallowed"]
old_info["sp_k"] += new_info["sp_k"]
old_info["sp_bb"] += new_info["sp_bb"]
old_info["rp_hra"] += new_info["rp_hra"]
old_info["rp_hitsallowed"] += new_info["rp_hitsallowed"]
old_info["rp_k"] += new_info["rp_k"]
old_info["rp_bb"] += new_info["rp_bb"]
old_info["pa"] += new_info["pa"]
old_info["gs"] += new_info["gs"]
old_info["g"] += new_info["g"]
old_info["fielding_gs"] += new_info["fielding_gs"]
old_info["runsscored"] += new_info["runsscored"]
old_info["rbi"] += new_info["rbi"]
old_info["intentionallywalked"] += new_info["intentionallywalked"]
old_info["sacflies"] += new_info["sacflies"]
old_info["sacbunts"] += new_info["sacbunts"]
old_info["gidp"] += new_info["gidp"]
old_info["stolenbases"] += new_info["stolenbases"]
old_info["caughtstealing"] += new_info["caughtstealing"]
old_info["ab"] += new_info["ab"]
old_info["sp_ip"] = add_ip(old_info["sp_ip"], new_info["sp_ip"])
old_info["sp_war"] += new_info["sp_war"]
old_info["rp_ip"] = add_ip(old_info["rp_ip"], new_info["rp_ip"])
old_info["rp_war"] += new_info["rp_war"]
old_info["fieldingip"] = add_ip(old_info["fieldingip"],
new_info["fieldingip"])
old_info["bwar"] += new_info["bwar"]
old_info["wraa"] += new_info["wraa"]
old_info["batr"] += new_info["batr"]
old_info["sp_bf"] += new_info["sp_bf"]
old_info["rp_bf"] += new_info["rp_bf"]
old_info["games"] += new_info["games"]
old_info["gamesstarted"] += new_info["gamesstarted"]
old_info["wsb"] += new_info["wsb"]
old_info["ubr"] += new_info["ubr"]
old_info["baserunningruns"] += new_info["baserunningruns"]
old_info["bsrunchances"] += new_info["bsrunchances"]
old_info["stealchances"] += new_info["stealchances"]
old_info["zr"] += new_info["zr"]
old_info["strikeouts"] += new_info["strikeouts"]
old_info["walks"] += new_info["walks"]
old_info["homeruns"] += new_info["homeruns"]
old_info["hits"] += new_info["hits"]
old_info["doubles"] += new_info["doubles"]
old_info["triples"] += new_info["triples"]
old_info["routinebiz"] += new_info["routinebiz"]
old_info["routinebizfield"] += new_info["routinebizfield"]
old_info["likelybiz"] += new_info["likelybiz"]
old_info["likelybizfield"] += new_info["likelybizfield"]
old_info["evenbiz"] += new_info["evenbiz"]
old_info["evenbizfield"] += new_info["evenbizfield"]
old_info["unlikelybiz"] += new_info["unlikelybiz"]
old_info["unlikelybizfield"] += new_info["unlikelybizfield"]
old_info["veryunlikelybiz"] += new_info["veryunlikelybiz"]
old_info["veryunlikelybizfield"] += new_info["veryunlikelybizfield"]
old_info["impossiblebiz"] += new_info["impossiblebiz"]
old_info["totalbiz"] += new_info["totalbiz"]
old_info["totalchances"] += new_info["totalchances"]
old_info["assists"] += new_info["assists"]
old_info["putouts"] += new_info["putouts"]
old_info["stolenbaseattempts"] += new_info["stolenbaseattempts"]
old_info["runnersthrownout"] += new_info["runnersthrownout"]
old_info["errors"] += new_info["errors"]
old_info["doubleplays"] += new_info["doubleplays"]
old_info["sp_playershitbypitch"] += new_info["sp_playershitbypitch"]
old_info["rp_playershitbypitch"] += new_info["rp_playershitbypitch"]
old_info["timeshitbypitch"] += new_info["timeshitbypitch"]
old_info["totalplaysfielded"] += new_info["totalplaysfielded"]
old_info["sp_wildpitches"] += new_info["sp_wildpitches"]
old_info["sp_pitch_stealchances"] += new_info["sp_pitch_stealchances"]
old_info["sp_pitch_basesstolen"] += new_info["sp_pitch_basesstolen"]
