The opinions expressed in these forums do not represent those of C2, and any discussion of profit/loss
is not indicative of future performance or success.
There is a substantial risk of loss in trading. You should therefore carefully consider
whether such trading is suitable for you in light of your financial condition. You should read,
understand, and consider the Risk Disclosure Statement that is provided by your broker
before you consider trading. Most people who trade lose money.
AH you are correct, thanks for pointing that out. We have a client that thinks in monthly/quarterly percent alpha and that was the gear my brain was in glancing at those percents.
Actually, as I think about that, I THINK that C2 calculates Monthly Return % using the Raw P/L divided by Suggested Minimum Capital, NOT the current account balance.
So, in this instance, the November +15.8% gain on $60K (Suggested Minimum Capital) IS making up all of September’s losses of -15.6% on $60K (Suggested Minimum Capital).
If not, how does C2 calculate it, so everyone (including me) will know how it IS done rather than just how it’s NOT done?
As C2 calculates “Suggested Minimum Capital” on its own, I assumed it used this value to calculate % monthly returns, since everyone’s actual account balance will vary. It seems like you’d need one constant (like Suggested Minimum Capital) to uniformly calculate %'s against.
It doesn’t impact my trading strategies, but since you seem to know definitively how it’s done, please share.
Well, hard to argue with that infallible observation and prediction.
That was sort of my point when I said “Tempting to turn everything off and take the 16% for November…” the rest of the month is a big unknown (as it always is), tempting to take the +16% and take the rest of the month off…
I need to think of a way to teach our models that traders fear crossing psychological support/resistance prices like the plague. I watch almost every tick of price movement of S&P futures throughout the day, and the level of fear and capitulation observed at committing to breaking through 4400 right now (and in early October) is almost comical. Like a kid who crosses a line at the playground they’re not supposed to, and then darts back to safety before anyone notices.
The problem is our strats don’t care about the psychology of 4400. They see everything else line up in a solid run up to 4400 and say “cool, next stop, 4408-4420”. Except they don’t know about the fear of traders being on the wrong side of the market being rejected at 4400 (because, you know, it’s a nice round number), and everyone running back to the safety of <4400 as soon as the price crosses over…
Suggestions? Something as simple as “don’t go long if the price is xx9x and don’t go short if the price is xx0x” seems rudimentary, but who knows, it could be that simple…
We’re currently long on the run up this morning with our first profit target @ 4408, the market touches 4407.75 and it’s like it realized “oh crap, I’m over 4400! Retreat!”
So, these initial results merit more testing. For one of our long-only strats, I added simple logic that says “Don’t go long if current “hundreds” price is >88 (i.e. 4289-4299, 4389-4399…)”. The screenshots below are the “as-is” strat backtest for 2023 (trading 1 ES contract) followed by the test with the above new logic:
Yup, you guessed it, Jerome Powell just started speaking…
I need an automated “Fed-speak” signal that says “dump any existing long positions right before ANY Fed president speaks anywhere publicly, and (maybe) buy them back 10-15 minutes later…”
On the plus side, today is the last scheduled market manipulation, err, I mean Fed speech, for the rest of the year. That doesn’t mean one of them won’t say something within microphone range, but the only two remaining Fed calendar events left for 2023 is the FOMC Minutes release on 11/21 and the press conference on 12/13 following their two-day meeting in December…
A sideways day can be murder for us. The market ends flat while we notch four straight losing trades.
With no stop losses we end the day in the green. A love/hate relationship with stop losses as they kill us most of the time, until they save us of course…
As you can see, with one exception (trade time, not trade price), all of the backtested trade times are within a minute of live trade times, and all trade prices are within 1-2 ticks (due to slippage) of each other. We do factor slippage into our backtest and tuning P/L calculations.
We do have one strategy we are watching, as it is the only one not 100% lockstep with backtesting. Not sure why yet, but that one strat lines up about 90% of the time, but 1-2 out of every ten or so trades is out of sync with backtesting AND real-time paper trading. I thought it might be that exit prices were getting momentarily touched but our limit orders not getting filled by the market, but we have our strats set to market order out of a position if a limit order exit price is touched but our limit order not filled within 10 seconds.
Sideways just hurts us when the market bounces intraday just enough to trigger entries/stop losses before reversing, over and over…
That’s all intra-day trend-following strategies suffer in this type of market. We have to be patient waiting for the day suitable for our strategies to come.
This is a note I posted to our non-C2 clients this morning. Is everyone here who leverages automated algorithmic trading as fanatical about live trade signals matching backtest trade signals? If not, why not?
A little “behind the scenes” of some of the things we’re constantly auditing/monitoring.
An ABSOLUTE NON-NEGOTIABLE 100% requirement of ANY strategy we trade live is it HAS to line up with backtesting, PERFECTLY. EVERY trade, +/- 1 candle (minute), has to match with backtesting. Fill prices can vary due to slippage and other factors (which we factor into our backtesting and tuning P/L calculations), but the trade signals need to be a 100% MATCH.
ANY unexplained difference between live trade signals and backtest trade signals for a strategy, no matter how profitable the strategy is live, means that it is unreliable, and we are running something live other than what we have backtested/tuned.
Yesterday afternoon (1:39PM to be exact) I temporarily reduced the trade quantity for one of our strategies across the board, even though it’s profitable, as we have now observed several instances of this particular strategy having different trade signals than in backtesting.
When watching your account trade balance in real time, you will observe your open position quantity changing, but not always just buying/selling 100% of the open position. This is because, on the backend, we have multiple instances of multiple strategies all trading in your account, with different trade quantities allowed, entry/exit criteria, etc.
In the screenshot below for one of our live accounts, you’ll see four different strategies trading in this time period (since yesterday afternoon). The one with a trade quantity of 1 traded twice since yesterday afternoon, but you’ll notice in the first trade the time for the exit signal varied from the backtest by 18 minutes, also affecting ROI. That is a RED FLAG.
If it happens just once in a long blue moon, that can happen due to our market data provider sending subsequent corrected candles (which backtest then uses, but could be different than what we first received and ran live against). But if it happens multiple times (as it has now with this strategy in the last couple of weeks), EVEN IF PROFITABLE, it gets reduced or turned off completely until (and if) we can figure out and resolve the trade signal discrepancies.
Seems very wise to constantly check how well the live is matching backtests of the same period. I take it that both of these strategies have matched backtests closely since inception here as they did for this week. If so that has got to be of some comfort during drawdowns.
it may repeat September later. Or not.
Hi-ho, hi-ho, a stopping-out we go… 
