def handle_race(self, ex, race):
if race['event'] == TO_BE_PLACED or race['n_runners'] < 3:
return
runners = race['selection']
if np.all(self.hm.get_runs(runners) >= self.min_races):
prices = ex.get_market_prices(race['market_id'])
prices = dict(map(lambda x: (x['selection'], x), prices))
odds = np.array(
map(lambda r: prices[r]['back_prices'][0]['price'], runners))
implied = get_implied_from_odds(odds)
p = self.hm.pwin_trapz(runners)
rel = p / implied - 1.0
t = 0.05
p[rel < -t] = implied[rel < -t] * 0.95
p[rel > t] = implied[rel > t] * 1.05
w = risk.nwin1_l2reg(p, odds, self.risk_aversion)
returns = risk.nwin1_bet_returns(w, odds)
if np.any(returns <= -self.max_expsoure):
logging.warning('Maximum exposure limit of %.2f reached!' %
self.max_expsoure)
logging.warning(
'Ignoring bets w=%s runners=%s with potential returns=%s' %
(w.tolist(), runners, returns.tolist()))
else:
logging.info(
'Betting on market_id=%s: |exposure|=%.2f collateral=%.2f'
% (race['market_id'], np.sum(np.abs(w)), np.min(returns)))
bets = ({
'selection_id': prices[r]['selection_id'],
'amount': w[i],
'data': {
'p': p[i],
'implied': implied[i]
}
} for i, r in enumerate(runners))
ex.place_exchange_bets(race['market_id'], bets)
if 'ranking' in race:
self.hm.fit_race(race)
def handle_race(self, ex, race):
if race['event'] == TO_BE_PLACED or race['n_runners'] < 3:
return
runners = race['selection']
if np.all(self.hm.get_runs(runners) >= self.min_races):
prices = ex.get_market_prices(race['market_id'])
prices = dict(map(lambda x: (x['selection'], x), prices))
odds = np.array(map(lambda r: prices[r]['back_prices'][0]['price'], runners))
implied = get_implied_from_odds(odds)
p = self.hm.pwin_trapz(runners)
rel = p / implied - 1.0
t = 0.05
p[rel < -t] = implied[rel < -t] * 0.95
p[rel > t] = implied[rel > t] * 1.05
w = risk.nwin1_l2reg(p, odds, self.risk_aversion)
returns = risk.nwin1_bet_returns(w, odds)
if np.any(returns <= -self.max_expsoure):
logging.warning('Maximum exposure limit of %.2f reached!' % self.max_expsoure)
logging.warning('Ignoring bets w=%s runners=%s with potential returns=%s'
% (w.tolist(), runners, returns.tolist()))
else:
logging.info('Betting on market_id=%s: |exposure|=%.2f collateral=%.2f' %
(race['market_id'], np.sum(np.abs(w)), np.min(returns)))
bets = ({'selection_id': prices[r]['selection_id'],
'amount': w[i],
'data': {
'p': p[i],
'implied': implied[i]
}
} for i, r in enumerate(runners))
ex.place_exchange_bets(race['market_id'], bets)
if 'ranking' in race:
self.hm.fit_race(race)
def calculate_collateral(group):
return np.min(
risk.nwin1_bet_returns(group.amount.values, group.odds.values))
def calculate_collateral(group):
return np.min(risk.nwin1_bet_returns(group.amount.values, group.odds.values))