old_info["sp_pitch_stealattempts"] += new_info["sp_pitch_stealattempts"]
old_info["sp_pitch_basesstolen_wild_pitch"] += new_info[
"sp_pitch_basesstolen_wild_pitch"]
old_info["sp_earned_runs"] += new_info["sp_earned_runs"]
old_info["sp_pitch_non_hr_hits"] += new_info["sp_pitch_non_hr_hits"]
old_info["sp_bb_hp"] += new_info["sp_bb_hp"]
old_info["sp_pitcherdoubleplays"] += new_info["sp_pitcherdoubleplays"]
old_info["rp_wildpitches"] += new_info["rp_wildpitches"]
old_info["rp_pitch_stealchances"] += new_info["rp_pitch_stealchances"]
old_info["rp_pitch_basesstolen"] += new_info["rp_pitch_basesstolen"]
old_info["rp_pitch_stealattempts"] += new_info["rp_pitch_stealattempts"]
old_info["rp_pitch_basesstolen_wild_pitch"] += new_info[
"rp_pitch_basesstolen_wild_pitch"]
old_info["rp_earned_runs"] += new_info["rp_earned_runs"]
old_info["rp_pitch_non_hr_hits"] += new_info["rp_pitch_non_hr_hits"]
old_info["rp_bb_hp"] += new_info["rp_bb_hp"]
old_info["rp_pitcherdoubleplays"] += new_info["rp_pitcherdoubleplays"]
if old_info["totalbiz"] > 0:
old_info[
"playpct"] = old_info["totalplaysfielded"] / old_info["totalbiz"]
old_info[
"routinefieldpct"] = old_info["routinebiz"] / old_info["totalbiz"]
old_info["routinebizpct"] = old_info["routinebizfield"] / old_info[
"routinebiz"] if old_info["routinebiz"] > 0 else 0
old_info["likelybizpct"] = old_info["likelybizfield"] / old_info[
"likelybiz"] if old_info["likelybiz"] > 0 else 0
old_info["evenbizpct"] = old_info["evenbizfield"] / old_info[
"evenbiz"] if old_info["evenbiz"] > 0 else 0
old_info["unlikelybizpct"] = old_info["unlikelybizfield"] / old_info[
"unlikelybiz"] if old_info["unlikelybiz"] > 0 else 0
old_info["veryunlikelybizpct"] = old_info[
"veryunlikelybizfield"] / old_info["veryunlikelybiz"] if old_info[
"veryunlikelybiz"] > 0 else 0
return old_info
def regress_rto(player_data):
avg_cabi = 0.0
avg_carm = 0.0
avg_ip = 0.0
attempts_by_arm = {}
rto_by_arm = {}
for player_info in player_data.values():
if player_info["runnersthrownout"] < 1 or player_info[
"cera"] < 2.0 or player_info["carm"] < 15 or player_info[
"fieldingip"] < 10.0:
continue
carm = player_info["carm"]
cabi = player_info["cabi"]
avg_carm = (avg_carm * avg_ip +
carm * ip_to_num(player_info["fieldingip"])) / (
avg_ip + ip_to_num(player_info["fieldingip"]))
avg_cabi = (avg_cabi * avg_ip +
cabi * ip_to_num(player_info["fieldingip"])) / (
avg_ip + ip_to_num(player_info["fieldingip"]))
avg_ip += ip_to_num(player_info["fieldingip"])
if carm in attempts_by_arm:
attempts_by_arm[carm] = (player_info["stolenbaseattempts"] +
attempts_by_arm[carm][0],
add_ip(player_info["fieldingip"],
attempts_by_arm[carm][1]))
rto_by_arm[carm] = (player_info["runnersthrownout"] +
rto_by_arm[carm][0],
player_info["stolenbaseattempts"] +
rto_by_arm[carm][1])
else:
attempts_by_arm[carm] = (player_info["stolenbaseattempts"],
player_info["fieldingip"])
rto_by_arm[carm] = (player_info["runnersthrownout"],
player_info["stolenbaseattempts"])
X = []
y = []
for carm in attempts_by_arm:
if attempts_by_arm[carm][1] == 0:
continue
X.append(carm)
y.append(attempts_by_arm[carm][0] / attempts_by_arm[carm][1])
# initial attempt - we still need to throw out outliers
att_model = sm.OLS(y, sm.add_constant(X))
att_results = att_model.fit()
att_influence = att_results.get_influence()
att_cooks_distance = att_influence.cooks_distance[0]
