You're going to want to create some system of hysteresis for profit switching that takes a number of factors into account. For instance, HashCows offen jumps to coins which are *slightly* more profitable than the current coin, only to jump back a minute later with no gains to show for the switch, when it would be more profitable to simply stay on the current coin. Properly filtering this will need to bias the coin choices based on past history. Some other things to consider include:

- Price at multiple exchanges
- Time of exchange confirms (cryptsy can take hours, and end up trading the coins at a much lower rate)
- Average reject rate for users (for instance, FastCoin = high reject rate on my rig)
- Average profitability of the coin over time
- Coin parameters (finishing blocks vs switching now, etc)

Likely scenario would be to score all of these factors as a low-pass filtered average over several time windows, then assign weights to them for each time window and finally weight them against each other. Rather than determining the weight manually, you should consider a system that evaluates different weightings and finds the ideal one.

For instance, imagine we construct min/average/max profitability for each coin for various time windows (say, 20 minutes, hourly, 6 hour, days, etc, based purely on exchange value). Assuming we're using cryptsy, we can look at the average time between transfer and trade to determine a possible variance on price over that time (based on our min value during a time period matching the average trade time for cryptsy). This gives us a reasonable projection of price and variance by the time we get to the exchange. Factor in the historical reject rate for the coin (averaged by users), and perhaps the slope of the coin over a long time window (for instance, doge has been going down for days), and you get a pretty decent score for your coin. More importantly, it's unlikely that a small spike in the data (single bogus trade, etc) is going to cause the system to thrash over to the wrong coin.

Anyway, you get the idea..

Even though I mine altcoins, I really think the proliferation of alt-coins is bad for Bitcoin and the eco-system in general. Any value stored in altcoins is essentially an inflation on digital currencies as a whole, as the price of the dominant digital currency (BTC right now) is dictated by the supply and demand, so alt-coins are effectively a way to increase supply beyond the original mechanism BTC introduced. Of course, this all hinges on people believing they are valuable - much like any currency.

In the end, I think it's best if we have one dominant currency which has a reasonable distribution across the user base. Right now, due to the nature of the mining/difficulty process, all of the digital currencies tend to have skewed distributions where a very small number of users hold most of the coins. (This is the primary reason we see alt-coins created every day - launching a coin puts you in a profitable position). This trend will likely increase with the life of a currency, just as it has with USD (which is majority held by a small number of people), unless some form of redistribution system is added, such as a progressive tax (coin rot?). But that presents a major issue, since no one would move from BTC to something else if such a system was added. So, unless there is a compelling reason not to - I suspect we will see the continued launch of alt-coins, and if those alt-coins become somewhat successful (as LTC has), they will act as an inflationary mechanic on the dominant coin.