We sold more Early Adopter Dakota R/T licenses than we expected, but we would like to provide you one last chance to pick up Dakota R/T, the intraday version of Dakota, at an affordable early price.

Effective midnight Monday, July 24th, the Dakota R/T Early Adopter pricing will be gone... forever...  Order ›

Due to popular request, the end-of-day Dakota now shows trading bot swarms in three dimensions as they adapt to changing market conditions.  The trading bots that are currently profitable are green, un-profitable red and those with zero equity performance are black.  This graphic really doesn't do it justice, but you can select and view up to 3 adaptable parameters and see a 3D plot of the trading bots' positions relative to their parameter values and a color coding of their recent equity performance.  You can drag to rotate, zoom, pan and reset back to the default positioning.  This is useful to see regions of trading system parameter values that are currently profitable and those that are not.  Some users use this information to tighten or expand the limits of their parameter adaptation.

Optimization vs. Adaptation
I get a lot of questions about system optimization vs. adaptation.  The distinction isn't obvious.  Simply stated, trading system optimization seeks the singular best set of trading system parameters using history and uses those values going forward.  On the other hand, adaptation gradually moves the parameters towards improved equity performance bar by bar.  Optimization is like "jump on best" and adaptation is "move towards best".  You might be saying, "If optimization finds the optimal, why wouldn't you use that?".  I say because optimal isn't necessarily optimal, and here's a few reasons why:

• Equity performance is very noisy.  What is optimal right now, perching your money on one of many transient peaks, is very likely not going to be optimal in the near term.  Dakota's adaptation of multiple trading "bots" (systems) tends to smooth through spurious equity peaks.
• Reversion to the mean may hurt "optimal" systems.  There is a theory held by some people that a given trading system has a base performance on a ticker.  Deviance from that base performance in the near term may cause a reversion back to that base performance.  That is, system parameters that happen to yield recent historical high performance may actually be poor performers in the coming bars as they decline to revert to their mean performance.  According to this theory, jumping to the "optimal" all the time may actually result in continual losses.  I believe that Dakota's adaptation of multiple trading "bots" (systems) tends to capitalize on all the signals while enhancing on-going equity performance on a broader, global basis.
• Jumping around breaks up signal continuity.  If you were to frequently re-optimize your systems, your parameter values would tend to jump around in a very discontinuous fashion and the continuity (smoothness) of your signals is broken, as at one instant, you should be long... no, now short,... no, now out... no, long.  With Dakota's adaptation, signals get a chance to smoothly transition and trading systems can do their work without being "jerked around".

Ponder those points for a minute and then think about what may be happening in your "traditional" trading tools when "optimizing" over years of history or even bar-to-bar.  Dakota users are already discovering "why best isn't", sharing their ideas in our private discussion forums.

Dakota does adaptation by default, although you could do optimization if you like, with a bit of coding.  Whether you want to is still open to debate.