att_cutoff = 4.0 / ((len(X) - 2) if len(X) > 3 else 1)
att_old_len_X = len(X)
X = []
y = []
i = 0
for carm in attempts_by_arm:
if att_cooks_distance[i] < att_cutoff or att_old_len_X < 4:
X.append(carm)
y.append(attempts_by_arm[carm][0] / attempts_by_arm[carm][1])
i += 1
# Real prediction
att_model = sm.OLS(y, sm.add_constant(X))
att_results = att_model.fit()
X = []
y = []
for carm in attempts_by_arm:
if attempts_by_arm[carm][1] == 0:
continue
X.append(carm)
y.append(rto_by_arm[carm][0] / rto_by_arm[carm][1])
# initial attempt - we still need to throw out outliers
rto_model = sm.OLS(y, sm.add_constant(X))
rto_results = rto_model.fit()
rto_influence = rto_results.get_influence()
rto_cooks_distance = rto_influence.cooks_distance[0]
rto_cutoff = 4.0 / ((len(X) - 2) if len(X) > 3 else 1)
rto_old_len_X = len(X)
X = []
y = []
i = 0
for carm in attempts_by_arm:
if rto_cooks_distance[i] < rto_cutoff or rto_old_len_X < 4:
X.append(carm)
y.append(rto_by_arm[carm][0] / rto_by_arm[carm][1])
i += 1
# Real prediction
rto_model = sm.OLS(y, sm.add_constant(X))
rto_results = rto_model.fit()
# attempts by arm, r2, rto/att by arm, r2
return (att_results.params, att_results.rsquared, rto_results.params,
rto_results.rsquared, avg_carm, avg_cabi)
def calculate_defensive_stats(tourney_data, cards):
positions = ["C", "1B", "2B", "3B", "SS", "LF", "CF", "RF"]
position_players_data = {
"C": {},
"1B": {},
"2B": {},
"3B": {},
"SS": {},
"LF": {},
"CF": {},
"RF": {}
}
position_biz_breakdown = {
"C": { "totbiz": 0, "routinebiz": 0, "likelybiz": 0, "evenbiz": 0, "unlikelybiz": 0, "veryunlikelybiz": 0, "routinebizfield": 0, "likelybizfield": 0, "evenbizfield": 0, "unlikelybizfield": 0, "veryunlikelybizfield": 0 },
"1B": { "totbiz": 0, "routinebiz": 0, "likelybiz": 0, "evenbiz": 0, "unlikelybiz": 0, "veryunlikelybiz": 0, "routinebizfield": 0, "likelybizfield": 0, "evenbizfield": 0, "unlikelybizfield": 0, "veryunlikelybizfield": 0 },
"2B": { "totbiz": 0, "routinebiz": 0, "likelybiz": 0, "evenbiz": 0, "unlikelybiz": 0, "veryunlikelybiz": 0, "routinebizfield": 0, "likelybizfield": 0, "evenbizfield": 0, "unlikelybizfield": 0, "veryunlikelybizfield": 0 },
"3B": { "totbiz": 0, "routinebiz": 0, "likelybiz": 0, "evenbiz": 0, "unlikelybiz": 0, "veryunlikelybiz": 0, "routinebizfield": 0, "likelybizfield": 0, "evenbizfield": 0, "unlikelybizfield": 0, "veryunlikelybizfield": 0 },
"SS": { "totbiz": 0, "routinebiz": 0, "likelybiz": 0, "evenbiz": 0, "unlikelybiz": 0, "veryunlikelybiz": 0, "routinebizfield": 0, "likelybizfield": 0, "evenbizfield": 0, "unlikelybizfield": 0, "veryunlikelybizfield": 0 },
"LF": { "totbiz": 0, "routinebiz": 0, "likelybiz": 0, "evenbiz": 0, "unlikelybiz": 0, "veryunlikelybiz": 0, "routinebizfield": 0, "likelybizfield": 0, "evenbizfield": 0, "unlikelybizfield": 0, "veryunlikelybizfield": 0 },
"CF": { "totbiz": 0, "routinebiz": 0, "likelybiz": 0, "evenbiz": 0, "unlikelybiz": 0, "veryunlikelybiz": 0, "routinebizfield": 0, "likelybizfield": 0, "evenbizfield": 0, "unlikelybizfield": 0, "veryunlikelybizfield": 0 },
"RF": { "totbiz": 0, "routinebiz": 0, "likelybiz": 0, "evenbiz": 0, "unlikelybiz": 0, "veryunlikelybiz": 0, "routinebizfield": 0, "likelybizfield": 0, "evenbizfield": 0, "unlikelybizfield": 0, "veryunlikelybizfield": 0 }
}
total_fielding_ip_per_pos = {}
# Use these to calculate avg play%
tot_biz_per_pos = {
