We need to distinguish between two different situation :
a) Training an African athlete. Altitude is their normal ambience, so their training is based on that constant. For example, when I use intervals of 1000m at 10000m race pace (connected with the PB in altitude), I recovery about 1'30" using 200m jogging (Wilson Kiprop, 27'26" last year, 10x1000m in 2'43" / 2'45" rec. 1'30"). If I connect the same type of workout with their hypothetical speed at sea level (in the above case, I supposed 26'40", therefore 2'40" per km), I open the recovery till 2'30", and the athlete walks. Of course, since not all the altitudes are the same and have the same effect, we need to give different parameters, according to the different situation, using this rule : after 2000m of altitude, we have to add to the recovery time about 10% for every 100m (for example, if we use 2' at 2000m, at 2500m recovery becomes 120" + 50% = 3'), of course running the same speed.
Instead, if we want to work inside a full aerobic area, we can maintain the same recovery time, slowing down the speed of 0.1 for every 100m (2'40" at 2000m becomes 2'45" at 2500). I never tried to go over 2700m, so I can't say if this rule can work also going higher. In any case, personally I'm not favourable to train at an altitude too much high, if not using only long easy run, in order to reach a better aerobic efficiency, for the reasons I explained in the other thread about doping in athletics.
When these athletes go to sea-level, I have only to maintain their high aerobic power (they don't stay at sea level longer than 3 weeks, so we don't need workouts too hard in that direction), I maintain the same recovery, and I ask them to run faster (for example, 10x1000 in 2'35" with 2' recovery, while I'm not interested in 10x1000 in 2'42" with 1'15").
In other words, I use sea level only for incresing the SPECIFITY, all the fundamental and special training was already carried out in altitude.
b) Athletes living and training usually at sea-level.
In this case, I recommand the following behavior when they go to altitude :
1) To use the first week for adaptation only, starting with easy run, and increasing every day volume and intensity. After one week, the athlete must be able to run the same volume used at sea level, with only little difference in the speed referred to the classic Aerobic Threshold of 2 mml, and a difference of about 5% in the speed referred to the classic AnT of 4 mml.
2) After one week, the athlete can start some interval. In this case, the rule is :
* for short intervals (200/300/400m) the athlete can use same speed and same recovery, but the total volume must not be higher than 70% (so, 10x300 in 43" rec. 1'30" can become 7x300).
* for medium intervals (500-800m) the athlete can reduce his speed of about 0.3 % referred on 100m speed
* for long intervals (1000-3000m) the athlete can reduce his speed about 0.5% referred on 100m (if you run 3000m in 9' = 18" per 100m, 0.5% is 18"9 per 100m, that means 3'09" per km = 9'27")
Of course, these are suggested data for athletes not use to train in altitude.
But, because our muscle fibres have specific "memory", after the first time every period in altitude becomes a little bit easier, and the athlete, in order to receive correct and useful stimula, need to increase his speed a little bit.
In any case, remember that altitude can be dangerous if you carry out lactic workouts with short recovery. The main effect of training in altitude is to enhance the AnT and consequentely to increase the Aerobic Power. All the tough lactic training must be carried out at sea level.