Frankly, my experience, along with several other Alta owners here in SoCal, is that the Alta MXR is already competitive with the current crop of 450s power and speed wise. At least until you get into the really fast guys. If you are talking making the same power over a longer period (in MX that's moto minutes, others may consider miles the measurement stick), well that's another story.
I really don't know how long the stock battery actually lasts in the hands of an A or above class rider as thermal limiting almost always precedes a low SOC pack, sometimes at 30-50% mark. So, the first task is to deal with the thermal issues. That's why I spent 6 months designing, fabricating and experimenting with a liquid cooled pack. That's also why I know that simply cooling a pack with water (like the motor or Inverter) simply won't work on a hot day as there's simply not enough difference in temperature to pull that much heat out. Remember, an ICE motor runs at 200F or so, that's a 100F difference on a hot day. Li-Ion (depending on who you talk to) doesn't like over 130-140F. I've got data logs to prove it. That's why I went to pre-chilling of the pack. Start at a lower cell temp and you have much more headroom to grow in heat before they hit the limit. It also eliminates the need for radiators, pumps, etc, that only adds weight. On cold days, the water cooling alone works great, however. But why add all that weight when all you need is a water chiller (all shown in several of my posts).
Another important item to consider is how quickly any particular cell can move heat from its core to its case. Li-Ion, at least the cylindrical types, are like jelly rolls, with electrodes, current collectors and isolators in-between. The isolators are great at keeping heat from moving radially to the case, so by far the best path is axially. That's why Alta (and Varg) use the negative ends to transfer heat to their respective heatsinks. Note: The positive ends are also good radiators, but the cell connections make this difficult, although I was able to obtain dual ended heat transfer on several of my personally built packs. BUT EVEN this process (single ended) is extremely inefficient as only a tiny portion of the negative electrode current collector (the good heat conductor) is actually in contact with the bottom of the cylindrical case. It's a thermal choke point. Cells with higher C rating typically use thicker current collectors and more contact area so they can conduct more heat. But something's gotta give as there's now less space for the actual electrodes and chemicals that generate the power. Luckly, chemically the cells are getting slightly better and this often offsets the less space, so it's a net improvement thermal wise and the capacity doesn't suffer (much). The VERY BEST cylindrical cells thermally use an extended negative current collector which "bunch up" at the cell bottom making the contact area orders larger and increases thermal conductivity greatly. There's a special name for these, but I forgot . . . However, that takes space too and the opportunity for shorting (among other issues) is always there. Plus, they are very expensive when you can find them.
So, what's better thermally, 18650 or a larger cell format? I haven't done any simulations, but there's some logical conclusions. Since the larger cells (in dia and length) have longer and/or a more thermally resistive paths, you'd think they'd be worse. On the other hand, by shear mass they may be able to absorb more thermal energy before they reach the limit. But then watch out, as without any assistance the thermal momentum will cause the cell to overshoot and then it's trouble. The BMS and good cell characterization can predict and prevent this, but by that time is the extra thermal capacity worth it? I still like the 18650. It's been proven time after time to be the most efficient and effective size. However. it's more expensive to make a pack and that pack may not be as reliable due to the many more connections. The type and quality of the connections will determine total end to end resistance, but everything equal, less conns are better, advantage larger cells in this case.
Now energy is energy. If you want HP and a reasonable range (reasonable is different for everyone), you'll need a certain amount of it regardless of what cell or fuel you choose. With electric, it's all about converting that potential energy into the good kind of kinetic energy (HP) as opposed to heat. Modern electric motors/controllers are getting pretty darn good, 85-90%. Varg may claim they have an efficiency advantage, but there's not much left to squeeze and no system is 100% efficient.
But to make a long story longer, if I were to try and design an Alta pack for myself (my needs and speed), I'd probably aim for maybe 4.6kwh using 18650s with those "special" type cells I mentioned (if I could find them). I'd still need to do some pre-chilling as just getting the heat to the cells case is only half the battle, you gotta ultimately get the heat to the air somehow, particularly difficult on hot days. So, my case mods would need to be done. I'd figure out a way to isolate the P groups so that a very efficient TIM could be used between the cell and HS and add a simple safety feature to prevent shorting. I'd build the modules using the stock Alta honeycombs but build my own BMS system. Might have to charge through a separate port on the pack and "fake out" the system to think everything was nominal. Possibly a new SOC indicator would be required. With some help I'd figure a weight loss of maybe 10-12 lbs.
Anyway, all theory as it would take far more time and money to do than I'd want to invest!