"C": 0,
"1B": 0,
"2B": 0,
"3B": 0,
"SS": 0,
"LF": 0,
"CF": 0,
"RF": 0
}
tot_fieldedbiz_per_pos = {
"C": 0,
"1B": 0,
"2B": 0,
"3B": 0,
"SS": 0,
"LF": 0,
"CF": 0,
"RF": 0
}
# Used to calculate outputs
regression_data = {
"C": {},
"1B": {},
"2B": {},
"3B": {},
"SS": {},
"LF": {},
"CF": {},
"RF": {}
}
r2_data = {
"C": {},
"1B": {},
"2B": {},
"3B": {},
"SS": {},
"LF": {},
"CF": {},
"RF": {}
}
progress_bar = ProgressBar(len(tourney_data.keys()), "Reading tourney players for fielding data")
for (key, pi) in tourney_data.items():
progress_bar.increment()
player_positions = 0
for position in positions:
if pi[position.lower() + "x"] > 0:
player_positions += 1
if player_positions > 1:
continue
if pi["pos"] not in position_players_data:
continue
tot_biz_per_pos[pi["pos"]] += pi["totalbiz"]
tot_fieldedbiz_per_pos[pi["pos"]] += pi["totalplaysfielded"]
position_biz_breakdown[pi["pos"]]["totbiz"] += pi["totalbiz"]
position_biz_breakdown[pi["pos"]]["routinebiz"] += pi["routinebiz"]
position_biz_breakdown[pi["pos"]]["likelybiz"] += pi["likelybiz"]
position_biz_breakdown[pi["pos"]]["evenbiz"] += pi["evenbiz"]
position_biz_breakdown[pi["pos"]]["unlikelybiz"] += pi["unlikelybiz"]
position_biz_breakdown[pi["pos"]]["veryunlikelybiz"] += pi["veryunlikelybiz"]
position_biz_breakdown[pi["pos"]]["routinebizfield"] += pi["routinebizfield"]
position_biz_breakdown[pi["pos"]]["likelybizfield"] += pi["likelybizfield"]
position_biz_breakdown[pi["pos"]]["evenbizfield"] += pi["evenbizfield"]
position_biz_breakdown[pi["pos"]]["unlikelybizfield"] += pi["unlikelybizfield"]
position_biz_breakdown[pi["pos"]]["veryunlikelybizfield"] += pi["veryunlikelybizfield"]
if pi["pos"] not in total_fielding_ip_per_pos:
total_fielding_ip_per_pos[pi["pos"]] = 0
total_fielding_ip_per_pos[pi["pos"]] = add_ip(pi["fieldingip"], total_fielding_ip_per_pos[pi["pos"]])
position_players_data[pi["pos"]][key] = pi
progress_bar.finish()
for position in position_biz_breakdown:
for key in position_biz_breakdown[position]:
# Per game stats
position_biz_breakdown[position][key] = position_biz_breakdown[position][key] / ip_to_num(total_fielding_ip_per_pos[position]) * 9
position_info = [
("C", [
("adj_play_pct", [
"carm",
"cabi"
]),
("adj_zr", [
"carm",
"cabi"
]),
("adj_totbiz", [
"ifrng"
])
]),
("1B", [
("adj_play_pct", [
"ifrng",
"ifarm",
"iferr",
"tdp",
"height"
]),
("adj_zr", [
"ifrng",
"ifarm",
"iferr",
"tdp",
"height"
]),
("adj_totbiz", [
"ifrng"
])
]),
("2B", [
("adj_play_pct", [
"ifrng",
"ifarm",
"iferr",
"tdp"
]),
("adj_zr", [
"ifrng",
"ifarm",
"iferr",
"tdp"
]),
("adj_totbiz", [
"ifrng"
])
]),
("3B", [
("adj_play_pct", [
"ifrng",
"ifarm",
"iferr",
"tdp"
]),
("adj_zr", [
"ifrng",
"ifarm",
"iferr",
"tdp"
]),
("adj_totbiz", [
"ifrng"
])
]),
("SS", [
("adj_play_pct", [
"ifrng",
"ifarm",
"iferr",
"tdp"
]),
("adj_zr", [
"ifrng",
"ifarm",
"iferr",
"tdp"
]),
("adj_totbiz", [
"ifrng"
])
]),
("LF", [
("adj_play_pct", [
"ofrng",
"ofarm",
"oferr"
]),
("adj_zr", [
"ofrng",
"ofarm",
"oferr"
]),
("adj_totbiz", [
"ofrng"
])
]),
("CF", [
("adj_play_pct", [
"ofrng",
"ofarm",
"oferr"
]),
("adj_zr", [
"ofrng",
"ofarm",
"oferr"
]),
("adj_totbiz", [
"ofrng"
])
]),
("RF", [
("adj_play_pct", [
"ofrng",
"ofarm",
"oferr"
]),
("adj_zr", [
"ofrng",
"ofarm",
"oferr"
]),
("adj_totbiz", [
"ofrng"
])
])
]
progress_bar = ProgressBar(len(position_info) * len(position_info[0][1]), "Calculating fielding regressions")
for position, pos_regs in position_info:
_calculate_normalized_play_pct(position_players_data[position], position_biz_breakdown[position])
for (outcome_stat, pos_ratings) in pos_regs:
reg, r2 = regress_defensive_stats(position_players_data[position], pos_ratings, outcome_stat)
regression_data[position][outcome_stat] = reg
r2_data[position][outcome_stat] = r2
progress_bar.increment()
progress_bar.finish()
for position in r2_data.keys():
for outcome_stat in r2_data[position].keys():
print(position, outcome_stat, "r2:", r2_data[position][outcome_stat])
print()
outs_above_average_per_162 = {}
zr_per_162 = {}
for position in regression_data.keys():
outs_above_average_per_162[position] = _get_oaa_fn(position, regression_data, position_biz_breakdown)
zr_per_162[position] = _get_zr_fn(position, regression_data)
positions = ["C", "1B", "2B", "3B", "SS", "LF", "CF", "RF"]
progress_bar = ProgressBar(len(cards), "Calculate in defensive stats to cards")
for card in cards:
for position in positions:
card[position + "_expected_zr"] = zr_per_162[position](card)
card[position + "_expected_outs_above_avg"] = outs_above_average_per_162[position](card)
progress_bar.increment()
progress_bar.finish()
def _calc_woba_factors(player_data):
X = []
y = []
X_wsb = []
y_wsb = []
total_ip = 0
total_runs = 0
pa = 0
walks = 0
hbp = 0
singles = 0
doubles = 0
triples = 0
homeruns = 0
ibb = 0
progress_bar = ProgressBar(len(player_data.keys()), "Reading wOBA values")
for player in player_data.values():
pa += player["pa"]
walks += player["walks"]
hbp += player["timeshitbypitch"]
singles += player["hits"] - (player["homeruns"] + player["doubles"] +
player["triples"])
doubles += player["doubles"]
triples += player["triples"]
homeruns += player["homeruns"]
ibb += player["intentionallywalked"]
total_ip = add_ip(add_ip(total_ip, player["sp_ip"]), player["rp_ip"])
total_runs += player["runsscored"]
progress_bar.increment()
if player["pa"] < 20:
continue
X.append([
1, player["walks"], player["timeshitbypitch"], player["hits"] -
(player["homeruns"] + player["doubles"] + player["triples"]),
player["doubles"], player["triples"], player["homeruns"]
])
y.append(player["woba"] *
(player["pa"] - player["intentionallywalked"]))
X_wsb.append([
player["stolenbases"], player["caughtstealing"],
((player["hits"] -
(player["homeruns"] + player["doubles"] + player["triples"])) +
player["walks"] + player["timeshitbypitch"] -
player["intentionallywalked"])
])
y_wsb.append(player["wsb"])
progress_bar.finish("\n")
results = sm.OLS(y, X).fit()
wsb_results = sm.OLS(y_wsb, X_wsb).fit()
avg_woba = (results.params[0] + results.params[1] * walks +
results.params[2] * hbp + results.params[3] * singles +
results.params[4] * doubles + results.params[5] * triples +
results.params[6] * homeruns) / (pa - ibb)
return {
"lg_woba": avg_woba,
"woba_intcpt": results.params[0],
"walks_factor": results.params[1],
"hbp_factor": results.params[2],
"singles_factor": results.params[3],
"doubles_factor": results.params[4],
"triples_factor": results.params[5],
"homeruns_factor": results.params[6],
"r_2": results.rsquared,
"runSB": wsb_results.params[0],
"runCS": wsb_results.params[1],
"lgwSB": wsb_results.params[2],
"wsb_r_2": wsb_results.rsquared,
"outs_per_run": (wsb_results.params[1] - 0.075) / 2,
"runs_per_win": 9 * (total_runs / ip_to_num(total_ip)) * 1.5 + 3,
}